1 | K|NGP|N | 32836 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-10980XE Extreme Edition Processor | |
Detailed Result |
|||||
2 | Xtreme Addict | 29476 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
3 | Slinky PC | 28853 | NVIDIA Titan Xp (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
4 | wilsto27 | 27837 | NVIDIA Titan Xp (4x SLI) | Intel Core i9-7960X Processor | |
Detailed Result |
|||||
5 | TooFurrious | 27622 | NVIDIA Titan Xp (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
6 | qRigs | 27062 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
7 | Olli-2 | 26712 | NVIDIA Titan Xp (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
8 | racoon | 26498 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
9 | excesseve | 26129 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
10 | closest.num.2.0 | 25422 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
11 | babotow-2 | 25362 | NVIDIA Titan Xp (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
12 | VIBOX | 25237 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
13 | TarekGT | 25140 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
14 | 1000D | 25111 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
15 | Menthol | 25014 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
16 | Frankythecat | 24825 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-7900X Processor | |
Detailed Result |
|||||
17 | Kaapstad-2 | 24656 | NVIDIA Titan Xp (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
18 | Finnsk3 | 24592 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | AMD Ryzen Threadripper 1950X | |
Detailed Result |
|||||
19 | Alwaystrue | 24586 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
20 | BASE303 | 24568 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
21 | Adrian Bufan | 24551 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
22 | der8auer | 24335 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
23 | rog_333 | 24231 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
24 | moostasha | 24049 | NVIDIA Titan Xp (4x SLI) | Intel Core i9-7900X Processor | |
Detailed Result |
|||||
25 | ThirtyIR | 23966 | NVIDIA Titan Xp (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
26 | yee245 | 23874 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-9900KF Processor | |
Detailed Result |
|||||
27 | SmokeTechTips | 23742 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
28 | WinterMagician | 23498 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
29 | Zurv | 23381 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
30 | Gorod | 23284 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
31 | Charles Too Far | 23249 | NVIDIA GeForce RTX 2080 Ti (4x SLI) | Intel Core i9-9900K Processor | |
Detailed Result |
|||||
32 | X-COMputers.eu | 23185 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
33 | TriOx | 23128 | NVIDIA Titan Xp Collector’s Edition (4x SLI) | Intel Core i7-9800X Processor | |
Detailed Result |
|||||
34 | Dr.Jamitzky | 23101 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
35 | HRC Informatik GmbH | 23100 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
36 | HatsuneMiku | 23036 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
37 | aomsin2526 | 23017 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
38 | 66racer | 22868 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
39 | Sherick Yu | 22806 | NVIDIA Titan Xp (3x SLI) | Intel Core i9-7960X Processor | |
Detailed Result |
|||||
40 | tim.takala-2 | 22770 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-7900X Processor | |
Detailed Result |
|||||
41 | motherisis | 22544 | NVIDIA Titan Xp Collector’s Edition (4x SLI) | Intel Core i9-7900X Processor | |
Detailed Result |
|||||
42 | Windaway | 22475 | NVIDIA Titan Xp Collector’s Edition (4x SLI) | Intel Core i9-7900X Processor | |
Detailed Result |
|||||
43 | sam nelson | 22466 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-5960X Processor | |
Detailed Result |
|||||
44 | Pegazus | 22311 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
45 | gubben | 21924 | NVIDIA GeForce GTX 1080 Ti (3x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
46 | Rauf | 21881 | NVIDIA GeForce RTX 2080 Ti (2x SLI) | Intel Core i9-9900K Processor | |
Detailed Result |
|||||
47 | tatransky | 21813 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-7900X Processor | |
Detailed Result |
|||||
48 | 룡이만세 | 21792 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | AMD Ryzen Threadripper 1950X | |
Detailed Result |
|||||
49 | TWENTYFO4R | 21757 | NVIDIA GeForce GTX 1080 Ti (3x SLI) | Intel Core i9-10980XE Extreme Edition Processor | |
Detailed Result |
|||||
50 | Addiecool | 21664 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
51 | Cheddle | 21578 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i7-5930K Processor | |
Detailed Result |
|||||
52 | Khars | 21444 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6900K Processor | |
Detailed Result |
|||||
53 | twentyonemore | 21162 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-9900K Processor | |
Detailed Result |
|||||
54 | wolik | 21160 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i7-5930K Processor | |
Detailed Result |
|||||
55 | Gunslinger | 21038 | NVIDIA GeForce GTX 1080 Ti (3x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
56 | …::: God of Nukes :::… | 21024 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
57 | jiacheng.liu15 | 20981 | NVIDIA Titan Xp (3x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
58 | Xanvast | 20859 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-4960X Processor | |
Detailed Result |
|||||
59 | kmarch8411 | 20792 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
60 | babotow | 20778 | NVIDIA Titan Xp (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
61 | IQuit | 20730 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | AMD Ryzen Threadripper 2990WX | |
Detailed Result |
|||||
62 | KEYS | 20707 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Xeon Processor E5-2696 v4 | |
Detailed Result |
|||||
63 | mirkoj.animator | 20415 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i7-5960X Processor | |
Detailed Result |
|||||
64 | dbb1987 | 20397 | NVIDIA Titan X (Pascal) (3x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
65 | mgr5001 | 20390 | NVIDIA Titan Xp (3x SLI) | Intel Core i9-7900X Processor | |
Detailed Result |
|||||
66 | Adrian Bufan | 20371 | NVIDIA GeForce GTX 1080 Ti (3x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
67 | Gosu — Extreme Hardware | 20235 | NVIDIA GeForce RTX 2080 Ti (4x SLI) | Intel Core i9-7960X Processor | |
Detailed Result |
|||||
68 | M249MINIMI | 20228 | NVIDIA Titan Xp (3x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
69 | jason-2630 | 20192 | NVIDIA Titan X (Pascal) (3x SLI) | Intel Core i9-7960X Processor | |
Detailed Result |
|||||
70 | Kandagar | 20155 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
71 | minimax690 | 20146 | NVIDIA GeForce GTX 1080 Ti (3x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
72 | 老喵实验室 | 20082 | NVIDIA Titan Xp Collector’s Edition (4x SLI) | Intel Core i9-7900X Processor | |
Detailed Result |
|||||
73 | 副会長 兼 魔法使い@独身の会 | 20049 | NVIDIA Titan Xp (3x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
74 | Twinsen-4 | 20023 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
75 | ElitePanzerAce | 19889 | NVIDIA Titan RTX (2x SLI) | Intel Core i9-10980XE Extreme Edition Processor | |
Detailed Result |
|||||
76 | ssh6314 | 19735 | NVIDIA GeForce GTX 1080 Ti (3x SLI) | Intel Core i9-7940X Processor | |
Detailed Result |
|||||
77 | ttp202016 | 19698 | NVIDIA GeForce GTX 1080 (4x SLI) | Intel Core i7-5960X Processor | |
Detailed Result |
|||||
78 | 8 Pack | 19661 | NVIDIA GeForce GTX 1080 (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
79 | MasterJ | 19654 | NVIDIA Titan X (Pascal) (3x SLI) | Intel Core i9-10980XE Extreme Edition Processor | |
Detailed Result |
|||||
80 | svhartman | 19566 | NVIDIA GeForce GTX 1080 (4x SLI) | Intel Core i7-5960X Processor | |
Detailed Result |
|||||
81 | 494940501 | 19566 | NVIDIA Titan Xp Collector’s Edition (4x SLI) | AMD Ryzen Threadripper 1950X | |
Detailed Result |
|||||
82 | NEMA6-15 | 19545 | NVIDIA Titan Xp (4x SLI) | Intel Core i7-5960X Processor | |
Detailed Result |
|||||
83 | tdgames | 19518 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | AMD Ryzen Threadripper 1950X | |
Detailed Result |
|||||
84 | 흙수저의반란 | 19480 | NVIDIA Titan RTX (2x SLI) | Intel Xeon W-3175X Processor | |
Detailed Result |
|||||
85 | ZedVile | 19382 | NVIDIA GeForce RTX 2080 Ti (2x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
86 | wesleyxy | 19338 | NVIDIA GeForce GTX 1080 Ti (3x SLI) | AMD Ryzen Threadripper 1950X | |
Detailed Result |
|||||
87 | OGS | 19304 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
88 | Carlos Gonzales | 19251 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6850K Processor | |
Detailed Result |
|||||
89 | Vindictive000 | 19186 | NVIDIA Titan X (Pascal) (3x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
90 | KalmTraveler | 19116 | NVIDIA Titan RTX (2x SLI) | Intel Core i9-7960X Processor | |
Detailed Result |
|||||
91 | Radox | 19093 | NVIDIA GeForce GTX 1080 Ti (3x SLI) | Intel Core i7-5960X Processor | |
Detailed Result |
|||||
92 | Hyuuga Hanabi | 19059 | NVIDIA Titan Xp (3x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
93 | Dansv8 | 18998 | NVIDIA GeForce GTX 1080 Ti (3x SLI) | Intel Core i9-7900X Processor | |
Detailed Result |
|||||
94 | V!leAMD | 18995 | NVIDIA GeForce RTX 2080 Ti (2x SLI) | AMD Ryzen 9 3950X | |
Detailed Result |
|||||
95 | GWAUGH | 18991 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | AMD Ryzen Threadripper 1950X | |
Detailed Result |
|||||
96 | cryptonaut | 18945 | NVIDIA Titan X (Pascal) (4x SLI) | Intel Core i7-6850K Processor | |
Detailed Result |
|||||
97 | yuyijun22 | 18880 | NVIDIA GeForce RTX 2080 Ti (2x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
98 | jpmboy | 18859 | NVIDIA Titan RTX (2x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
99 | dismantler300 | 18829 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i7-6950X Processor | |
Detailed Result |
|||||
100 | My Little Pwny-2 | 18807 | NVIDIA GeForce RTX 2080 Ti (2x SLI) | AMD Ryzen 9 3950X | |
Detailed Result |
3DMark Basic Edition Download (2022 Latest)
3DMark is a popular gaming performance benchmark used by millions of people, hundreds of hardware review sites, and many of the world’s leading technology companies. Detailed charts show you how the CPU and GPU temperatures, clock speeds, and frame rate changed during your benchmark run. Explore your PC’s performance limits with custom benchmark settings.
You can change the rendering resolution and quality across a range of settings. Can your PC handle 4K gaming? Fire Strike Ultra’s 4K Ultra HD rendering resolution is four times larger than the 1080p resolution used in Fire Strike. A 4K monitor is not required, but your graphics card must have at least 3GB of memory to run this monstrously demanding benchmark. Download NOW!
It includes everything you need to benchmark your PC and mobile devices in one app. Whether you’re gaming on a desktop PC, laptop, notebook, or tablet, 3DMark for PC has a benchmark designed for your hardware.
3DMark Basic Edition includes seven benchmark tests:
Port Royal (NEW!) is the world’s first dedicated real-time ray tracing benchmark for gamers. You can use Port Royal to test and compare the real-time ray tracing performance of any graphics card that supports Microsoft DirectX Raytracing. As well as benchmarking performance, 3DMark Port Royal provides a realistic and practical example of what to expect from ray tracing in upcoming games.
Night Raid is a DirectX 12 benchmark for mobile computing devices with integrated graphics and low-power platforms powered by Windows 11 or Windows 10 on ARM. Run Night Raid 3D Mark 2022 Latest to test and compare laptops, notebooks, tablets, and the latest Always Connected PCs. The test is especially suitable for DirectX 12 systems that cannot achieve high frame rates in the more demanding Time Spy benchmark.
Wild Life is a cross-platform benchmark for Windows, Android, and Apple iOS. Use 3DMark Wild Life to test and compare the graphics performance of the latest notebook computers, tablets, and smartphones. Wild Life uses the Vulkan graphics API on Windows PCs and Android devices. On iOS devices, it uses Metal. You can compare benchmark scores across platforms.
Time Spy is a showcase DirectX 12 benchmark test for high-performance gaming PCs running Windows 10. With its pure DirectX 12 engine, built from the ground up to support new features like asynchronous compute, explicit multi-adapter, and multi-threading, Time Spy is the ideal benchmark for testing DirectX 12 performance. DirectX 12 benchmark for Windows 10. Renders at 2560 × 1440 resolution. Includes a Demo, Graphics tests, and a CPU test.
Time Spy Extreme is a new 4K DirectX 12 benchmark test, available in the app Advanced and Professional Editions. You don’t need a 4K monitor to run it, but you will need a GPU with at least 4 GB of dedicated memory. With its 4K Ultra HD rendering resolution, Time Spy Extreme is an ideal benchmark test for the latest high-end graphics cards. The CPU test has been redesigned to let processors with 8 or more cores perform to their full potential.
Fire Strike is a showcase DirectX 11 benchmark designed for today’s high-performance gaming PCs. It is very demanding, even for the latest graphics cards. If your frame rate is low, use Sky Diver instead.
Sky Diver is a DirectX 11 benchmark for gaming laptops and mid-range PCs. Use it to benchmark systems with mainstream graphics cards, mobile GPUs, or integrated graphics. It is especially suitable for systems that cannot achieve more than single-digit frame rates in the more demanding Fire Strike test.
Cloud Gate is designed for Windows notebooks and typical home PCs with integrated graphics. The benchmark uses a DirectX 11 engine limited to Direct3D feature level 10 making it suitable for testing DirectX 10 compatible hardware.
Ice Storm — Use 3D Mark Ice Storm to compare basic tablets and smartphones. Ice Storm includes two Graphics tests to measure GPU performance and a Physics test to stress CPU performance. Ice Storm uses DirectX 11 feature level 9 on Windows. On Android and iOS, it uses OpenGL ES 2.0. You can compare scores across platforms.
Use Ice Storm Extreme to compare low-cost smartphones and tablets running Android, iOS, and Windows (compatible with Windows 11). Ice Storm Extreme raises the rendering resolution from 1280 × 720 to 1920 × 1080 and uses higher quality textures and post-processing effects in the Graphics tests.
Feature Tests
API Overhead test
Games make thousands of draw calls per frame, but each one creates performance-limiting overhead for the CPU. APIs with less overhead can handle more draw calls and produce richer visuals. With the API Overhead feature test, you can compare the performance of Vulkan, DirectX 12, and DirectX 11. See how many draws calls each API can handle before the frame rate drops below 30 FPS.
DirectX Raytracing feature test
Real-time ray tracing is incredibly demanding. The latest graphics cards have dedicated hardware that’s optimized for ray-tracing. The 3DMark DirectX Raytracing feature test measures the performance of this dedicated hardware. Instead of using traditional rendering techniques, the whole scene is ray-traced and drawn in one pass. The result of the test depends entirely on ray-tracing performance.
PCI Express feature test
PCI Express (PCIe) is a standard interface that provides high-bandwidth communication between devices in your PC. New PCIe 4.0 interfaces provide up to twice the bandwidth of PCIe 3.0. With more bandwidth, games can transfer more data, reduce loading times, and support more complex scenes. The PCI Express feature test measures the bandwidth available to your GPU.
VRS feature test
Variable-Rate Shading (VRS) is a DirectX 12 feature that can improve performance by reducing detail in parts of the frame where it is unlikely to be noticed. The 3DMark VRS feature test helps you compare differences in performance and image quality when using Tier 1 and Tier 2 VRS. An interactive mode lets you change VRS settings on the fly and export frames for comparison.
NVIDIA DLSS feature test
Deep Learning Super Sampling (DLSS) is an NVIDIA RTX technology that uses the power of deep learning and AI to improve game performance while maintaining visual quality. The NVIDIA DLSS feature test helps you compare performance and image quality using DLSS 2 or DLSS 1. With DLSS 2, you can choose between three quality modes—Quality, Performance, and Ultra Performance.
System Requirements
All the currently supported benchmarks and feature tests in 3DMark are compatible with Windows 11.
To run Time Spy and Night Raid benchmarks, you need Windows 10, a graphics card that supports DirectX 12, and a processor that supports SSSE3. To run Port Royal, you must have the Windows 10 October 2018 Update (1809) and a graphics card that supports Microsoft DirectX Raytracing. Wild Life requires a GPU compatible with Vulkan 1.1.
The DirectX Raytracing feature test requires Windows 10 Version 2004 or later and a DirectX 12 compatible graphics card with driver support for DirectX Raytracing Tier 1.1. The PCI Express feature test requires a DirectX 12 compatible discrete graphics card. The VRS feature test requires Windows 10 version 1903 or later and a DirectX 12 GPU that supports Variable-Rate Shading. The NVIDIA DLSS feature test requires an NVIDIA graphics card that supports DLSS.
Note: Limited functionality in the unregistered version.
Download 3DMark Basic Edition Latest Version
1 | OGS | 63361 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
2 | biso biso | 62389 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
3 | K|NGP|N | 61206 | NVIDIA GeForce GTX 1080 Ti (4x SLI) | Intel Core i9-10980XE Extreme Edition Processor | |
Detailed Result |
|||||
4 | Seby9123 | 58860 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
5 | Tacha | 56716 | AMD Radeon RX 6950 XT | Intel Core i9-12900KS Processor | |
Detailed Result |
|||||
6 | [Phantom] | 56554 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
7 | Dragon Soop | 55727 | NVIDIA GeForce RTX 2080 Ti (2x SLI) | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
8 | nvidiafan | 55688 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
9 | jroche673 | 54769 | AMD Radeon RX 6950 XT (2x CrossFireX) | Intel Core i9-12900KS Processor | |
Detailed Result |
|||||
10 | snakeeyes | 54767 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
11 | Slinky PC | 54521 | AMD Radeon RX 6950 XT | Intel Core i9-12900KS Processor | |
Detailed Result |
|||||
12 | wimpievanzyl | 54169 | AMD Radeon RX 6950 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
13 | mantiz | 54149 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
14 | RSR | 54038 | AMD Radeon RX 6900 XT | Intel Core i9-12900KS Processor | |
Detailed Result |
|||||
15 | Butanding1987 | 53702 | AMD Radeon RX 6950 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
16 | Jimshown LMHF | 53528 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
17 | ericc64 | 53491 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
18 | dartagnan77 | 53107 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
19 | Rauf | 53081 | NVIDIA GeForce RTX 2080 Ti (2x SLI) | Intel Core i9-9980XE Processor | |
Detailed Result |
|||||
20 | Auh50004 | 53062 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
21 | Claviger | 52985 | AMD Radeon RX 6950 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
22 | Gawl86 | 52705 | AMD Radeon RX 6900 XT | Intel Core i9-12900KS Processor | |
Detailed Result |
|||||
23 | Kathum | 52427 | AMD Radeon RX 6950 XT | Intel Core i9-12900KF Processor | |
Detailed Result |
|||||
24 | ROSCOE_IS_HIGH | 52333 | NVIDIA GeForce RTX 2080 Ti (2x SLI) | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
25 | CosimoCarlino | 52316 | AMD Radeon RX 6950 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
26 | earphonelnwshop | 52217 | AMD Radeon RX 6900 XT | Intel Core i9-12900KS Processor | |
Detailed Result |
|||||
27 | bl4ckdot | 52040 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
28 | Jigfio | 51880 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
29 | D3LTA KING | 51617 | AMD Radeon RX 6950 XT | Intel Core i9-12900KS Processor | |
Detailed Result |
|||||
30 | Gunslinger | 51418 | NVIDIA GeForce RTX 2080 Ti (2x SLI) | Intel Core i9-9980XE Processor | |
Detailed Result |
|||||
31 | LunerK9 | 51396 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
32 | MCKarlMarx | 51091 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
33 | bengtibacken | 51028 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
34 | kotaro2 | 51028 | AMD Radeon RX 6900 XT | Intel Core i9-12900KS Processor | |
Detailed Result |
|||||
35 | papersleevesJTF | 50994 | AMD Radeon RX 6950 XT | Intel Core i9-12900KF Processor | |
Detailed Result |
|||||
36 | Table21 | 50969 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
37 | Kawarius | 50921 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
38 | -=Popeye=- | 50673 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
39 | Kirederf | 50594 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
40 | 흙수저의반란 | 50461 | NVIDIA Titan RTX (2x SLI) | Intel Xeon W-3175X Processor | |
Detailed Result |
|||||
41 | bob(nz) | 50366 | NVIDIA GeForce GTX 1080 Ti (2x SLI) | Intel Core i9-7980XE Processor | |
Detailed Result |
|||||
42 | XOC_JOE | 50011 | AMD Radeon RX 6900 XT | AMD Ryzen 7 5800X | |
Detailed Result |
|||||
43 | $SOLID$ Necro | 50008 | AMD Radeon RX 6950 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
44 | VShpotov | 49946 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
45 | K0obies | 49815 | NVIDIA GeForce GTX 1080 Ti (2x SLI) | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
46 | Mahoro Hosizuki | 49749 | AMD Radeon RX 6900 XT | Intel Core i9-12900KF Processor | |
Detailed Result |
|||||
47 | rootWAShere | 49726 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
48 | itried | 49693 | AMD Radeon RX 6950 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
49 | Black_Death223 | 49683 | AMD Radeon RX 6950 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
50 | Secret Nugget | 49621 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
51 | RebelHubby | 49583 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
52 | NicklasAPJ @ Hwt.dk | 49569 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
53 | ccdragon6 | 49477 | AMD Radeon RX 6900 XT | Intel Core i9-12900KS Processor | |
Detailed Result |
|||||
54 | caliente134 | 49472 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
55 | GaryHTX | 49458 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
56 | OMG de CCL | 49371 | NVIDIA GeForce RTX 2080 Ti (2x SLI) | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
57 | Madoka-2 | 49310 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
58 | sugi0lover | 49298 | NVIDIA GeForce RTX 3090 Ti | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
59 | passionne | 49167 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
60 | navigator_111 | 49055 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
61 | Taelso | 48984 | AMD Radeon RX 6950 XT | Intel Core i9-12900KS Processor | |
Detailed Result |
|||||
62 | YOYH | 48823 | AMD Radeon RX 6950 XT | Intel Core i7-12700K Processor | |
Detailed Result |
|||||
63 | johnksss | 48822 | NVIDIA GeForce RTX 3090 | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
64 | Russ739 | 48809 | AMD Radeon RX 6900 XT | Intel Core i9-12900KF Processor | |
Detailed Result |
|||||
65 | ZEROGOGO | 48776 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
66 | MsiMax_Overclocker | 48773 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
67 | PilsenPowerLab | 48677 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
68 | PCtoPCGames | 48654 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
69 | [OCN] xR00Tx | 48588 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
70 | geriatricpollywog | 48579 | NVIDIA GeForce RTX 3090 | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
71 | VikTOR | 48500 | AMD Radeon RX 6800 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
72 | NorCal-George | 48496 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
73 | redstone_sam123 | 48484 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
74 | RG_Devil-Child | 48401 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
75 | spit986 | 48378 | AMD Radeon RX 6900 XT | Intel Core i9-12900KS Processor | |
Detailed Result |
|||||
76 | SoldierSnubby | 48352 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
77 | XE0 | 48350 | NVIDIA GeForce RTX 3090 | Intel Core i9-12900KF Processor | |
Detailed Result |
|||||
78 | Nando Gomes | 48344 | AMD Radeon RX 6900 XT | Intel Core i9-12900KF Processor | |
Detailed Result |
|||||
79 | nicolai141 | 48318 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
80 | finnOC | 48220 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5900X | |
Detailed Result |
|||||
81 | wisdom-3 | 48198 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
82 | mmO | 48195 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
83 | Xpand3 | 48195 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
84 | Dreadzone | 48162 | AMD Radeon RX 6900 XT | Intel Core i9-12900KF Processor | |
Detailed Result |
|||||
85 | jiacheng.liu15 | 48126 | NVIDIA Titan Xp (2x SLI) | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
86 | Kingjohn | 48030 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
87 | 中野五月-2 | 48012 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
88 | ♥KuroNeko♥ | 47956 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
89 | thethirdeye462 | 47949 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
90 | Umbra_Venator | 47904 | AMD Radeon RX 6900 XT | AMD Ryzen 7 5800X3D | |
Detailed Result |
|||||
91 | Subamax | 47899 | NVIDIA GeForce RTX 3090 Ti | Intel Core i9-12900KS Processor | |
Detailed Result |
|||||
92 | LtMatt81 | 47875 | AMD Radeon RX 6950 XT | AMD Ryzen 7 5800X3D | |
Detailed Result |
|||||
93 | DroiDMester | 47859 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
94 | affxct | 47850 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
95 | hrloveforever | 47830 | AMD Radeon RX 6900 XT | Intel Core i9-12900K Processor | |
Detailed Result |
|||||
96 | 普普通通Tony大叔 | 47757 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
97 | True Monkey | 47755 | AMD Radeon RX 6900 XT | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
98 | SΑFEDISK | 47725 | NVIDIA GeForce RTX 3090 | AMD Ryzen 9 5950X | |
Detailed Result |
|||||
99 | jay73040700 | 47724 | NVIDIA GeForce RTX 3090 Ti | Intel Core i9-12900KS Processor | |
Detailed Result |
|||||
100 | zippyt3k | 47687 | NVIDIA GeForce RTX 3090 | Intel Core i9-12900KS Processor | |
Detailed Result |
Download 3DMark Basic Edition Download (2020 Latest) for Windows 10, 8, 7 — Heaven32
3DMark is a popular one game performance benchmark used by millions of people, hundreds of hardware review sites and many of the world’s leading technology companies. Detailed graphs show how CPU and GPU temperatures, clock speeds, and frame rate changed during your run. Explore your PC’s performance limits with custom reference parameters. You can change the resolution and rendering quality in various settings. Can your PC handle 4K games? Fire Strike Ultra’s 4K Ultra HD rendering resoluti on is four times higher than the 1080p resolution used in Fire Strike. You don’t need a 4K monitor, but your graphics card must have at least 3GB of memory to run this monstrously demanding benchmark. Download now!
It includes everything you need to Compare Your PC | and mobile devices in a single application. Whether you’re playing on a desktop, laptop, notebook or tablet PC, 3DMark for PC has a benchmark designed for your hardware.
3DMark Basic Edition includes seven reference tests:
Night Raid (NEW) it is a DirectX 12 reference for mobile computing devices with integrated graphics and low-power platforms powered by Windows 10 on ARM. Run Night Raid 3D Mark 2020 Latest news to test and compare laptops, notebooks, tablets and the latest PCs always connected. The test is especially suitable for DirectX 12 systems that cannot reach high frame rates at the more demanding Time Spy benchmark.
Time Spy is a benchmark test of DirectX 12 for high-performance gaming computers running Windows 10. With its pure DirectX 12 engine, built from the ground up to support new features like asynchronous computing, multiple adapter explicit and multi-threading, the ideal benchmark for testing the performance of DirectX 12. DirectX 12 reference for Windows 10. It renders at a resolution of 2560 × 1440. Includes a demo, graphics testing, and CPU testing .
Time Spy Extreme is a new 4K DirectX 12 reference test, available in the Advanced and Professional Editions application. You don’t need a 4K monitor to run it, but you’ll need a GPU with at least 4GB of dedicated memory. With its 4K Ultra HD rendering resolution, Time Spy Extreme is an ideal benchmark for the latest high-end graphics cards. The CPU test has been redesigned to allow processors with 8 or more cores to run at full capacity.
Fire strike is a DirectX 11 reference showcase designed for today’s high-performance gaming computers. It is very demanding, even for the latest graphics cards. If the frame rate is low, use Sky Diver.
Sky Diver it is a DirectX 11 benchmark for laptops for gaming and mid-range computers. Use it to compare systems with conventional graphics cards, mobile GPUs, or integrated graphics. It is especially suitable for systems that cannot reach frame rates greater than one digit on the most demanding Fire Strike test.
Cloud Gate is designed for Windows laptops and typical home computers with integrated graphics. The benchmark uses a DirectX 11 engine limited to the level 10 Direct3D function, making it suitable for testing DirectX 10 compatible hardware.
Ice storm – Use 3D Mark Ice Storm to compare basic tablets and smartphones. Ice Storm includes two graphics tests to measure GPU performance and a physics test to highlight CPU performance. Ice Storm uses the Level 9 DirectX 11 feature in Windows. On Android and iOS, use OpenGL ES 2.0. You can compare scores between platforms.
Use Extreme ice storm to compare low-cost smartphones and tablets in operation Android, iOS and Windows. Ice Storm Extreme increases the rendering resolution from 1280 × 720 to 1920 × 1080 and uses superior quality textures and post-processing effects in graphics testing.
General API function test: Games make thousands of calls per frame, but each creates an overload that limits CPU performance. Less overhead APIs can handle more drawing calls and produce richer visuals. The 3D Mark 3D API Feature Test is a standalone test to compare the performance of Vulkan, DirectX 12, and DirectX 11. See how many handling calls your PC can handle with each API before the frame rate drops below 30 FPS.
Stress tests is a great way to check the reliability and stability of your system after buying or building a new PC, upgrading your graphics card or overclocking your GPU. It can help you identify faulty hardware or the need for better cooling. If the GPU crashes, hangs, or produces visual artifacts during testing, it may indicate a reliability or stability issue. If it heats up and turns off, you may need more cooling on your computer.
Note: Functionality limited to the unregistered version.
3DMark (Advanced Edition) » FREE DOWNLOAD
11359
Here you can download 3DMark (Advanced Edition) for free! On this page you will find information about 3DMark (Advanced Edition) and how you can download the software for free. Here you get the direct link (from different filehoster) or a torrent download. The link to the free download can be found at the bottom of the page.
3DMark includes everything you need to benchmark your hardware. With its wide range of benchmarks, you can test everything from tablets and notebooks to the latest 4K gaming PCs.
3DMark is for gamers, overclockers and system builders who want to get the most out of their hardware. When you buy 3DMark on Steam you unlock all the tests, settings, and features in 3DMark Advanced Edition.
Better benchmarking with 3DMark
The right test every time
We’ve made it easy to find the right test for your PC. 3DMark automatically scans your hardware and recommends the best benchmark for your system.
Choose your tests
3DMark grows bigger every year with new tests. When you buy 3DMark from Steam, you can choose to install only the tests you need.
Hardware monitoring
Detailed charts show you how the CPU and GPU temperatures, clock speeds, and the frame rate changed during your benchmark run.
Custom settings
Explore your PC’s performance limits with custom benchmark settings. Make the tests more or less demanding by changing the resolution and other settings.
DirectX 12 Benchmark Tests
Time Spy – DirectX 12 benchmarks for Windows 10 PCs
With its pure DirectX 12 engine, Time Spy is the ideal benchmark for testing the DirectX 12 performance of modern gaming PCs. Time Spy Extreme, the world’s first 4K DirectX 12 benchmark test, is perfect for benchmarking the latest high-end graphics cards and new processors with 8 or more cores.
Night Raid – DirectX 12 test for integrated graphics
Night Raid is a DirectX 12 benchmark for testing and comparing PCs with integrated graphics and Always Connected PCs powered by Windows 10 on ARM. This is an ideal test for laptops, notebooks, tablets and other mobile computing devices.
Port Royal – Real-time ray tracing benchmark
3DMark Port Royal is the world’s first real-time ray tracing benchmark for gamers. Use Port Royal to test and compare the real-time ray tracing performance of graphics cards that support Microsoft DirectX Raytracing.
DirectX 11 Benchmark Tests
Fire Strike – For high-performance gaming PCs
Fire Strike is a showcase DirectX 11 benchmark designed for high-performance gaming PCs. Use Fire Strike to test dedicated gaming PCs, Fire Strike Extreme for high-end systems with multiple GPUs (Crossfire/SLI), or see if your PC is ready for 4K gaming with Fire Strike Ultra.
Sky Diver – For gaming laptops and mid-range PCs
Sky Diver is a DirectX 11 benchmark that is ideal for testing mainstream graphics cards, mobile GPUs, integrated graphics and other systems that cannot achieve double-digit frame rates in the more demanding Fire Strike test.
DirectX Legacy Benchmark Tests
Use Cloud Gate to test the performance of older notebooks and home PCs with DirectX 10 compatible hardware. Use Ice Storm and Ice Storm Extreme to compare DirectX 9 compatible notebooks and tablets.
Feature Tests
NVIDIA DLSS feature test – Compare DLSS performance and image quality
Deep Learning Super Sampling (DLSS) is an NVIDIA RTX technology that uses the power of deep learning and AI to improve game performance while maintaining visual quality. This feature test helps you compare performance and image quality with and without DLSS.
API Overhead feature test – Compare the latest graphics APIs
The 3DMark API Overhead feature test is the world’s first independent test for comparing the performance of DirectX 12, DirectX 11 and Vulkan. See how many draw calls your PC can handle with each API before the frame rate drops below 30 FPS.
Stress Tests
Check the stability of your system
Stress testing is a good way to check the reliability and stability of your system after buying or building a new PC, upgrading your graphics card, or overclocking your GPU. Stress testing can also identify faulty hardware or the need for better cooling.
Main features
- Easy to use, no technical know-how needed.
- Works with all types of PC, from desktops to tablets.
- Includes all benchmarks, stress tests, and feature tests.
- Explore your PC’s performance limits with custom runs.
- In-depth hardware monitoring graphs.
- Automatically save your results offline.
Demo features
- Test your PC with Time Spy, Night Raid, Fire Strike, Sky Diver, Cloud Gate, and Ice Storm.
- Other benchmarks, features, and settings are locked.
- OS:Windows 7 SP1
- Processor:1.8 GHz dual-core
- Memory:2 GB RAM
- Graphics:DirectX 9*
- Storage:3 GB space
FREE DOWNLOAD
3DMark Download Latest Version – {Updated In 2022}
Download the version 3DMark Basic for Windows – 3DMark is the most popular benchmark in the world. 3DMark is the latest version of a benchmark designed to measure computer hardware performance, this update adds the new DirectX 12 Time Spy benchmark. This release includes different tests, each designed for a specific type of hardware, from smartphones to high-end gaming computers.
3DMark is a gaming performance benchmark popular used by millions of people, hundreds of hardware review sites and most of the world’s leading technology companies . Detailed graphs show how CPU and GPU temperatures, clock rates and frame rate changed during the benchmark run. Explore your PC’s performance limits with custom benchmark settings . You can change the rendering resolution and quality in a variety of settings. Can your PC play games in 4K? Fire Strike Ultra’s 4K Ultra HD rendering resolution is four times the 1080p resolution used in Fire Strike. A 4K monitor is not required, but your graphics card must have at least 3GB of memory to run this extremely demanding test. Download now!
3DMark works by running intensive graphics and computational tests on your hardware. The more powerful your hardware, the smoother the tests will run. Don’t be surprised if your frame rates are low, as 3DMark tests are very demanding. Each test provides a score that you can use to compare similar devices and systems.
3DMark Port Royal
Real-time ray tracing promises to bring new levels of reality to the game’s graphics. Port Royal uses DirectX Raytracing to enhance reflections, shadows and other effects that are difficult to achieve with traditional rendering techniques. In addition to its benchmark performance, 3DMark Port Royal is a practical and realistic example of what can be expected from ray tracing in upcoming games – ray tracing effects running in real-time at reasonable frame rates at 2560× resolution. 1440. 3DMark Port Royal was developed in collaboration with AMD, Intel, NVIDIA and other leading technology companies. We worked closely with Microsoft to create a premium implementation of the DirectX Raytracing API. Port Royal will work on any graphics card that has DirectX Raytracing compatible drivers.
3DMark Time Spy
Developed in collaboration with AMD, Intel, Microsoft, NVIDIA and other members of our Benchmark Development Program, 3DMark Time Spy is one of the first DirectX 12 applications to be built from the ground up “correctly” to achieve full performance. what the new API offers wins. Introduced with Windows 10, DirectX 12 is a low-level graphics API that reduces processor overhead. With less overhead and better use of modern GPU hardware, the DirectX 12 game engine can draw more objects, textures and effects on the screen. How much more? Take a look at the chart below comparing Time Spy to Fire Strike, a top-notch DirectX 11 test. 3DMark works by running intensive graphics and computational tests on your hardware. The more powerful your hardware, the smoother the tests will run. Don’t be surprised if your frame rates are low, as 3DMark tests are a lot of work. Each test provides a score that you can use to compare similar devices and systems. With its pure DirectX 12 engine supporting new API features such as asynchronous computing, open multi-adapter and multi-threading, 3DMark Time Spy is an ideal benchmark for testing the DirectX 12 performance of the latest graphics cards.
3DMark – Player Benchmark
3DMark is a popular benchmarking application used by millions of gamers, hundreds of hardware review sites and many leading technology companies. With its wide range of tests, you can compare anything from tablets and laptops to the latest high-quality 4K gaming computers. Time Spy is available as a free upgrade for all versions of 3DMark for Windows, including 3DMark Basic Edition and the Steam Demo.
3DMark Basic Edition / Steam Demo – FREE
Compare your PC with various tests like Time Spy and Fire Strike.
This is a small update to fix issues reported by some users. Benchmarks are not affected by an exception – see the section on custom Fire Strike runs below for details.
We have released a major update to 3DMark for Windows that adds Night Raid, a new DirectX 12 benchmark for laptops, laptops and tablets with integrated graphics. Night Raid also has native ARM support for the latest Always Connected computers supported by Windows 10 on ARM. Night Raid is available as a free update to all 3DMark users. However, please note that custom benchmark settings and the Night Raid Stress Test are only available in 3DMark Advanced Edition and 3DMark Professional Edition.
- DirectX 12 testing requires a computer running an updated version of Windows 10 Technical Preview (10041 or higher), 4 GB of system memory, and DirectX 11_0 feature level compatible hardware with at least 1 GB of graphics memory.
- The DirectX 11 test requires DirectX 11_0 feature level compatible hardware with at least 1 GB of graphics memory and 4 GB of system memory.
3DMark Basic Edition Download (Latest) for Windows 11 10 8 7
3DMark is a popular gaming benchmark that is used by millions of people, hundreds of hardware review sites, and many of the world’s top technology companies. Detailed charts show how CPU and GPU temperatures, clock speeds, and frame rates changed during the test. Explore the performance limits of your PC with custom test settings. You can change the resolution and rendering quality with a number of settings. Can your computer support 4K games? The resolution of 4K Ultra HD rendering in Fire Strike Ultra is four times the resolution of 1080p used in Fire Strike. A 4K monitor is not required, but your graphics card must have at least 3GB of memory to perform this monstrously demanding test. Download now!
It includes everything you need to test your PC and mobile devices in one app. Whether you’re playing on a desktop PC, laptop, laptop or tablet, 3DMark for PC has a test designed for your hardware.
3DMark Basic Edition includes seven benchmarks:
Port Royal (NEW!) is the world’s first real-time ray tracing benchmark for gamers. You can use Port Royal to test and compare the real-time ray tracing performance of any graphics card that supports Microsoft DirectX Raytracing. Performance testing aside, 3DMark Port Royal is a realistic and practical example of what to expect from ray tracing in future games.
Night Raid is a DirectX 12 benchmark for mobile computing devices with integrated graphics and low-end platforms running Windows 10 on ARM. Launch Night Raid 3D Mark 2021 Latest to test and compare laptops, laptops, tablets, and the latest Always Connected PCs. The test is particularly suited to DirectX 12 systems that cannot achieve high frame rates in the more demanding Time Spy test.
Wild Life is a cross-platform test for Windows, Android and Apple iOS. Use 3DMark Wild Life to test and compare the graphics performance of the latest laptops, tablets and smartphones. Wild Life uses the Vulkan graphics API on Windows PCs and Android devices. On iOS devices, it uses Metal. You can compare test results across different platforms.
Time Spy is a DirectX 12 benchmark demo for high-end gaming PCs running Windows 10. Time Spy is a pure DirectX 12 engine built from the ground up to support new features such as asynchronous computing, explicit multithreading, and multithreading. perfect benchmark for testing DirectX 12 performance. DirectX 12 test for Windows 10. Rendering at 2560 × 1440 resolution. Includes demo, graphics tests, and processor test.
Time Spy Extreme is a new 4K DirectX 12 performance benchmark available in the Advanced and Professional Edition applications. It doesn’t need a 4K monitor to run, but you will need a GPU with at least 4GB of dedicated memory. Time Spy Extreme with 4K Ultra HD rendering resolution is the perfect benchmark for the latest high performance graphics cards. The CPU test has been redesigned to allow processors with 8 or more cores to reach their full potential.
Fire Strike is a DirectX 11 demo benchmark designed for today’s high-end gaming PCs. This is very demanding even for the latest graphics cards. If you have low frame rates, use Sky Diver instead.
Sky Diver is a DirectX 11 test for mid-range gaming laptops and PCs. Use it to test systems with conventional graphics cards, mobile GPUs, or integrated graphics. It’s particularly suited to systems that can’t deliver more than single digit frame rates in the more demanding Fire Strike test.
Cloud Gate is designed for Windows laptops and regular home PCs with integrated graphics. The benchmark uses the DirectX 11 engine, limited to Direct3D 10 features, making it suitable for testing DirectX 10 compatible hardware. Ice Storm includes two graphics benchmarks to measure GPU performance and a physical benchmark to verify CPU performance. Ice Storm uses DirectX 11 feature level 9on Windows. On Android and iOS it uses OpenGL ES 2.0. You can compare scores across platforms.
Use Ice Storm Extreme to compare low cost smartphones and tablets running Android , iOS and Windows. Ice Storm Extreme upscales the render resolution from 1280×720 to 1920×1080 and uses better textures and post-processing effects in graphics benchmarks.
Function tests Service data test
API
Games make thousands of draw calls per frame, but each one creates overhead that limits CPU performance. APIs with less overhead can handle more draw calls and produce richer visuals. With the API overhead testing feature, you can compare the performance of Vulkan, DirectX 12, and DirectX 11. See how many draw calls each API can handle before the frame rate drops below 30 fps.
Testing the DirectX Raytracing feature
Real-time ray tracing is incredibly tricky . The latest graphics cards have dedicated hardware optimized for ray tracing. The 3DMark DirectX Raytracing test measures the performance of this specialized hardware. Instead of using traditional rendering methods, the entire scene is traced and rendered in one pass. The result of the test depends entirely on the performance of ray tracing.
Feature test
PCI Express PCI Express (PCIe) is a standard interface that provides high-speed communication between devices on your PC. The new PCIe 4.0 interfaces offer twice the bandwidth of PCIe 3.0. With more bandwidth, games can transfer more data, reduce load times, and support more complex scenes. The PCI Express feature test measures the bandwidth available to your GPU.
VRS Feature Test
Variable-Rate Shading (VRS) is a DirectX 12 feature that can improve performance by reducing detail in parts of the frame that are unlikely to be seen. The 3DMark VRS feature test helps you compare performance and image quality differences between Level 1 and Level 2 VRS. Live mode allows you to change VRS settings on the fly and export footage for comparison.
NVIDIA DLSS 9 Feature Test0004
Deep Learning Super Sampling (DLSS) is an NVIDIA RTX technology that uses deep learning and AI to improve game performance while maintaining image quality. The NVIDIA DLSS Functional Test helps you compare performance and image quality using DLSS 2 or DLSS 1. With DLSS 2, you can choose from three quality modes — Quality, Performance, and Ultra Performance.
System requirements
To run the Time Spy and Night Raid tests, you need Windows 10, a DirectX 12 capable graphics card, and a processor that supports SSSE3. To run Port Royal, you must have Windows 10 October 2018 Update (1809) and a graphics card that supports Microsoft DirectX Raytracing. Wild Life requires a Vulkan 1.1 compatible GPU.
DirectX Raytracing feature test requires Windows 10 version 2004 or later and a DirectX 12 compatible graphics card with DirectX Raytracing Tier 1.1 drivers. PCI Express feature test requires DirectX 12 compatible discrete graphics card. VRS feature test requires Windows 10 version 1903 or later and a DirectX 12 GPU that supports variable rate shading. The NVIDIA DLSS feature test requires an NVIDIA graphics card that supports DLSS.
Note : Limited functionality in unregistered version.
3DMark Gained Gaming Performance News — NVIDIA WORLD
The latest update to UL Benchmarks’ benchmarking utility includes a new feature that allows users to evaluate their system’s performance in various games.
Intel Arc Alchemist mobile graphics cards have started to appear in laptops, but in reality they can be found in the only model — Samsung Book 2 Pro, which is sold exclusively in Korea.
In addition to Arc A350M graphics, this laptop also features a Core i7-1260P processor, 1080p AMOLED screen, 32GB DDR5 memory and a 1TB SSD. The Arc A350M GPU contains 6 Xe cores, 6 trace units, clocked at 1150 MHz. The video memory subsystem is limited to a 64-bit bus and 4 GB GDDR6.
GPU Intel Arc
But what about the performance of this graphics? It turned out that this card is in the expected range. At 30W, it was 14% slower than the NVIDIA GTX 1650 in 3DMark Time Spy and 22% slower in Fire Strike. At the same time, the Arc A350M turned out to be 17% faster than the entry-level MX450 card from NVIDIA, while the TDP is also 17% higher. In the 3DMark Port Royal test, the Intel card scored 200 points.
Arc 350M test results
Of course, for a 6 nm accelerator, the performance turned out to be quite mediocre. However, the company continues to work on driver optimization, increasing frequencies, so by June, when these accelerators become more affordable, their speed may noticeably increase.
Testing GPUARCINTEL ARC 350M3DMARKVIDEOKOKOTKYMARKICARY
Comment similar news
GURU
UL Benchmarks produces a new add -ons, which is designed to test the SSDs, using the SSDs, using the SSDs, using the SSD, which is used to test the actual gaming. in the game environment.
Thanks to the new tool, users will be able to really assess the speed of the drive. For testing, 7 tasks are performed, all of them are performed three times to reduce the measurement error. These tasks include loading times for games, recording 1080p gameplay using OBS, installing a game on a PC, saving a game, and copying a folder from an external SSD to the system drive. By using this tool, gamers can evaluate the impact of SSDs on gaming performance.
Drive Benchmark by 3DMark
The new tool also gives testers access to reviews of upcoming SSDs. With this analysis, it will be possible to highlight the best SSDs for gamers using real data.
As mentioned above, 3DMark Storage Benchmark uses seven ways to evaluate performance:
- Battlefield V loading from startup to main menu;
- Call of Duty: Black Ops 4 loading from launch to main menu;
- loading Overwatch launch before the main menu;
- 1080p 60fps gameplay video recording with OBS (Open Broadcaster Software) while playing Overwatch;
- installing The Outer Worlds from the Epic Games Launcher;
- copying the Counter-Strike: Global Offensive Steam folder from an external SSD to the system drive.
testing3DMarkSSDbenchmarks
comment on related news
Overclock 3D
Many people are looking forward to the release of Valve’s Steam Deck portable computer based on a custom AMD processor. And someone from China, probably close to production, got hold of a copy of this machine to run various game tests on it.
It is noted that the software part of the new machine is far from ready. The interface is not translated into other languages, but there is English.
The hardware is exactly the same as the early rumors. The main thing is that computer performance tests in games have been carried out, and this is very interesting.
Steam Deck
4 games were tested: Shadow of the Tomb Raider, Doom, Cyberpunk 2077 and DOTA 2. Cyberpunk 2077 showed strange performance. The frame rate fluctuated from 20 to 30 fps, however, at high settings. Other games worked much better. So, DOTA 2 on high settings gave out 47 fps and 80 fps on low ones. DOOM hit 60 fps on medium, and when I increased the settings, the frequency dropped to 46 fps. Finally, Shadow of the Tomb Raider ran decently as well, running at 36fps on the built-in benchmark, reaching 60fps on optimal settings.
We can conclude that the performance of the Steam Deck was acceptable, and given that the quality is of secondary importance on a 7” screen, we can expect an excellent frame rate.
testcomputersSteam DeckgamesSteambenchmarks
comment on related news
The evaluation is carried out with a different number of computational threads, which allows you to find out the performance of your CPU in various tasks.
Maximum thread test — is a test of the full performance of the processor and demonstrates its full computing potential.
3DMark CPU Profile
- 16-thread test — 16 threads is a good measure of performance in heavy tasks such as digital content creation and 3D rendering.
- 8-Thread Test — Evaluates performance in modern DirectX 12 games.
- 4-Thread and 2-Thread Test — Evaluates performance in older DirectX 9 games that use mostly 2 or 4 processing threads.
- 1-thread test is a fundamental measure of processor performance.
You can get new tests as part of 3DMark by installing the package or updating it.
testingUL3DMarkprocessorsbenchmarksutilities
comment on related news
TweakTown
Benchmarks of the upcoming NVIDIA RTX 3060 Ti graphics card have leaked online, showing that the card has amazing performance.
Published screenshots of map testing in 3DMARK Fire Strike and 3DMARK Time Spy. The NVIDIA RTX 3060 Ti card is rumored to launch on December 2nd for a suggested price of under $400, making it the most affordable Ampere-based card to date. At the same time, the accelerator has remarkable performance. If the pictures are correct, then the RTX 3060 Ti even beats the RTX 2080 Super, which had a starting price of $700.
NVIDIA RTX 3060 Ti
Based on the above images, Videocardz concluded that the RTX 3070 is only 11-13% faster than the RTX 3060 Ti. And with factory overclocking, that gap can close even further, making the RTX 3060 Ti almost perfect for gamers.
NVIDIA RTX 3060 Ti test in 3DMark Fire StrikeNVIDIA RTX 3060 Ti test in 3DMark Time Spy time rather than being sold out in minutes, as is the case with most gaming systems these days.
testingrumors3DMarkvideo cardsNVIDIAGeForce RTX 3060 Tibenchmarks
comment on related news Last week we reported that the overclocker Lucky_n00b set a record for its performance in the air. It was only a matter of time before someone applied liquid nitrogen.
The new leader of the 3DMark Fire Strike leaderboard has the nickname Takukou. He overclocked the RX 6800 XT card to 2.80 GHz and the memory to 2150 MHz (17.2 Gb/s). This configuration allowed us to set a Fire Strike performance record of 48,890 points.
Overclocking RX 6800 XT
In another attempt, I overclocked the card to 2. 75 GHz and recorded 49,456 points in HWBot 3DMark Fire Strike. In both cases, he used liquid nitrogen to cool the graphics card. It also cooled the AMD Ryzen 9 5950X processor, which ran on all active cores at a frequency of 5.60 GHz. MSI MEG X570 GODLIKE motherboard with 32 GB DDR4-3800 memory was used as a platform.
RX 6800 XT overclocking results RX 6800 XT overclocking results
The overclocker seems to have used a reference design graphics card, and it may well be that the factory overclocked versions under liquid nitrogen will show even better performance.
Highlighting 3dmarkradeon RX 6800 XTBENChMARKI
Comment similar news
TechPowerUp
1111
UL Benchmarks introduced its next 3DMARK bikemark, which is .
Once the new test is released, everyone will be able to test and compare their own system’s performance in ray tracing tasks. The launch of the test is scheduled for the day the AMD Radeon RX 6000 graphics cards are released, November 18th.
Scene from 3DMark DirectX Raytracing test
Enabling ray tracing has a negative effect on the graphics performance of the system, which is why UL Benchmarks created a special ray tracing test.
«The 3DMark DirectX Raytracing test is designed to make tracing performance the limiting factor,» , says the blog post. «Instead of relying on traditional rendering, the entire scene is traced and rendered in one pass. In this test, camera rays are traced across the field of view with small random shifts to simulate the effect of depth of field. The frame rate is determined by the time spent tracing and shading a set number of samples for each pixel, combining the results with previous samples, and displaying the output image on the screen.0236 . Users can change the number of samples in a test when rendering a scene at 2560×1440 pixels.
In addition to the benchmark mode, this test can also work in the entertainment mode, which offers the possibility of free movement around the scene with a change in focus.
3DMark DirectX Raytracing test scene
The 3DMark DirectX Raytracing benchmark will be available as a free upgrade to 3DMark Advanced Edition on Steam and the UL Benchmarks website.
testing3DMarkray tracingbenchmarks
Comment Related news
Hexus
Given the performance of the EVGA RTX 30xx KINGPIN graphics card series, it’s no surprise that the older model set a new performance record in 3DMark Port Royal.
According to the published tweets, the GPU voltage was 1069 mV when the record was set. Vince Lucido, a well-known overclocker, got to 2.58 GHz, while the memory was running at 21.5 Gb / s.
Overclocking EVGA RTX 3090 KINGPIN
Using a huge amount of liquid nitrogen for cooling, the overclocker was able to set a record and score 16,673 points in 3DMark Port Royal.
Surprisingly, when the record was set, the GPU frequency was increased by 50%, while the memory frequency increased by only 10%. At the same time, the benchmark result was 30% higher than the standard RTX 3090 card.
More information can be found in the 3DMark database.
overclocking 3DMark video cards EVGAGeForce RTX 3090benchmarks
comment on related news
Fudzilla
A new NVIDIA graphics card appeared in the 3DMark Time Spy test.
Most likely, we are talking about the new top-end video card of the Ampere family, because it demonstrates a 30% performance increase in DirectX 12 compared to the NVIDIA GeForce RTX 2080 Ti Founders Edition.
3D Mark Time Spy
As usual for graphics cards that haven’t been announced yet, the specs shown in the test are a little odd. So, the GPU frequency is 1935 MHz, and memory — 6 GHz. Of course, this is a strange frequency. Reviewers note that it is caused by the lack of «calibration» of the driver, and in the final version everything should be correct.
Testing an unknown Ampere video card in Time Spy
The specifications of three video cards based on the GA102 processor were also published. These models are:
- 2nd Gen TITAN RTX: GA102-400-A1, 5376 CUDA, 24GB GDDR6X, 17Gbps, 384bit, 816GB/s.
- RTX 3090: GA102-300-A1, 5248 CUDA, 12GB GDDR6X, 21Gbps, 384bit, 1008GB/s.
- RTX 3080: GA102-200-Kx-A1, 4352 CUDA, 10GB GDDR6X, 19Gbps, 320bit, 760GB/s.
3DMarkAmpereNVIDIAbenchmarks
Comment on related news
Hexus
The end of support for Windows 7 will mark the end of an era. The operating system will be unsupported starting January 14, and support for a number of utilities from UL will end on the same day.
Thus, UL Benchmarks announced that 3DMark 11, PCMark 7, Powermark, 3DMark Cloud Gate and 3DMark Ice Storm benchmarks will no longer be supported starting January 14, 2020. These tests are no longer up to date.
3DMark 11
All benchmarks have a life cycle. The ones listed above were issued between 2011 and 2013. During this time, a lot has changed. CPUs are focusing on multithreading, while GPUs are getting more and more interested in new technologies like ray tracing, making old benchmarks irrelevant.
Thus, starting January 14, UL Benchmarks stops selling old versions of tests and ends their support program for users. However, enthusiasts will be able to continue using these tests completely free of charge, but their operation is not guaranteed.
Proposed Replacements for Legacy UL
Benchmarks The above is a list of UL Benchmarks products that will be closed, as well as proposed replacements for modern solutions.
testing PCMarkUL3DMarkbenchmarks
comment on related news
The new test has become more flexible. It allows you to select different shading levels when rendering, while Tier 1 can be set to a single shading level per render request. Thus, lower shading levels can be applied to areas where the level of detail is low or negligible, such as in shadow areas.
3DMark
Below are two images that demonstrate how the technology works. As you can see, there is almost no difference between the pictures with VRS on and off.
The updated test also offers an interactive mode. In it, you can make changes to the shading level on the fly, and see how this affects performance.
Frame before applying VRS
In addition to VRS, the new version also offers a fix for hardware data monitoring synchronization problems, as well as a solution to the annoying problem of the test crashing before displaying the results.
Frame after applying VRS
VRS Tier 2 technology is only supported by NVIDIA Turing graphics cards, while Tier 1 can also be tested on the integrated graphics of Intel Ice Lake processors.
Benchmark update is available free of charge for 3DMark Advanced Edition and Professional Edition. VRS technology works on Windows 10 version 1903 and later.
testingVariable-Rate Shading3DMarkbenchmarksutilities
comment on related news
Neowin
UL Benchmarks announces a new computer performance test that will be included in the 3DMark suite. This test is designed to measure the effectiveness of Variable-Rate Shading (VRS).
Variable shading is a DirectX 12 feature that allows developers to improve performance by selectively reducing the level of detail in different parts of the frame where such degradations would be invisible or barely noticeable.
3DMark VRS test. VRS disabled
The new 3DMark VRS test is designed to evaluate performance and image quality when using VRS. The test will be a forest scene with floating lanterns. On the first run, VRS will be disabled for basic performance information.
3DMark VRS test. VRS enabled
With VRS, one pixel shading operation can be applied to a block, such as block shading 4×4 in one operation, instead of 16 operations, which will save resources.
The 3DMark VRS test will apply three degrees of variable shading: 4×4 for distant objects (blue zone in the image), 2×2 for medium distance (green zone) and 1×1 for nearby objects (red zones).
3DMark VRS test. VRS Application Map
The test reports the average frame rate per pass and calculates performance with VRS enabled.
The new 3DMark Variable-Rate Shading benchmark will be available from August 26th.
testing3DMarkDirectX 12benchmarks
comment related news
So, a test of an unknown video card appeared in the 3DMark Time Spy database, but given the Boost frequency of 1878 MHz, we should be talking about the AMD Radeon RX 5700 XT. The benchmark was first discovered by the WCCFTech website. The journalist noted that the tested PC is similar to their test system. The installed i7-8700K processor ran at 5 GHz on all cores, and the memory was clocked at 3200 MHz.
AMD Radeon RX 5700 XT
This configuration achieved 8575 points, of which 8719 points for the GPU test, and 7843 points for the CPU. This allows us to say that the tested graphics accelerator turned out to be faster than the Radeon RX Vega 64 and NVIDIA GTX 1080. Also, the video card turned out to be slightly slower than the RTX 2070 Founders Edition, which gives out 8901 points. This means that without AMD dumping pricing, it will be very difficult to sell the Radeon RX 5700 XT.
testingrumors3DMarkAMDRadeon RX 5700 XTbenchmark
1 comment related news
Overclock 3D
UL Benchmarks announced the development of a PCI-Express 4.0 test for its 3DMark benchmark, which should be available this summer.
The 3DMark PCI-Express benchmark, as the name suggests, tests the PCI-Express (PCIe) bus, which allows devices within a computer to communicate at high speeds. This version 4.0 bus has twice the bandwidth of 3.0. This allows games to transfer more data, reduce loading times, and enable more complex scenes to be displayed.
Screenshot of PCIe
benchmark In fact, the 3DMark PCI-Express test measures the speed of data transfer to the video card. This will allow you to compare throughput across different generations of PCIe, since this test will be extremely easy to use. In addition, this test will probably allow you to find out how demanding the video card is in terms of bandwidth, and whether it can be used on older generations of PCIe without performance loss.
PCI-Express speed measurement will be available in 3DMark this summer. For all 3DMark Advanced Edition and 3DMark Professional Edition license holders, the upgrade will be free.
testing3DMarkPCI Expressbenchmarks
comment on related news
Dark Vision Hardware
Despite the fact that the release of the Radeon VII video card will take place on February 7, the first tests of its performance have already appeared on the Web.
Several bloggers, including Tum_apisak and Komachi, reported that the Radeon VII graphics card scored 19,210 in 3DMark Fire Strike, 12,224 in Fire Strike Extreme, and 6707 in Fire Strike Ultra. These are data from the 3DMark database.
Radeon VII
video card Twitter also reports that the card scored 8700 points in the Time Spy benchmark. This means that the Radeon VII outperforms the RTX 2080 without any problems. NVIDIA solutions. So, the new AMD card turned out to be slower than the GTX 1080 Ti and barely beat the RTX 2070 at 2560×1440 at standard quality settings. At high quality and the same resolution, it doesn’t even come close to the GTX 1070 Ti.
Do not forget that these are only the first test results, and those are unofficial. This may mean both that the results in Final Fantasy 15 are not real, or that AMD still has a week to finalize the driver.
testingrumorsFinal Fantasy3DMarkvideo cardsAMDRadeon VIIbenchmarks
3 comments related news So far, it is not possible to get it due to problems with suppliers, but some enthusiasts still got it, despite the price of $ 2,500.
Naturally, after the appearance of the video card, the first benchmarks appeared. Of course, enthusiasts immediately overclocked their boards, so we have not pure performance.
NVIDIA Titan RTX
graphics card However, the graphics card has a standard hardware configuration that includes a TU102 GPU consisting of 4608 CUDA cores, 288 texture units, 96 ROPs, 576 tensor cores, 72 RT cores. The board also houses 24GB of GDDR6 memory at 672GB/s.
Overclocked to 2070 MHz GPU and 2025 MHz memory, the card scored 41,109 in 3DMark.
Titan RTX 3DMark Fire Strike test results
Amazing result. It is worth the fate that this is one of the first tests of a video card, which means that in the future professional overclockers will be able to achieve much better performance.
testing3DMarkvideo cardsNVIDIAGeForce TITAN RTXbenchmarks
comment on similar news
Eteknix
UL Benchmarks, a week after the announcement, released a 3DMark Port Royal benchmark that measures system performance under a real-time ray tracing load.
The 3DMark Port Royal benchmark is the world’s first ray-traced system speed benchmark designed for gamers. The application is used to test the performance of any graphics card that supports Microsoft DirectX Raytracing.
3DMark Port Royal
This technology provides a new level of visualization in games, raising the realism of images to a level never seen before. In addition to the speed of work, the new benchmark also demonstrates the capabilities of the new technology.
3DMark Port Royal — Real-time ray tracing benchmark — Full Demo 1440p
They promise to release the test on January 8th.
testing3DMarkray tracingbenchmarks
comment on similar news
Dark Vision Hardware
Following rumors that UL was preparing an update to the Time Spy test with ray tracing support, the company contacted global observers to clarify the situation.
As it turns out, the developer will not update the Time Spy test suite with ray tracing technologies. The fact is that this will require a complete rewrite of the benchmark, which will be counterproductive and lead to a change in the test results in the normal mode, without Raytracing.
UL test with ray tracing
As a result, UL decided to develop a completely new benchmark built from the ground up using Microsoft DirectX Raytracing (DXR). This new test will be added to the 3D Mark package as updates. The new test will produce its own results, just like Fire Strike and Time Spy do, and will be another measure of record-breaking performance.
No other details yet, which is understandable. The new test should appear around September 20, along with NVIDIA 20-series graphics cards.
testingDXR3DMarkray tracingbenchmarks
comment on related news
TechPowerUp
UL, which bought 3D Mark, announced the development of a change to the Time Spy test.
The updated test, scheduled for the end of September, will receive support for ray tracing in the high performance preset.
Time Spy ray tracing test
The new version of the benchmark will be tentatively released with the next update of Windows 10, currently called Redstone 5. This means that the release will occur after the launch of the 2000 series of NVIDIA RTX, which will take place on September 20th.
Time Spy ray tracing test
One can only hope that a preview of the test will be available to reviewers so that we can finally answer the question of whether ray tracing is really critical to performance and how effective the new tensor cores are for ray tracing. and how many “gigarays per second” are actually needed.
testingrumors3DMarkray tracingNVIDIA GeForce RTX 2080benchmarks
comment on similar news
TechPowerUp
The last few days, all attention has been riveted on the new NVIDIA video cards, but what they will be capable of in terms of speed has remained a mystery. Now the company has clarified the situation by presenting the results of testing in games.
During the Editors’ Day at Gamescom, the company presented performance graphs for various games in 4K resolution, but without the RTX effect. According to reports, the effect of ray tracing drastically reduces the speed of work. Users noted that in Tomb Raider with ray tracing, the RTX 2080 Ti cannot consistently maintain 60 fps even at FullHD resolution.
Performance of the GeForce RTX 2080 in games
But in standard settings, the new video cards show a good increase of 30-50% compared to the GTX 1080. But when using the new Deep Learning Anti-Aliasing anti-aliasing technology, the increase exceeds 200%. This anti-aliasing method uses artificial intelligence, it is not as accurate as SSAA, but it shows excellent results in terms of speed.
RTX 2080 vs. GTX 1080 performance comparison
Unfortunately, we don’t know what quality settings were set during testing, and given the marketing, these results should be viewed with some skepticism.
testing gamesNVIDIAbenchmarks
comment on related news
These are the first test results of the new Intel CPU for the LGA1151 socket, and they are impressive. The chip contains 8 cores and 16 threads. It was tested by the famous Thai overclocker TUM APISAK.
Core i9 benchmark results-9900K in 3DMark
In the CPU test, the Core i9-9900K scored 10719 points with a total score of 9862 points on the GeForce GTX 1080 Ti graphics card. This score is 2500 points higher than the 6-core 12-thread Core i7-8700K and about 1500 points higher than the 8-core 16-thread AMD Ryzen 7 2700X.
CPU specification in test
In testing, the processor had all 8 cores with 16 threads, a base clock speed of 3.10 GHz and a boost frequency of 5.00 GHz.
testing3DMarkCore i9CPUIntelbenchmarks
1 Comment Related news
TechPowerUp
PCGamesN shared interesting benchmark results for NVIDIA and AMD graphics cards in the as-yet unreleased game Battlefield V.
is being developed in collaboration with NVIDIA Battlefield V02 , but it seems that the cooperation did not bear fruit. The fact is that according to the published report, the AMD RX 580 8 GB video card (the only one from the red ones) outperformed the NVIDIA GTX 1060 6GB by a noticeable amount.
Battlefield V
There was a 30% speed difference between 1080p and 1440p on Ultra settings. Also, when switching to DX12 rendering, AMD traditionally turned out to be faster. Perhaps the old collaboration between DICE and AMD is taking its toll, but while the game is in pre-alpha testing, it may still change.
Test GTX 1060 and RX580 in Battlefield V at 1080pTest GTX 1060 and RX580 in Battlefield V at 1440p
Battlefield V scheduled to release on 1
The 2400G and 2200G processors use Vega 11 and Vega 8 graphics, respectively. According to the results of 3DMark 11, the older solution turned out to be comparable to the discrete card NVIDIA GT 1030 and RX 550. Thus, the Ryzen 5 2400G processor scored 5042 points in the benchmark, while the RX 550 Mobile demonstrates 5568 points.
As for the Ryzen 3 2200G, it scored 3950 points. But the RX Vega M GL, which AMD created for Intel, is incomparably faster. With 9967 points in 3DMark 11, this chip even outperforms the GTX 1050Ti with 9694 points. This means that the graphic part in laptops based on Intel hybrids will be almost twice as productive as in AMD solutions. If, of course, these rumors are true.
Testing 3DMARKAPURX VEGARYZEN 2000AMDBENCHMARKS
Comment similar news
Kit GURU
Brazilian overclock Ronaldo Buassali revealed a method for caring for rolids of benchmark results, which allows you to overtake the test results.
It is noteworthy that the trick he came up with is not unique to 3DMark, and can be used for many other benchmarks. The essence of cheating the testing program is to change the driver settings right before starting the test. Thus, by changing the settings after driver validation, you can get results that are unattainable in any other way.
To demonstrate his method, Ronaldo scored 12638 on the graphics test and 13278 total on Time Spy. By changing the settings, he got 12951/13800 points respectively.
He explained that the trick does not always increase the result. Its effectiveness depends on the software, but the results in benchmarks from Futuremark, Unigine and Allbenchmark (Catzilla) turned out to be questionable.
Olha isso (Look at)! Melhore seu score (Improve your score) — LOD
As for Futuremark, the developers acknowledged the possibility of bypassing the systems for validating results, and closed their Hall of Fame until the problem was resolved. A fix is promised soon.
overclocking 3DMarkUniginebenchmarks
comment on similar news
Dark Vision Hardware
A few months ago, the first benchmark of the Radeon RX Vega video card appeared on the Web. However, now that the card has gone through a series of optimizations, the result of new testing has appeared, which turned out to be noticeably higher.
In 3DMark 11, the Vega RX performed slightly better than the stock GeForce GTX 1080. In addition, AMD still has time to fine-tune the driver, as it should be announced at Siggraph 2017 on July 30th.
RX Vega graphics card was tested at 1630MHz GPU and 8GB HBM2 VRAM at 945MHz. As a result, the Radeon RX Vega (device ID 687F:C1) scored 3DMark 11 P21179 points in 3DMark 11, and the net graphics component is 31873 points, which is higher than the stock GeForce GTX 1080. However, it is slower than the overclocked version of the same card.
Traditionally, AMD wins by lowering the price compared to its counterparts from NVIDIA, however, given the high cost of HBM2, it is difficult to say whether the company will be able to repeat its traditional strategy. 998 (at that time the company was still called MadOnion), and since then the benchmarks of this company have become virtually the only unshakable standard for measuring the performance of video cards. It has always been a challenge to produce tests that all manufacturers of the corresponding hardware like. There were also scandals related to specific optimizations in drivers specifically for 3DMark, and criticism from users and manufacturers never ceased. But although this affected the reputation of the test, there were still no alternatives, and year by year 3DMark is only gaining popularity.
Why are these tests so common? In addition to nice graphics and the use of modern technologies, one of the basic requirements for a 3D benchmark is ease of testing and good repeatability of results. The package as a whole satisfies these requirements. It is easier for the user to run 3DMark than to test several games and get an average score. A good 3D test should match the load and technologies of modern games and projects of the near future as much as possible. At the same time, released games are often too simple for modern GPUs, especially given the widespread use of multi-platform projects.
Futuremark develops its tests taking into account the opinions of companies such as Nvidia and AMD, listening to everyone little by little with varying success, so their benchmarks are relatively honest, without a clear bias in anyone’s direction. Since 3DMark is widely used by users and manufacturers, the results database contains numbers from millions of different systems. This feature makes this test package very convenient for comparing performance against other systems, including overclocking, and is often used by overclockers. Yes, and the specialized press considers 3DMark one of the mandatory tests for their materials, even taking into account the presence of gaming applications.
The company updates the 3D test suite from time to time to meet changing requirements. They were going to release the update back in December last year, but delayed its release first until January, and then until February. And now, after a two-year break (the previous version was eloquently called 3DMark 11), Futuremark has released a new version of its 3DMark graphics benchmark suite. This is exactly what the ninth benchmark of the series is called — without any indication of the year or version number, for the first time in many years. The name was shortened to 3DMark because this test suite is the same for many hardware and software platforms and will be updated with more and more new subtests.
Perhaps the new product is the most versatile 3D test among all the company’s packages. Following the trends of the times, Futuremark turned not only to desktop PCs, but also to the most widespread mobile devices: smartphones, tablets, etc. Previously, the company made separate tests for desktop and mobile PCs based on x86 processors and separate benchmarks for mobile solutions , but the released 3DMark has become a cross-platform benchmark for devices running various operating systems: Windows, Windows RT, Android, and even iOS (Apple iPhone and iPad). Even though this cross-platform approach mostly belongs to only one of the 3DMark tests, even such an opportunity to compare very different devices in terms of 3D rendering speed is already very pleasing. The rest of the tests require more sophisticated hardware and will only be available on systems running Windows for now. It is this version that is the most complete and contains tests that use the capabilities of DirectX 11, DirectX 10 and DirectX 9. So far, only the desktop Windows version has been released, all others will be released later.
Released for Windows, the package includes three tests, each of which is designed for a specific range of devices up to heavy-duty gaming PCs and systems for overclockers. When testing various devices, it is necessary to choose the right subtest based on the power and hardware capabilities of the systems being compared. Each of the tests at default settings gives the final result in the form of a score, by which the system can be compared with similar ones. But, unlike previous versions, there is no general result for the entire 3DMark package now, only the results of specific tests remain. It is planned that over time the number of tests will increase — new ones will simply be added.
The test scenes in the new 3DMark are written specifically for this benchmark, using tessellation, volumetric lighting, advanced post-processing, and GPU physics effects. The 3DMark package uses the Bullet engine as a physical engine, both universal CPU cores and specialized GPU cores are used for calculations. In this article, we will try to figure out how worthy the continuation of the Futuremark test package series turned out to be, and first we will look at its interface.
Versions, settings, and interface
Minimum requirements to run the package:
- DirectX 11 compatible graphics card (support for all Direct3D 11 hardware features is optional)
- Windows Vista, Windows 7 or Windows 8 operating system
- 1.8 GHz or faster
- 2 GB of system memory (4 GB recommended)
- Approximately 3 GB of free storage space
- Any sound card
Quite logically, the minimum requirements to run the Windows version of 3DMark include an operating system starting with Windows Vista (with DirectX 11 platform update installed), Windows 7 and Windows 8 are also supported. The requirements for 3D tests on the processor are not high by today’s standards — any dual-core AMD or Intel CPU with a frequency of 1. 8 GHz is sufficient. The minimum amount of RAM is 2 GB, but 4 GB is recommended (at the current price of RAM modules, this recommendation is also quite justified), and about 3 GB of disk memory is required.
The most important thing here is the support of DirectX hardware capabilities by the video card. The minimum GPU requirement includes Direct3D9 compatibility, but running all benchmarks will require a Direct3D11 capable GPU and at least 1 GB of VRAM. The minimum level of support for DirectX 9 includes support for Shader Model 3.0 shaders, a minimum of 128 MB of video memory, and WDDM 1.1 drivers. The last requirement negates, for example, video cards of the Radeon X1000 series, for which WDDM 1.1 drivers simply do not exist. Thus, the minimum graphics cards for 3DMark are the AMD Radeon HD 2000 series (two tests are running: Ice Storm and Cloud Gate), the Nvidia Geforce 7000 series (Ice Storm only), and the integrated Intel GMA X4500 graphics core (Ice Storm only).
Depending on the 3DMark subtest, the requirements for video memory are also different. So, Ice Storm requires 128 MB of local video memory, the Cloud Gate test already requires 256 MB, Fire Strike uses 1 GB of video memory in normal mode, and 1.5 GB of memory and 4 GB of RAM in Extreme mode. And although on systems with video cards with less video memory, the tests may run, but they will show too low performance.
As far as mobile versions are concerned, we’re promised that 3DMark will run on all Windows RT devices, but most of them won’t be able to run all tests yet. On Android devices, the minimum required OS version is Android 3.1, the amount of memory is from 1 GB, the amount of storage is 300 MB. The video core must have support for OpenGL ES 2.0, and the requirements for the processor have not yet been announced — they promise to provide details a little later. Apple devices running iOS require at least the fifth version of this system and the same free 300 MB of storage. The «Apple» version supports devices ranging from the iPhone 4, iPad 2 and the 5th generation iPod Touch. As a result, all tests (Ice Storm, Cloud Gate and Fire Strike) can run on Windows and Windows RT, while only Ice Storm can run on Android and iOS.
3DMark Versions and User Interface
As always, 3DMark comes in three versions: the free Basic Edition , Advanced Edition (for $24.99), and Professional Edition (for $995). Even in the free Basic Edition , all tests are available, but only with standard settings — without the possibility of changing them. Also, there is no support for Extreme mode for the Fire Strike test. That is, in this version, it is possible to run three tests in the default settings, and the results are shown on a special HTML page only when the Internet connection is active.
Advanced — version adds an Extreme mode for the Fire Strike test, and also allows you to change all fine test and resolution settings, as well as run subtests separately. Of course, if you change the settings, it will not be possible to get the final result in standard points, but this may be useful for specific studies. It is also possible to «loop» the work of the benchmark, which is useful for stress testing, but the most important addition to this version can be considered the saving of test results and their output in 3DMark itself. By the way, the new mode for displaying results in the form of graphs is very convenient, and what prevented it from being done earlier is completely incomprehensible. Interestingly, Advanced has risen in price since 3DMark 11: then it cost $19.95 and now $24.99. Inflation!
As for the Professional version for wild money (many companies buy them for their own purposes — this is what Futuremark lives on), it is intended for all types of commercial use and offers all the features of the Advanced Edition, and in addition gives the quality comparison functionality images (rendering individual frames), allows you to run tests using the command line and save the results in XML format — which is very convenient for testers.
Advanced Edition will also be available for free with the purchase of certain graphics cards from Futuremark sponsors MSI and Galaxy, as can be seen from the logos that appear in the demos. You can also save on the purchase of the paid version if you purchase the Steam version of the regular version or by entering the serial key from 3DMark 11 Advanced Edition during the purchase process — in these cases, the new test suite will cost 25% less. In addition, the Steam version of the benchmark has a more convenient way to install, automatically update and publish results. By the way, the keys to the Advanced Edition are suitable for both the Steam version and the regular one — they are stored in the Windows registry.
The appearance of the 3DMark window and the overall interface of the program has changed somewhat, although the style remains the same — a flash-like interface made according to modern standards (it will probably be different in mobile versions). As before, the Custom and Professional pages, which are not available to free users, are highlighted separately. On the first page there are only welcome words and the only button is to run all tests:
On the second page you can read detailed information about each of the tests and even about its subtests (graphic, physical and combined), as well as select the display of demo videos and a profile Extreme for Fire Strike test:
The third page is the most interesting: it contains user settings, separate for each test. Note that they are located on the Custom page for the Windows version and on the Settings screen for Android and iOS. We also repeat that for the Windows version of the package, they are available only for paid versions of the package — Advanced and Professional editions. Here you can change numerous test-specific settings to suit your needs. Naturally, the final result with the changed settings is displayed only in the form of FPS, and the standard 3DMark score is not displayed in this case.
For example, in the Fire Strike settings, you can select the method and quality of texture filtering, the type of full-screen anti-aliasing and the number of multisampling samples, the split ratio of primitives for objects to which tessellation is applied, the maximum possible split ratio, resolution and number of samples from shadow maps, quality volumetric lighting (sampling step in the «ray marching» algorithm when calculating volumetric lighting is an important parameter, large values of which reduce performance), the number of samples per pixel when calculating indirect lighting in the ambient occlusion algorithm, as well as the quality of one of the types of post-processing — imitation of the depth effect sharpness depth of field. This setting controls the size of the bokeh texture used in post-processing for objects in the scene, while the Color saturation setting changes only the color saturation of the final image without affecting performance.
The fourth page is only available for the expensive «professional» edition. It allows you to draw only single frames (or a sequence of them) to check the quality of rendering, including a variant of the Microsoft reference rasterizer (requires installation of the DirectX SDK), and also provides export of saved results to XML format.
The fifth page is used to display information about the test results, and it is made very spectacular and convenient. The most interesting feature of the new 3DMark was the output of results in the form of graphs. In the new package, in addition to individual values of the average frame rate in the subtests and the final results in 3DMark points, a small graph of FPS (and other parameters) is also shown:
Diagrams and other system information can be viewed in more detail by clicking the Details button in the corner. This mode is thought out in detail and quite convenient. In the screenshot below you see the result of a PC with an old Intel Core i7-990X processor and an Nvidia Geforce GTX 480 video card. As you can see, in addition to the already familiar score in 3DMark points, it also provides detailed frames per second graphs, divided into subtests: two tests graphics, physics test and combined.
It is also possible to display other parameters: frequencies, CPU and GPU temperatures, etc. Apparently, this implies further improvement of the benchmark, since these parameters were not shown on several different systems. In real use, the page is even more convenient than on the screenshot: when you move the mouse cursor over the chart, additional information is displayed (for example, instant FPS at this point), you can also zoom in or out on one or another section of the chart. The only downside to this new feature is that it’s only available to users of the pro version, although it would be quite useful for overclocking enthusiasts as well. After all, they buy the test for half the western price of a full-fledged game ($25) — it would be possible to add graphics to the Advanced Edition…
Well, the sixth page contains everything that was not included in the previous ones. It shows information about the 3DMark edition, software version, there are settings related to registration, and some system parameters:
Register section — if you are a happy owner of a license (bought or received it along with a video card, for example) for 3DMark Advanced or Professional Edition , then it is on this page that you can register the package. You can also, if necessary, unregister the application — in order to transfer the license to another system, for example.
Automatically view results online — automatic viewing of test results in the browser. The option is selected by default and cannot be removed in the free version of 3DMark Basic Edition. Automatically hide results online — the setting allows you to prohibit sending results to the 3DMark website — these results will not be visible to other users and will not be placed in the Futuremark Overclocking Hall of Fame list. The setting is disabled by default and cannot be enabled in the free version.
Scan SystemInfo — setting includes collection of data about the system by a component of the SystemInfo package — the option is enabled by default and is required to obtain a valid test result. SystemInfo hardware monitoring — includes the collection of information such as CPU temperature, clock speed, etc. This setting is also enabled by default.
Demo audio — the setting allows you to turn off the sound when watching demos. GPU count — Forcing the number of GPUs that the package should identify. By default, the number of GPUs is determined automatically, and this determination does not work only in extremely rare cases. Scaling mode — selection of the image scaling mode on the screen. By default, the Letterbox mode is selected, which does not change the aspect ratio, adding black bars at the top and bottom to the rendering resolution. Selecting the Stretched mode will enable stretching the rendered frame to fill the screen, changing the aspect ratio. This only affects the output of the image and does not affect performance. Language — select the interface language: English, German, Traditional Chinese and Simplified Chinese.
In general, although the 3DMark interface and the set of settings have clearly become better, they are far from ideal — for example, even in the professional version there is no easy way to run only tests without running demo videos (Ice Storm, Cloud Gate or Fire Strike can be run without them separately, but not all together, in batches). However, when using the command line utility, you can do anything at all, but this does not apply to the graphical interface.
Graphic tests
So, modern 3DMark consists of three different tests: Ice Storm, Cloud Gate and Fire Strike. Ice Storm is a graphics benchmark that utilizes DirectX 9-level hardware capabilities and the equivalent capabilities of Android, Windows RT, and iOS-based solutions. Cloud Gate already belongs to the DirectX 10 hardware support level and will only run on Windows and Windows RT — it is designed for testing laptops and low-cost home PCs. Fire Strike is the most complex test of the package and uses almost all the main hardware capabilities of the most modern video chips that support DirectX 11 — it is designed for powerful gaming PCs. This test has an additional «extreme» mode «Extreme», introduced for enthusiasts with super-powerful overclocked systems based on two or more GPUs, and has a certain margin of safety for the near future.
All tests in 3DMark use the same DX11 engine (so the package does not work on Windows XP), but with different capabilities: Ice Storm is limited to Direct3D feature level 9_1, Cloud Gate uses Direct3D feature level 10_0, and Fire Strike is capable of show all the features of DirectX 11. The applied effects and techniques are described in detail below, in the corresponding section.
As far as DirectX 11.1 add-ons are concerned, their use in 3DMark is minimal, and features introduced in DirectX 11.1 are used solely to accelerate rendering techniques using DirectX 11.0. There is no visual difference between the picture obtained on DX11 and DX11. 1 video chips in 3DMark, and there is a practical difference in performance only in the Ice Storm test on Windows and Windows RT systems with low-power GPUs. For example, one DX11.1 optimization slightly improves performance on a very limited number of tablets with graphics chips that use tile rendering, or laptops with two GPUs. The 16-bit texture formats supported in DX11.1 are used in Ice Storm tests to store intermediate results in post-processing. If there is no such support, then 32-bit formats are used.
Ice Storm
This is the only fully cross-platform package test that runs on all devices. It’s great for comparing the 3D performance of smartphones, tablets, ultraportable laptops, and older or weaker PCs. Ice Storm includes two graphics subtests that measure the performance of graphics cores (GPUs) and one physical test that uses the computing capabilities of a general purpose processor (CPU).
On Windows systems, the Ice Storm test uses the DirectX 11 engine, which is limited to Direct3D 9 (feature level 9), while on Android and iOS, the OpenGL ES 2. 0 engine is used, so this is a true multi-platform test. Since all resources (textures, models, etc.), resolution and settings in the Ice Storm test are the same on all platforms, relatively equal testing conditions allow us to compare various devices running the specified operating systems.
Ice Storm’s default settings include a render resolution of 1280×720 (720p) and 128MB of video memory. The image in Ice Storm is calculated at a fixed resolution and then scaled to the «native» resolution of the output device. This technique is called off-screen rendering and is often used to compare systems with different display capabilities.
Graphics test 1
The first graphics test from Ice Storm is designed to test vertex processing speed (geometric performance), while the load on pixel-by-pixel processing is relatively small here. This test determines the processing capabilities of geometry and pixels. In an average frame, 530 thousand vertices and 180 thousand triangles are processed (including geometry for calculating shadow maps) The number of pixels processed in a frame is approximately 4. 7 million. particles.
Graphics test 2
The second Ice Storm graphics test determines the GPU’s ability to handle large numbers of pixels. It makes heavy use of texture fetches, per-pixel calculations, and numerous writes to off-screen buffers. On average, 12.6 million pixels are processed per frame, and there are only about 75 thousand vertices in the frame in this test. The amount of geometry is much lower than in the first test, since the models are less complex and the shadows are not drawn. And additional pixel load is provided by the rendering of particle systems and post-processing effects: bloom, streaks and motion blur.
Physics test
The physics subtest of this section of 3DMark tests the system’s CPU capabilities for physics calculations. The load on the video core here is minimal, but all its abilities are required from the CPU — the test uses the Bullet Open Source Physics Library. It simulates four «worlds», each containing two rigid and two soft bodies colliding with each other. One thread per CPU core is used to run the simulation. All physical calculations are done on the CPU, and the graphics load is very low — the backdrop is pre-rendered and is a static image.
Ice Storm on all platforms uses the same engine with support for the same functionality: traditional rendering with one pass per light, lighting model — Blinn Phong, several types of lights (point light without shadow calculation, directional light with shadows and a pre-calculated environment cubemap), scene updating and object visibility calculation are parallelized, draw calls use a single thread, static and skinned geometry is supported, translucent objects and particles are supported, 16-bit formats are used in light buffers if hardware support is available .
In the Windows and Windows RT versions of the package, the Ice Storm test requires support for Direct3D feature level 9_3 or 9_1 with support for hardware shadow filtering. Android versions do not use specific OpenGL ES 2. 0 extensions, textures are compressed in ETC format, and textures with an alpha channel are uncompressed. Naturally, an absolutely equal comparison of systems when working under DirectX and OpenGL ES will not work. For example, if DirectX has built-in support for working with PCF-filtered shadow maps, then in OpenGL this is solved by extensions from hardware manufacturers that … are not used in 3DMark. Therefore, versions for OpenGL ES emulate working with shadow maps programmatically and lose some performance.
It is also interesting that this test takes into account the spread of tiled GPUs in single-chip systems (SoC), and corresponding optimizations are made in its «mobile» versions to help such solutions show higher performance. On the other hand, even in the simple Ice Storm test, post-processing filters and filtered shadow maps are used, the rendering of which is unprofitable for tiling GPUs due to the need to switch buffers (render target). In general, it looks like another attempt to please both yours and ours, which is in the spirit of Futuremark.
Cloud Gate
The second 3DMark test exists (so far) only in the Windows version and is intended for laptops and mid-range home PCs with an integrated video graphics subsystem. The Cloud Gate test contains two graphic subtests and one physical subtest. The test uses the DirectX 11 engine, which is limited by Direct3D 10 capabilities. Accordingly, Cloud Gate is best suited for testing modern entry-level systems and legacy DX10-enabled GPUs.
Cloud Gate in the new 3DMark roughly corresponds to the level of one of the previous versions of the package — 3DMark Vantage. The difference of the new engine is that it uses DirectX 11, but is limited to Direct3D feature level 10, unlike the DirectX 10 engine in the old version. This approach to game engines will be the most common — you can support a single version of DirectX, but support previous generations of hardware — up to the DirectX 9 level. Specifically, Cloud Gate is designed for the DirectX 10 level.0009
The default test settings are as follows:
The resolution is the same — 1280×720, but the memory is already required twice as much. You can increase the settings, but the final score in this test will not be shown in this case.
Graphics test 1
The first graphics test from Cloud Gate loads geometry processing blocks using simple shaders. On average, 3 million vertices and 450 thousand geometric primitives are processed per frame, and as a result of the geometry shader, 1.1 million triangles are obtained, including shadow maps. Volumetric lighting is disabled in this test, but a number of effects with particle systems are present. Post-processing is also used in the form of bloom and depth of field effects. On average, about 18 million pixels are processed in a frame.
Graphics test 2
Cloud Gate’s second graphics test uses complex math shaders but significantly less geometry. On average, 1.8 million vertices, 340 thousand primitives and 690 thousand rasterized triangles are processed in a frame. Simple volumetric lighting is used without particle effects, and post-processing effects are similar to those used in the first graphics test.
Physics test
The physics test in Cloud Gate also relies on the capabilities of the CPU, not the GPU: the load on the graphics core is minimal, and the CPU works at full power. The test simulates 32 worlds, each containing four soft bodies. 4 hinges and 20 rigid bodies collide with each other. Rigid bodies are invisible, but affect soft bodies. Simulations are run in the amount of one thread per available CPU core, and the capabilities of the Bullet Open Source Physics Library are also used. All physical effects are calculated on the CPU, the load on the GPU in the subtest is minimal. Each world also contains one particle system, which is simulated on the CPU.
The engine in Cloud Gate is the same as in Fire Strike (see below), but uses a limited set of features, such as a simplified lighting model, as well as some features made with Direct3D 10 rather than D3D11. Accordingly, the Windows Cloud Gate test (full desktop and Windows RT) requires Direct3D feature level 10 hardware support. It differs from the tests in the 3DMark 11 suite in that it is even more demanding on GPU power. This test uses the multi-threaded DirectX 11 engine and includes not only two graphics and one physical subtests, but also a combined one — in which the load falls equally on the CPU and GPU. The paid editions of Advanced and Professional additionally include an Extreme profile for this test, which loads the video core even more and is intended for systems with multiple GPUs and future solutions. Interestingly, Fire Strike is available in versions of the package not only for Windows, but also for Windows RT — although so far only theoretically, since full DirectX 11 support is required by the graphics core of the operating system. It would be interesting to see a WinRT device capable of somehow running this test…
For high-end desktop PCs, the default settings for the Fire Strike test don’t seem all that maxed out — they’re more like «medium» in 3DMark terms, except that the rendering resolution is high: for top multiprocessor SLI and CrossFire video systems, as well as future GPUs. Extreme settings differ from the standard ones in that rendering is carried out to a higher resolution buffer, and the graphic test settings are raised to the maximum — geometric detail is increased, the quality of shadows, lighting and post-processing is improved. Naturally, the requirements for GPU power have increased significantly — all the juices are being squeezed out of the graphics system, this is the maximum 3D load in all of 3DMark.
Graphics test 1
Graphics test 1 from Fire Strike focuses on lighting and geometry calculations. Particles are rendered at half resolution and dynamic particle lighting is turned off. This subtest uses a scene with 100 spotlights and 140 spotlights without shadows. On average, 3.9 are processed per framemillion vertices containing 500k tessellation patches resulting in 5.1 million triangles per frame, including shadow maps. The compute shaders in the subtest are used to simulate particle systems and post-processing effects (except for the disabled depth of field simulation effect). Pixel loading averages about 80 million processed pixels per frame.
Graphics test 2
The second graphics test from Fire Strike contains the most particle system calculations and other GPU physics simulations. Particles are rendered at full resolution and dynamic particle lighting is enabled. The scene has 2 smoke sources simulated on the GPU, as well as 6 spotlights with shadow calculation and 65 point sources without shadows. On average, 2.6 million vertices are calculated per frame, containing 240 thousand patches for the tesselator and 1.4 million primitives generated by geometry shaders. In total, 5.8 million triangles are processed on average per frame. Computational shaders are used in particle and liquid (smoke) simulations, as well as in post-processing, including depth of field. On average, 170 million pixels are processed in the frame of this subtest.
Physics test
The Fire Strike physics test also uses the Bullet Open Source Physics Library to simulate physics effects on general-purpose CPU cores, and the load on the GPU is deliberately kept to a minimum. The test simulates 32 worlds, uses one thread per CPU core, does all the physics on the CPU, and updates the vertex data on the GPU every frame.
Combined test
The only combined test in 3DMark loads both the GPU and the CPU at the same time. The GPU work combines elements of graphics benchmarks 1 and 2, including tessellation, volumetric lighting, particle and fluid simulation, post-effects such as bloom and depth of field. The load on the CPU is provided by simulating the physics of solid objects of collapsing statues in the background. The scene has 32 worlds in separate threads, each containing a 113-piece statue. Additionally, the scene has 16 invisible solid bodies in each world — to simulate a collision with the statues. The Fire Strike combo test also uses the Bullet Open Source Physics Library.
Demo mode
Each of the above tests consists not only of several subtests, but also of a demo that contains some kind of recorded action shown to the music, serving purely for entertainment purposes. The demo mode is designed to show the capabilities of the Futuremark graphics engine with DirectX 11 support, and all three scenes are shown in it in much more detail, more interesting and longer than in performance tests, and even to the accompaniment of music specially written for 3DMark.
Ice Storm
Cloud Gate
Fire Strike
Turning on the demo mode in itself is a very good thing, but demo videos should not be boring, they need an interesting plot, the presence of a spectacular action and maybe even make some sense. In fact, only the last part of the demo, from Fire Strike, more or less corresponds to this description, and it is really not bad. But the first two parts of the demo are too boring, in our subjective opinion.
Scoring
With each new version of 3DMark, Futuremark revises the selected screen resolution for benchmarking. If earlier in 3DMark 2011 there were three different sets of settings (Entry, Performance and Extreme), which differed in resolution and test settings, now the approach is completely different. An important change compared to all previous versions of 3DMark was the «decoupling» of the rendering resolution from the screen resolution — the second can be anything, but the first is fixed. In other words, the package is made in such a way that it allows you to compare platforms with different screen resolutions on an equal footing, drawing graphics into internal buffers of a certain resolution. So in the Ice Storm test, both a smartphone with a screen resolution of 800×480, and a tablet with a screen resolution of 1280×800, and even a PC with 2560×1440 will work under the same conditions — all the same, rendering will be done to a buffer of the same resolution of 1280×720. And only then this buffer will be scaled to the screen resolution — in modern devices this is done with almost no performance loss.
Why is this necessary and why is it important, apart from setting different devices on an equal footing? Now 3DMark does not require 1920×1080 resolution support from a laptop, and it can be tested in any mode — theoretically, you can even run Fire Strike Extreme even on a netbook with integrated graphics (with DirectX 11 support, of course), if you want to see a slideshow of a couple of dozen frames. In all previous packages, even powerful mobile systems were limited to medium settings precisely because of the resolution of their displays.
Accordingly, the effect of the set resolution on performance also changes. If in 3DMark 11 a change in resolution immediately affected the frame rate and the overall score, then in 3DMark off-screen rendering is used with the display of the image on the screen, which is scaled from the rendering resolution to the display resolution. The internal resolution for tests is always the same, for Ice Storm and Cloud Gate — 1280×720, for Fire Strike — 1920×1080, and for Fire Strike Extreme — 2560×1440.
Modern GPUs scale the image with almost no performance loss, so the desktop resolution has no effect on the result. On the other hand, if the frame rate in the test is limited by the CPU, and it is quite high (as in the Ice Storm test on all modern desktop-level video cards or Cloud Gate on top-end GPUs), then this affects the result — FPS is shown lower than it can be , since it adds the overhead of scaling and timing the output to the display (which would not be the case with full off-screen rendering without displaying the image on the screen). And to achieve overclocking records, it’s better to set the desktop resolution equal to the resolution of the internal off-screen buffer (see above) in order to squeeze a couple more percent out of the result. In the case of the Professional and Advanced editions, you can change the internal rendering resolution, but, of course, the 3DMark score in points is not given.
It should be noted right away that since there is no total number of 3DMark points now, it is necessary to specify the test in which the result was obtained, and in the case of Fire Strike with the Extreme profile, also this mode. The 3DMark package includes three different tests, each made for specific hardware devices, and each of them gives its own score in points, which must be compared with the results of other systems in the same test — results from different tests and profiles cannot be compared.
The total score for each of the three tests of the package is calculated from the results of the graphic, physical and combined (if any) subtests as follows:
The weights (W) for each parameter are taken from the table, they differ for each test:
The contribution of tests to the total score is different, the difference between the weight of the graphic and physical subtests is especially important. In addition, the difference between the physical and graphical accounts can be used to judge the (im)balance of the system. If the graphics score is significantly higher, then you need to tighten up the CPU performance, and vice versa.
The final 3DMark score is the sum of the results of all subtests using special formulas. We will not describe it in detail, but simply give the relevant pages from the Futuremark documentation, which, if you wish, can understand everything:
Technologies and effects
Multi-threaded rendering rendering performance, became the possibility of parallelized rendering. All modern CPUs have multiple processing cores, and previous versions of DirectX were limited to running commands on a single thread. Previously, only one CPU core could issue commands to the GPU, which often caused performance to be limited by just one CPU core. Multi-threaded rendering in DirectX 11 makes it possible to use all the CPU cores to perform rendering tasks. This feature began to be used in games with DirectX 11 support, and it is quite logical that its support appeared in 3DMark 11.
Multi-threaded rendering in the benchmark is based on command lists. The test engine uses one thread per available CPU core in a DX11 test. One of the command threads is the main one, and it uses direct and deferred device contexts (immediate and deferred device context), and the rest of the worker threads (worker threads) use only deferred. The load is distributed between threads, each of them serves (calculation of transformation matrices, clipping of invisible objects, calculation of shader parameters and writing DX calls to the command list) approximately equal number of scene objects.
Tessellation
This is the most important innovation in DirectX 11, and in the graphics pipeline of all modern video chips. Tessellation allows you to render much more detailed geometry with a slight increase in computational resources and memory requirements. Together with efficient algorithms for changing the level of detail (LOD), they provide the ability to dynamically change the detail of objects depending on the distance between the surface and the camera. Which, in turn, is extremely convenient for scaling the load on GPUs of different capacities.
The 3DMark engine supports rendering with or without tessellation. Tessellation algorithms are used and the load on geometric blocks is much larger than those already used in games. The new 3DMark package uses three types of tessellation: based on displacement maps, PN triangles, and Phong tessellation. Tessellation of triangles and quadrangular primitives (quad) is supported.
The degree of partitioning of primitives is calculated based on the lengths of the triangle faces projected onto the screen plane. The LOD algorithm is adaptive, and the degree of tessellation depends on the distance of the primitive from the camera. Triangles should be approximately the same size in pixels at any distance, and at a certain distance from the object to the camera, tessellation is turned off for it altogether. As an additional optimization, we can consider discarding (setting the degree of splitting to 0) those patches that are not included in the display area, as well as «invisible» ones — oriented in the other direction. When generating shadow maps (shadow maps), tessellation is also used.
Lighting
Back in 3DMark 11, deferred shading was introduced, which allows you to render a large number of lights in a frame with acceptable performance. First, geometry attributes are drawn into the multilayer G-buffer. Then indirect lighting is calculated using the ambient occlusion algorithm, using information from the depth buffer and the normal buffer. And then the lighting buffer is rendered based on those attributes.
The engine supports two different shading models. A more complex model uses four textures and a depth texture in the form of a G-buffer, and a simple model uses two textures and a depth texture. Lighting models in 3DMark also use two: a combination of the Oren-Nayar diffuse reflectance model and the Cooke-Torrens glare model or the basic Blinn Phong reflectance model. The simple lighting model is used in tests and demos of DirectX 10, and the complex model is used in Fire Strike. Optionally, atmospheric attenuation can also be calculated, as well as indirect illumination using the horizon based screen space ambient occlusion (HBAO) method.
Different types of light sources are supported: point, directional and «spotlights» (spot), shadow maps can be calculated for the last two types. For «spotlights», the resolution of the shadow map is calculated depending on the volume of the light source in screen space.
Scene rendering uses volumetric lighting, a technique that simulates the propagation of light in an atmosphere that contains particles of water, dust, etc. It has been around for a long time in games, albeit in a much simpler form. In 3DMark 11, volumetric lighting has already been used in several tests, and it was especially noticeable in the «Deep Sea» scene in opaque water, illuminated by bathyscaphe lanterns, and in the «High Temple» scene — in the visible rays of the sun shining through the thickets.
Volumetric illumination uses a very complex algorithm that approximates the amount of light scattered by the medium between the surface and the camera for each illuminated pixel. The algorithm is based on the ray casting method and the Rayleigh-Mie scattering and attenuation model. The lighting calculation uses two off-screen buffers for HDR: one for surface lighting and one for volumetric lighting. Before combining these buffers, the volume result is softened with a blur filter.
A new effect in the Fire Strike test is particle lighting. Particle effects are rendered on top of the calculated surface lighting and simulated on the GPU. Particles can be self-luminous or illuminated by scene lights that are chosen specifically for this. Particles can cast shadows on surfaces and other particles. When rendering lighted particles, tessellation is used, hull and domain shaders are active. The degree of splitting is set so that the triangles on the screen are the same size.
Post-Processing Filters
3DMark benchmarks use several post-processing filters, including depth-of-field simulation, halo effects, and full-screen post-processing anti-aliasing. Another tricky post-processing method is used — image distortion based on particles (particle based distortion). The distortion field generated in the 3D texture is used to deform in post-processing.
Depth of field effect gives the rendered scene a sense of volume, simulating the focusing of a virtual camera at a certain distance. Also in this algorithm, a bokeh effect is added to out-of-focus areas of the image. The DOF technique used in 3DMark is much more complex than in most games. Before the release of DirectX 11, the effect of simulating depth of field in games was most often a primitive blur on distant and near objects, and the possibility of a more complex effect has appeared only recently. We’ve seen advanced depth of field effects in gaming applications like Metro 2033 and Just Cause 2. We’ll skip the technical details, but the algorithm uses geometry shaders and complex calculations.
The bloom effect has also been known for a long time from gaming applications, and in its simple forms it does not require many resources. In the case of 3DMark, a Fast Fourier Transform (FFT) is used with a compute shader. FFTs are calculated using textures with 32 bits per color, effects are performed on a four times smaller image, the post-processing filter simultaneously applies the effects: blur, streak, flare and color halo. Additionally, a post-filter is also used, which imitates optical reflections.
Although the benchmark engine also supports normal multisampling (MSAA) for full-screen image smoothing, it is also possible to use a post-processing filter in the form of FXAA, which is applied to the finished image after the rest of the post-processing, including tone mapping.
Smoke Simulation
3DMark’s advanced smoke simulation is based on Ronald Fedkiw’s «Visual Simulation of Smoke», with the addition of the viscosity calculation from Jos Stam’s «Stable Fluids», except for temperature, which is simply redundant here. Objects interacting with smoke are voxelized as velocity and density fields, and all smoke simulation is done entirely with DirectX 11 compute shaders. 0009
In the 3DMark 11 review, we did a detailed analysis of the time spent on rendering a frame for various operations. This time there will be no such analysis due to its extreme complexity, but since the 3DMark engine and the techniques used in the test are in many ways similar to those we saw in 3DMark 11, we can make some assumptions. In addition to the rendering calls related to lighting, tessellation, and post-processing, there were added calls related to the calculations of physical effects — simulations of particle systems and liquids (in this case, smoke). Naturally, the share of time spent on lighting has become somewhat smaller (in 3DMark 11 it was 35-40% of the total frame rendering time), but nevertheless, it is volumetric lighting that remains the main computational task in the new 3DMark.
Yes, tessellation and geometry block loading has been increased a bit, but it’s kept fairly simple for high-end Nvidia chips and better suited for competitor GPUs and mid-priced Nvidia solutions. Well, most of the frame rendering time is spent in lighting shaders, including volumetric ones. There are more physics executable on the GPU in the package, but still not enough, in our opinion. The Bullet engine allowed something to be calculated on GPUs, and you can say thanks for one test of GPU physics, but still, a generally accepted benchmark should promote the technologies of the future.
Test results
We made a small analysis of the results for a number of common video cards. These studies were carried out using our familiar test system, familiar to readers from i3D-Speed materials. We tested several graphics cards from both desktop graphics vendors in all 3DMark benchmarks. First, let’s take a look at the 3DMark results for the list of common video cards in the Ice Storm test — both in terms of average frame rates in all subtests, and in the form of a final score separately for 3DMark Ice Storm.
First of all, we note the spread of results in the «physical» subtest — it clearly lacks stability, the difference on one system can reach up to 10%. It is also curious that the results with AMD Radeon graphics cards in this subtest are slightly better than with Nvidia Geforce. But we are more interested in graphical subtests. We can say that Nvidia solutions already have some advantage in them. But this is not too important, because the result of about 1000-1500 frames per second indicates that the test is too simple for these video cards.
In general, Futuremark notes that the Ice Storm benchmarks can experience significant scatter when running on the same system. This happens because the test is too easy for modern PCs and is more suitable for smartphones and tablets. On powerful systems, the frame rate easily exceeds 1000 FPS, so this test makes little sense when testing desktop GPUs. There is a rule of thumb for choosing the right test — if the score in Ice Storm exceeds 50,000 points, then this system is not able to reveal all its capabilities in the test and you need to use the more difficult tests Cloud Gate or even Fire Strike. So let’s move on to the first of them.
The numbers in Cloud Gate are already somewhat more interesting, although the FPS results are still too high for top solutions. But the physical test is already clearly more stable — the difference in FPS between all the results does not exceed 1%. As for graphics tests, they are well suited for testing modern video cards from inexpensive and mid-price segments. And for top-end solutions, the 3D load is still rather weak. But the difference between solutions of different levels is immediately visible: Radeon HD 7970 and HD 7870. In this test, AMD and Nvidia video cards perform approximately the same.
In the case of the highest performing multi-GPU systems such as SLI and CrossFire, performance in Cloud Gate will again be limited by system and CPU power, and top-end graphics chips in Cloud Gate will simply not be able to show decent results. For such systems, Fire Strike tests should be used. In general, let’s clarify again that for all systems, you need to choose the appropriate test that will show the advantages and disadvantages of the GPU without being limited by other system components. It is impossible to do this in one test, so 3DMark included three different tests, and one of them also has extreme settings. Consider first a simpler option:
Fire Strike already includes four subtests, a combined one has been added — with simultaneous load on the CPU and GPU. Since we had the same CPU for all tests, the rendering speed in this subtest depended solely on the GPU. The third physical test again showed stable results with a spread of less than 1%, that is, it is quite suitable for testing the power of universal computing cores.
The Fire Strike graphics tests are quite applicable for testing powerful modern solutions — top cards show 27-37 frames per second on average, which can be considered the optimal mode. And for less powerful solutions, Fire Strike is suitable, but if you try the GPU a notch lower than our Radeon HD 7770 and Geforce GTX 650 Ti, then the frame rate will be too low. In this test, the solutions of the two companies are close to each other, but the Radeon HD 79 still has a slight advantage. 70. Let’s look at the results of these video cards in the «extreme» test mode:
It is very curious that the stable results of the physical subtest in the normal Fire Strike mode were replaced by not very clear numbers with a clearly larger spread in the extreme profile. Apparently, the result is somehow affected by the changed rendering resolution. As for the graphics subtests and the combined variant, it is clearly seen that Fire Strike Extreme is only suitable for the most powerful GPUs of the Radeon HD 79 level.70 and Geforce GTX 670/680. Yes, and they show only 12-18 FPS in the test, which is the lower limit of acceptable frame rates for testing. In this test, the top-end Radeon graphics card again has some advantage.
Slow solutions fall deep down, which can be explained by a seriously increased load on the GPU. For example, for the Geforce GTX 650 Ti and Radeon HD 7770, this Fire Strike test profile is definitely not suitable due to the extremely low performance of 0. 3-3 FPS. At the same time, the NVDIA video card drops FPS more due to the lack of video memory (it’s no secret that Nvidia solutions use slightly more video memory on average compared to AMD video cards). Therefore, Fire Strike Extreme is best used only in the case of top-end single-chip cards and systems based on multiple GPUs, and it is poorly suited for weaker video chips.
In general, a brief study of the results of the new 3DMark on an incomplete set of video cards of different levels from AMD and Nvidia showed that it can be used to test 3D performance. There are all the necessary conditions for this: graphics tests depend mainly on the power of the GPU (provided that the test for the video card level is chosen correctly, which should not be forgotten), and processor tests depend solely on the speed of the system’s central processor.
Conclusions
As the theater begins with a hanger, so we will start drawing conclusions from the name of the test package. The reasons for removing any reference to the release year or version of the package have been explained, but still not very clear. How to distinguish modern 3DMark from previous versions? The easiest way is to call it 3DMark 13 or 3DMark 2013, and fellow journalists from other publications agree with us.
The rest of the test package can be praised — it has clearly become better than the previous one. Firstly, cross-platform — cheers, now we can compare elephant with a whale smartphone with a powerful home PC! Secondly, the variety of tests in 3DMark clearly benefits it — now you don’t need to keep all these old versions of 3DMark, but you can use only the current package for any comparisons. Well, the presence of demos for each of the three tests allows you to increase public interest in the package — it is in them that the graphics capabilities of a certain level (DX9, DX10 and DX11).
Cross-platform package is very important, especially in these times, when the image of the PC has clearly faded behind the brilliance of all these numerous tablets and communicators. The reader is now interested in questions like: how much slower than an old laptop is my smartphone in games? Which is more powerful, Apple iPad 3 or Google Nexus 10? Very soon we will be able to answer most of these questions — we just need to wait for the release of the corresponding mobile versions of 3DMark. It would also be very nice to have some kind of pivot table in the results of different systems on the Futuremark site — and we think that this will definitely appear.
But the cross-platform implementation also has an obvious drawback: only one of the three tests exists in all mobile versions, which is clearly not enough. It is clear that Fire Strike is too heavy even for mobile solutions and most desktop solutions, and so far it does not need to be dragged to tablets and smartphones, but Ice Storm, in turn, is too light for all PCs and shows 1000 or more frames on the most powerful systems per second. So comparing the most powerful mobile systems (like the Apple iPad 4 or the upcoming Nvidia Project Shield) in terms of 3D rendering speed with existing PC systems would be nice in a test like Cloud Gate — the best mobile GPUs can pull this level.
As we have already mentioned, there are no ideal 3D test suites, but of the existing 3DMark developers from Futuremark, it has been the closest to this ideal for many years now. The results obtained in the new version on various video cards (so far we can’t say anything about mobile systems like smartphones and tablets based on the ARM architecture) are quite consistent with the level of their comparative performance in other applications. As in real games, a powerful GPU is more important than CPU speed in 3DMark, although games always have uses for «extra» CPU power (like artificial intelligence and physics effects).
The main innovations of DirectX 11 in the form of tessellation and computational shaders in 3DMark are used quite actively, but there are complaints about the effectiveness of the applied algorithms and techniques. Although compared to static geometry with tessellation in 3DMark 11, animated and tessellated characters appeared in the Fire Strike test, and there were also more physical calculations on the GPU. However, there are not too many modern computational shaders in the test, and the effectiveness of the existing ones is questionable — just remember the volumetric lighting, which is too heavy for the GPU and is not used in games so far. And while 3DMark benchmarks don’t look like the games of the future, this is Futuremark’s vision, and they have every right to be.
But we would like to see test scenes in 3DMark that would load computational shaders and geometry blocks separately from each other, and only in the most difficult test would they load everything together. In this case, one could draw some conclusions that one GPU is faster at processing geometry, while the other GPU is stronger when calculating complex compute shaders. Yes, and Futuremark synthetics have long been thrown out, it was only in 3DMark Vantage. In general, the 3DMark package is quite suitable as a 3D benchmark, but to study the performance of tessellation, it is better to take the same Unigine Heaven.
A study of the performance of several modern video cards in the new 3DMark showed that it is well suited for testing general 3D performance, the results in it depend mainly on the power of the GPU and are quite in line with comparative performance in modern games. The only clarification is that it is imperative to choose the right test from the three existing ones for a correct comparison. We also really liked the output of graphs with per-second display of performance, temperature, frequencies, etc., although not all of the corresponding functionality has been implemented so far.
The people at Futuremark were clearly trying to make this version of 3DMark better than 3DMark 11, and we think they succeeded. The test package will become the new universally recognized standard, even despite some of its shortcomings. It’s definitely better, more interesting, and more useful than 3DMark 11. 3DMark 2013 has everything you need to test any hardware with relatively new GPUs. Three tests of different levels and load allow you to compare smartphones, tablets, laptops and gaming PCs running various operating systems.
Nvidia GeForce RTX 2080 video card review and testing (page 3)
Rise of the Tomb Raider
Min|Avg.
1920 x 1080
Please enable JavaScript to see graphs
2560 x 1440
Please enable JavaScript to see graphs
3840 x 2160
Enable JavaScript to see graphs
3DMark
Settings:
- Presets: Fire Strike Ultra, Fire Strike Extreme, Fire Strike.
Futuremark 3DMark
Fire Strike Ultra, Fire Strike Extreme, Fire Strike
Graphic points
Deus Ex: Mankind Divided 2560 x 1440; 3840×2160;
FPS
Min|Avg.
1920 x 1080
Please enable JavaScript to see graphs
2560 x 1440
Please enable JavaScript to see graphs
3840 x 2160
Please enable JavaScript to see graphics
Total War: Warhammer
Settings:
- Resolution — 1920 x 1080; 2560 x 1440; 3840 x 2160;
- DirectX version — Dx12;
- Preset — Ultra.
FPS
Min|Avg.
1920 x 1080
Please enable JavaScript to see graphs
2560 x 1440
Please enable JavaScript to see graphs
3840 x 2160
Enable JavaScript to see graphs
Battlefield 1: Over The Top
Settings:
- Resolution — 1920x; 2560 x 1440; 3840 x 2160;
- DirectX version — Dx12;
- Preset — ultra.
FPS
Min|Avg.
1920 x 1080
Please enable JavaScript to see graphs
2560×1440
Please enable JavaScript to see graphs
3840 x 2160
Enable JavaScript to see graphs
Summarize
Average Frames
Medium | Medium, 1920 x 1080 |
Medium, 2560 x 1440 |
Secondary, 3840 x 2160 |
|
Radeon RX Vega 64 | 79. 7 | 112.2 | 81.5 | 45.5 |
GeForce RTX 2080 Ti | 125.0 | 157.7 | 135.5 | 81.9 |
GeForce RTX 2080 | 105.5 | 142.5 | 110.6 | 63.4 |
GeForce GTX 1080 Ti | 106.5 | 145.4 | 111.5 | 62.7 |
GeForce GTX 1080 | 83.4 | 119.1 | 83.7 | 47.4 |
Radeon RX Vega 64 OC | 86.1 | 119.3 | 88. 8 | 50.1 |
GeForce RTX 2080 Ti OC | 132.6 | 163.5 | 144.2 | 90.3 |
GeForce RTX 2080 OC | 112.7 | 148.9 | 120.2 | 69.0 |
GeForce GTX 1080 Ti OC | 111.5 | 148.9 | 118.1 | 67.7 |
GeForce GTX 1080 OC | 88.9 | 126.1 | 90.4 | 50.4 |
Depending on the load, the advantage of the GeForce RTX 2080 over the GeForce GTX 1080 ranges from 20 to 33% (average 25%). But overclocking the GeForce RTX 2080 brings it even more convincing superiority, up to 37% (average 27%). And although the most correct comparison is with the GeForce GTX 1080, we will do it differently. Since the cost of the GeForce RTX 2080 is equal to the price of the GeForce GTX 1080, then discarding all the innovations, let’s compare them face to face, and then the result will be more interesting.
According to new data, it turns out that the GeForce RTX 2080 is identical to the GeForce GTX 1080 Ti plus or minus a fraction of a percent. Thus, the GeForce RTX 2080 manages to step on the heels of the previous leader (we will not take into account the very specific Titan). We add here less power consumption and a lower operating temperature.
GeForce RTX 2080 benchmarks in new games
The ASUS ROG Swift PG27UQ is a 27″ 4K/Ultra-HD IPS gaming monitor with a high refresh rate (up to 144Hz). Thanks to quantum dot technology, it has a wide color gamut (DCI-P3), and HDR support means increased contrast, so this monitor delivers an incredibly realistic picture with rich colors. There is a built-in light sensor to automatically change the screen brightness according to the surrounding conditions. The appearance of the device can be personalized with Aura sync lighting and built-in projection elements. ROG Swift PG27UQ is the technology of the future, available now!
The test would not be objective if we did not take advantage of the situation and test DLAA and RTX technologies in games that are already available to us reviewers to better convey the meaning of Nvidia’s innovations.
Infiltrator Demo
TAA/DLSS comparison results (4K resolution):
- Avg. number of frames per second — 48/66;
- 1% min. number of frames per second — 14/37.
Final Fantasy XV
TAA/DLSS comparison results (4K resolution):
- Avg. number of frames per second — 32/45;
- 1% min. number of frames per second — 25/35.
This anti-aliasing method really improves performance without sacrificing quality.
SDR vs HDR
HDR performance has previously been relatively resource intensive for Nvidia graphics cards. And it was connected with Tone mapping, which was performed by shaders. Now everything happens in hardware.
Battlefield 1: Over The Top
The GeForce RTX 2080 in SDR produces a minimum (1%) of 69 fps, with an average of 83 fps. In HDR, the numbers are 65 fps and 82 fps, respectively.
Hitman
The GeForce RTX 2080 in SDR produces a minimum (1%) of 63 fps, an average of 79 fps. In HDR, the numbers are 60 fps and 73 fps, respectively.
Shadow of the Tomb Raider
SDR vs HDR: GeForce RTX 2080 in SDR produces a minimum (1%) of 38 fps, with an average of 46 fps. In HDR, the numbers are 32 fps and 46 fps, respectively.
As you can see, the percentage of performance drop has ceased to be depressing.
Conclusion
How and with what to compare the GeForce RTX 2080? On the one hand, now there is simply no competitor on the market with such a speed and such a set of capabilities. The only exception is the previously reviewed flagship of the Turing lineup, but the price of the GeForce RTX 2080 Ti can scare off potential buyers.
On the other hand, for the GeForce RTX 2080 it is 100% explained by: speed, which is comparable in already released games with the GeForce GTX 1080 Ti, minor innovations (Mesh shading, Variable Rate shading, Texure space shading), potentially useful things (ray tracing) , already giving a visually significant increase in speed (DLSS), and, of course, overclocking to 2 GHz.
It turns out that the GeForce RTX 2080 has more chances to become a hit, despite the lower speed compared to the GeForce RTX 2080 Ti. Yes, and the objective merits of this are confirmation. But the most important thing is the reasonableness of the price for new chips — only $ 50 on top.
Rasamaha (Dmitry Vladimirovich)
0138
Subscribe to our channel in Yandex.Zen or telegram channel @overclockers_news — these are convenient ways to follow new materials on the site. With pictures, extended descriptions and no ads.
Futuremark 3DMark 2.19.7216 Developer Edition (2021) PC
3DMark is a popular program that checks the stability of the processor, memory status, with a graphics card, tests its performance.
The utility will show how well the computer on which the user is playing the game will perform. The verification itself takes place in an environment close to the game.
PC diagnostic software — 3DMark is a paid service. If you turn on the demo version, then a couple of tests will be available to you.
Through the program it will be possible to view the state of the video card, with the processor.
In the material we will tell you how to use the free version of 3DMark?
Unigine Heaven
Summary
A computer benchmark developed by a Russian company based on its own engine. Benchmark uses the achievements of the Unigine engine to conduct stress tests of graphics accelerators. Interactive walks and cinematic scenes, created using the latest advances in 3D computer graphics, load any graphics card to the limit. This checks its stability and the quality of the cooling system.
Unlike similar programs, Unigine Heaven created the scenes in such a way that one would like to watch them: the laboratory of a mad professor at work, a tropical beach with palm trees and surf, a small village with paved roads, cozy buildings and a dragon statue, a gothic-style chapel, a forest array. Each scene must be downloaded separately. The program also contains algorithms that even top-end video cards cannot withstand.
In the settings you can specify the DirectX used, the quality of shaders, filtering, anti-aliasing, tessellation, resolution. The process is started with the «Run» button.
Panoramic view for loading video memory
Settings during the test
Download
Fire Strike
To run the Fire Strike test, you need various graphics adapter with full support for DirectX 11 features. serious computing resources of the GPU, and therefore it is expected that it will be used to evaluate the capabilities of charged gaming platforms.
It takes approximately 15 minutes to complete the full test cycle including demo scenes.
The basic free version of 3DMark Basic Edition allows you to evaluate the performance in all tests, however, they will be mandatory to run with default settings, and you will need a network connection to view the results. Fine-tuning capabilities appear in the 3DMark Advanced Edition, an upgrade to which will cost $24.99. In addition, the extended version has a whole list of features that make it easier to use the test suite. For example, you can get detailed results in the 3DMark application itself and save them locally.
A function with interactive graphs that show CPU/GPU temperatures, the number of frames per second at any time during which the test was running, as well as processor power consumption can also be useful (we will check with the developers on how to calculate it).
Of course, the advantage of the Advanced Edition is the ability to selectively select a specific stage and skip demos to reduce the time it takes to pass tests. The advanced version also allows access to Fire Strike’s Extreme mode and the ability to run certain steps in a loop to test system stability.
The $995 version of 3DMark Professional Edition provides a commercial use license, command line capability, and greater flexibility in managing results (saving to XML). In addition, this edition provides access to the Image Quality Tool, which allows you to evaluate and compare the rendering quality of specific frames on different adapters.
25% discount available for previous 3DMark 11 Advanced Edition owners. The same opportunity to save when buying within one week after the announcement of 3DMark will be available to users of the Steam service. Curiously, due to an oversight or with the intention to attract more users, but at the time of preparation of the material, the cost of the extended version on Steam was $ 11.24, i.e. The 25% discount was made from $15, not the planned $25. A curious feature of the Steam version is the presence of a number of funny «achievements».
MSI and Galaxy have become partners in the development of the new version of 3DMark, so some products from these manufacturers will be bundled with keys for free activation of the Advanced Edition.
Measure the performance of your PC with 3DMark
A test program is a program designed to reduce the maximum performance of the hardware and thus let us know its actual performance. There are all kinds of programs for analyzing computer performance, from general programs for measuring computer performance in general, such as AIDA64 or PCMark, to specialized programs in specific components, such as Cinebench for CPU or 3DMark for graphics cards.
3DMark
3DMark is a reference software developed by Futuremark (UL) that allows us to measure the performance of our computer while playing games. This program consists of a series of tests, from less demanding to more demanding, so that we can see in detail how far our processor and our graphics card can go. We can measure performance from mobile gaming to 4K gaming.
This program is ideal for comparing different computer configurations, measuring the difference in performance between two or more components, and above all, to determine how much our hardware improves when it is overclocked. It even has a stability test that will let us know if the overclock is stable or could cause problems.
This tool has specific tests for DirectX 12 , DirectX 11 and older versions such as 10 and up to 9. In addition, it has stress tests, stability tests, and even specialized benchmarks to measure GPU bandwidth and performance with Ray Tracing and DLSS.
Futuremark benchmark software has been running since 1998 for over 20 years and has always been up to date, including support for the latest technology and updating its benchmarks to the latest versions of DirectX.
Where to buy 3DMark for Windows
While until recently you could buy this software directly from its main website, now the best option we have is to buy it at Par. In this store we will find the best deals and in addition we can buy various accessories to improve the functions of this software.
The official price of the tool is $29.99 on the website. However, on Steam we can get a pack with all the accessories (even the latest Port Royale tests) for less than 25 euros. And if we wait for a good offer, we can get all Futuremark tools for less than 10 euros. Deal.
We can also Buy a 3DMark license from Steam sites like Instant Gaming, saving us money. For example, on this website this license costs 18 euros, but we must take into account that it does not come with all DLCs.
3DMark DLC for all test types
All add-ons in the 3DMark test suite include:
- Spy Time : for gaming computers with DirectX 12. We can choose between normal and «extreme».
- Night Raid: DirectX 12 test for computers with integrated graphics.
- Firestrike : For computers running DirectX 11. There are several editions: normal, ultra and extreme.
- Sky Diver: For gaming laptops and midrange computers.
- Cloud Gate: For basic laptops and midrange computers.
- Ice Storm: analyze smartphones and tablets.
- API Overhead: Measures the performance of the machine’s API.
- PCI Express: Measures the throughput of a PCI Express port.
- VRS: VRS test to measure performance and image quality.
- Port Royal : Test to measure RTX graphics performance when moving Ray Tracing.
- DLSS Benchmark: Performance test with and without Deep Learning. For RTX graphics.
The most important and specific for a gaming computer are Time Spy, Fire Strike and Port Royale (if we have an RTX image).
We will also have a trial version of VRMark, a virtual reality testing software.
If we buy the application ourselves, without a package like Steam, we can use all the tests except Time Spy and Port Royale, along with their included subtests (like DLSS).
How to measure the performance of my PC with 3DMark
Once we have this application (particularly in our Steam library), we download and install it as if it were any game. After loading, we execute it to go to the main program window.
In the main window we can see all the tests we need to measure the performance of our computer. If one is not installed, we can download it from this application as we need. In addition, if we do not have a purchased item, we can also buy it here.
When we select a test, we see its summary window. Each test is set up as a baseline to run with a configuration to keep the data as neutral and comparable as possible across different computers and components. In any case, if we want, we can run individual tests with the configurations we need.
When we run the test, the program downloads all the resources and executes them. Each test usually takes several minutes. And while they are working, it is better not to do anything with the computer so that the values \u200b\u200bare as reliable and real as possible.
When each of the tests ends, we see the total result. We can also see the values obtained in each of the individual tests, as well as a summary of the equipment used and other detailed information.
In addition, we can also see our results online. In this way, we can find out more information about them, about our PC and about the notes that other users make with similar equipment.
Furmark
The program offers a single test: the spinning furry ring scene. To start testing the GPU, click «Burn-in test». You can change the available settings beforehand:
- resolution;
- smoothing;
- dynamic camera and background;
- duration;
- GPU peak temperature.
The «GPU stress test» button will run an extra test lasting 15 minutes at the resolution you choose.
Furmark’s only test
During testing, the display shows the scene itself, a temperature graph of the device, and a number of information about the test. After a couple of minutes, the cooler will start to work noticeably louder, and after a few more — almost at maximum.
Download
3D Mark Performance Comparison
Having determined the characteristics of a video card, you probably want to know what it can do in comparison with others. A classic of the genre will help satisfy curiosity — a set of synthetic tests for video of any 3D Mark models.
3D Mark is good because it allows you to check the performance of the entire structure of the computer’s video subsystem. By the way, its latest (12) version includes a new DirectX 12 health check module.
The program is available in several editions, one of which is free. It includes the same set of tests as the main ones — paid ones, but the user is deprived of the opportunity to change their settings and run them separately. The results in the free edition are evaluated not in the application, but online on the developer’s website (Futuremark).
Paid license holders are spared this inconvenience. However, the most budget version of 3D Mark today costs about $30, and you will have to spend about $10 to upgrade one of the previous ones. With constant use of the program, this is justified, but for a one-two-time launch, for example, before and after overclocking the video card, you can get by with the free version.
By the way, with the help of 3D Mark you will not only know the true performance of your video card, but also check it for overheating and hidden defects. If flickering, ripples, loss of textures, twitching of the picture appear during the test, you should determine the temperature of the graphics chip, processor and other PC components. If it clearly tends to the maximum, stop testing and ensure that the devices are properly cooled. Insufficient heat dissipation can also cause the computer to shut down during a 3D Mark run. It is better not to allow this, because the hard drive may suffer from a sudden power outage.
If the appearance of artifacts is not accompanied by excessively high temperatures or other problems occur — blue screens of death (BSoD), freezes, loss of signal to the monitor, etc., you can suspect a hardware malfunction of the video system components. First of all — a graphics chip, video memory, elements in the circuits for forming the supply voltages of these nodes or the power supply itself. In the most favorable cases, this turns out to be a video driver error.
How to use 3DMark
There are many different programs for testing processor performance and stability, but one of the most popular is 3DMark. The main advantage of this program is the ability to check not only the processor, but the system as a whole, including the graphics card, RAM, and other parameters. The test is performed under conditions close to games, so the program allows you to most accurately assess how well the system will behave in games.
The 3DMark program is paid, but the developer provides several free tests as a demo version. They will be quite enough to test the processor and video card. This article will explain how to use 3DMark. And we will focus on the free version.
Installing 3DMark
Unlike other programs, you can’t install 3Dmark from the official site because there is no download link. There is only an opportunity to buy the program. However, you can download a demo version of the program from Steam. To do this, of course, you need to have Steam installed on your computer.
Open Steam and type in the search 3DMark :
Select the first version found and scroll down a little on the page that opens. On the right there will be a button Download demo version , you need it:
The program is quite large, so the download may take a long time, depending on the speed of your Internet. When the download is complete, the program will appear in your Steam library:
From here you can launch it.
Russian interface
By default, the program is delivered in English. However, in the settings you can also enable the Russian language. To do this, open the menu Options and in the item Language select Russian :
Now the program will be in Russian.
Launching the first test
Immediately in the main program window, you are prompted to launch the Time Spy test suite. This is a small scene in which a person walks around the museum with a special magnifying glass that allows you to look into the past and look for weapons. You can watch the full scene here:
To run the test just go to the main page and select Run or Run :
First you will see the demo scene itself: frame number:
After the testing is completed, the utility will give you an overall score of your computer for all tests:
It can be compared with the scores of other users on the Internet:
Running other tests
In addition to the Time Spy test, you have a few more scenes available in the free version. These are:
- TimeSpy DirectX 12 performance test;
- FireStrike DirectX 11 performance test;
- Nigth Raid — DirectX 12 test for computers with integrated graphics;
- Sky Driver — DirectX 11 test for computers that can’t achieve high scores in the FireStrike test;
- Cloud Gate — DirectX 11 and DirectX 10 test for laptops and regular home PCs;
- Ice Storm Extreme — for mobile devices and tablets;
- Ice Storm — for budget smartphones.
You can run any of these tests on the tab Tests :
Just open the desired test and click the button Run :
In the paid version, settings are also available that allow you to turn off the demo video, as well as select those tests that you want from the set. Basically, with the help of these tests, a 3dmark video card is checked, but the processor is also quite loaded.
Stability tests
You can use the stability tests to test the stability of your computer after building a new computer or after overclocking. But they are only available in the paid version on the tab stability tests:
You don’t have to use 3DMark for this purpose, you can use the same AIDA64 which has a free trial or Prime95 which is completely free.
View results
Also in the free version you can view the results of the 3dmark test at the last check. To do this, open tab Results :
Conclusions
In this article, we have discussed how to use 3DMark to test CPU and GPU performance. As you can see, the free version of the program is enough to understand what your computer is capable of. And if you need more, then you can use other free programs to test the processor in Windows.
, please select a piece of text and press Ctrl+Enter.
Source: https://te4h.ru/kak-polzovatsya-3dmark
How to test a video card and improve the results?
If you perceive information better with your eyes, watch this review video with instructions for working with the benchmark:
If your video card is old, with insufficient cooling, then we do not recommend testing it at all. This may cause it to break.
After running the test, you will receive a certain number of points. The more points, the more powerful your system. What can be done to improve performance?
First compare the scores with other computers with the same hardware. If your readings are lower, the problem may lie in insufficient cooling, in settings, or in defective components. Above-average numbers speak of overclocking potential.
Update Windows, install newer drivers, try different system settings. Recheck your score after any changes to help you understand your weaknesses. Think it’s time to schedule a computer upgrade?
System ⇒ Futuremark 3DMark 2.18.7185
Rate news / program! 4.8 out of 5, total assessments — 4 |
|
3DMARK
— a benchmark that combines the processor, tools for determination of the quality of the image. Futuremark Corporation, a well-known manufacturer, is perhaps one of the most popular test packages in its class for measuring the performance of modern personal computers and their components.
Futuremark
develops its tests taking into account the opinions of companies such as NVIDIA and AMD, listening to a little bit of everything with varying success, so their benchmarks turn out to be relatively honest, without a clear bias in anyone’s direction. Since 3DMark is widely used by users and manufacturers, the results database contains numbers from millions of different systems. This feature makes this test package very convenient for comparing performance against other systems, including overclocking, and is often used by overclockers. Yes, and the specialized press considers 3DMark one of the mandatory tests for their materials, even taking into account the presence of gaming applications.
3DMark
is available in three editions: free Basic Edition, Advanced Edition and Professional Edition. Even in the free Basic Edition, all tests are available, but only with standard settings — without the possibility of changing them. Also, there is no support for Extreme mode for the Fire Strike test. That is, in this version, it is possible to run three tests in the default settings, and the results are shown on a special HTML page only when the Internet connection is active.
Features:
- Includes three stunning tests that run in DirectX 9, DirectX 10 and DirectX 11 modes.
- Test everything from tablets to gaming PCs.
- Compare your results with other Windows, Android and iOS devices.
- Compete for the title of the fastest PC on Steam (built-in support for Steam Leaderboards).
- Easy to use, no technical knowledge or skills required.
OS:
Windows XP, Vista, 7, 8, 8.1 (64bit)
You have been denied access to this information or the block has been removed at the request of the copyright holder. In order to gain access, you need to Authorize on the site
Subscribe to news Futuremark 3DMark 2.18.7185
If the news changes, you will receive a notification by E-mail.
Subscribe
Already subscribed: 3
Thanks: 17
-=zlo=-, NaumVlad19881980, saby saba, workerrx580
, AlexDiakov, Stas159,
AlexAlex23
, Pipiska,
rm1508
, Alley, palych262, Konstantin2, joker.jx, oldjuri, matiyo37 , NadezdaKrypskaya, ksileks
Test results
We conducted a small study of 3DMark capabilities using a test platform and video cards available in the laboratory at the time of the announcement of the application. The current list of participants is not indicative, in the future the list will be expanded.
Core [email protected]/MSI Z77A-GD65/2×4 GB DDR3-2133, Windows 7 (x64), drivers — NVIDIA ForceWare 310.90, AMD Catalyst 13.1.
Nearly 1500 fps in Ice Storm provide an indescribable smoothness of the picture, but, alas, the human eye cannot appreciate all the delights of such a frame rate. The 1.5-fold difference in system performance between Radeon HD 7770 and HD 7970 is an obvious sign that this stage is of little use for evaluating the capabilities of desktop systems.
Cloud Gate, although it generates a noticeably higher load, it is obvious that this stage does not allow us to appreciate the benefits of powerful components. Although the spread of results is higher than in the previous case, but more than 200 fps (Radeon HD 7970) in the test scenes, this subtest will be relevant for comparing systems with integrated graphics and entry-level video cards.
Fire Strike is a completely different matter, where a platform with an overclocked Corei7-3770K processor and a Radeon HD 79 video card70 was able to produce 25-30 frames / s on the mountain, and for acceptable performance in Extreme mode, a dual-chip GeForce GTX 690 is desirable, or rather a couple of such video cards.
Note that in addition to a serious load on the graphics adapter, during subtests with calculations of the physical interaction of particles, serious resources are also required from the central processor. As an engine for calculating physics, the open Bullet Physics and a multi-threaded algorithm are used, which generates, if not the maximum possible load, then at least capable of well loading a quad-core CPU with Hyper-Threading.
GeForce GTX 1070 Ti 3DMark test results
Very soon, NVIDIA will announce official details about the GeForce GTX 1070 Ti, which will take an intermediate position between the GeForce GTX 1070 and GTX 1080. Formal announcement and start accepting pre-orders for the GTX 1070 Ti in retail stores are scheduled for next Thursday, October 26, and a week later, on November 2, the first reviews of the accelerator will be published and shipment of cards to customers will begin.
Meanwhile, enthusiasts continue to scour the Internet for performance information. This was found in the 3DMark Achievements database, where there are four results for the GeForce GTX 1070 Ti: two in the Fire Strike Extreme subtest (9449 and 9546 Graphics points), one each in Time Spy (6777 Graphics points) and in 3DMark 11 Extreme (9002 Graphics points) .
Futuremark test suites are usually not a reliable tool for determining the characteristics of video cards, and this case is no exception. In three entries in the database, the core frequency was 1886 MHz, and in one it was 1607 MHz (recommended values for the GTX 1070 Ti are from 1607 to 1683 MHz, excluding boost mode). The frequency of the GDDR5 buffer memory in 2003 (8012) MHz was determined correctly, as was its size of 8 GB.
Taking the 3DMark results from the Radeon RX Vega 56 review as a reference, we can roughly estimate what the upcoming new product is capable of. The average 3DMark Fire Strike Extreme score of 9498 points is 1% higher than the RX Vega 56, while the 3DMark Time Spy score of 6777 means a 7. 9% improvement over the same graphics card. Given the possible overclocking limitation of the GeForce GTX 1070 Ti adapter, it is too early to name the winner in its confrontation with a representative of the AMD Radeon family. Also note that the tested GTX 1070 Ti is closer to the GTX 1080 than the GTX 1070 (especially in Time Spy).
3DMark (Graphics Score) including preliminary results GeForce GTX 1070 Ti | ||||||
Resolution | NVIDIA GeForce GTX 1070 (1506/8008MHz, 8GB) 8 GB) | NVIDIA GeForce GTX 1080 (1607/10008 MHz, 8 GB) | AMD Radeon RX Vega 56 (1471/1600 MHz, 8 GB), Turbo | AMD Radeon RX Vega 64 (1546/1890 MHz, 8 GB), Turbo | ||
Fire Strike | 1920 × 1080 | 17 519 | 21 694 | 19 848 | 22 503 | |
Fire Strike Extreme | 2560 × 1440 | 8298 | 9498 | 10 264 | 9488 | 10 711 |
Fire Strike Ultra | 3840 × 2160 | 4079 | 5001 | 4774 | 5 400 | |
Time Spy | 2560 × 1440 | 5693 | 6777 | 7111 | 6281 | 7079 |
The entries in it show that for 9002 Graphics points it is enough to thoroughly overclock the GeForce GTX 1070 — to frequencies of about 2000-2100/2250-2375 (9000-9500) MHz for the core and memory, respectively.
Features
The software provides an objective and accurate assessment of the performance of video adapters and processors. 3DMark has 4 gaming benchmarks, several CPU and individual 3D component tests, and it also performs image quality tests.
List of other features of the utility:
- Support for DirectX technology of various generations.
- Availability of Cloud Gate, Ice Storm, and Fire Strike tests.
- Determining the number of FPS.
- Provide detailed information about the specifications, status, and health of your computer.
- Presence of a built-in game.
- Ability to register a free online account to view test results.
- Formation and analysis of graphs with information about the platform used.
- Access to adjust the screen resolution.
- Capable of testing laptops, tablets, and high performance gaming PCs in 4K resolution.
- Compare test scores with other Windows, iOS, and Android devices. Geekbench is a powerful utility for testing CPU and memory.
- UNIGINE Benchmarks — checking the power of the computer at maximum loads and the stability of the NVidia video card and its cooling.
- Furmark is a diagnostic tool for OpenGL and GPU performance using fur rendering algorithms. It also helps to check the quality of cooling.
- Cinebench is a set of test algorithms for your PC. Analysis of the performance of the processor and video card and comparison for different operating systems — Windows and Mac OS X.
- UserBenchMark is another computer performance tester.
- Name — under what name the card is defined in the current operating system.
- GPU is the code name for the graphics chip of a video card.
- Revision — GPU revision (NVIDIA only).
- Technology — technological process.
- Die size — core area.
- Release date — GPU release date.
- Transistors — the number of transistors in the chip.
- BIOS version — video BIOS version.
- Subvendor is the manufacturer of the final product (in our example, NVIDIA is the processor manufacturer, Gigabyte is the video card manufacturer).
- Device ID — identifiers of the GPU and chip manufacturer.
- ROPs/TMUs — number of rasterization/texturing units.
- Bus interface — PCI-e bus interface.
- Shaders — the number and type of pipelined (shader) processors.
- DirectX Support is the supported version of DirectX.
- Pixel Fillrate — pixel rendering speed.
- Memory type and bus width — video memory type and bus width for data exchange between memory and GPU.
- Memory size and Bandwidt — video memory size and bus bandwidth.
- Driver version — video driver version and operating system.
- GPU clock and memory — current clock frequency of the system bus and memory.
- Default clock — the same without overclocking.
- Multi GPU (SLI/Crossfire) — support and use of video collaboration technologies.
- Computing — support for computing technologies.
- Equipment test
- Custom test settings
- Several types of tests with the possibility of separate connection
- Saving measurements
- 4 tests to test system performance in gaming conditions
- processor (CPU) performance measurement
- graphics quality test (checking a video card and its drivers)
- analysis of shader capabilities
- Spy Time : for DirectX 12 gaming computers. We can choose between normal and «extreme».
- Night Raid: DirectX 12 test for computers with integrated graphics.
- Firestrike : For computers running DirectX 11. There are several editions: normal, ultra and extreme.
- Sky Diver: For gaming laptops and midrange computers.
- Cloud Gate: For basic laptops and midrange computers.
- Ice Storm: analyze smartphones and tablets.
- API Overhead: Measures the performance of the machine’s API.
- PCI Express: Measures the throughput of a PCI Express port.
- VRS: VRS test to measure performance and image quality.
- Port Royal : Test to measure RTX graphics performance when moving Ray Tracing.
- DLSS test: performance test with and without deep learning. For RTX graphics.
- Close other applications, including downloading files from the Internet.
- Run the program, it will determine the device model, available resolution and current temperature.
- Set the standard resolution in the «Resolution» field and open the settings by clicking the «Settings» button.
- In the test options, check the boxes for «Dynamic background» and «Burn-in». Do not use the «Xtreme burn-in» mode if you do not understand the intricacies of computer equipment — it can burn video memory.
- Return to the main screen, click «Burn -in test».
- Confirm that you agree with the possible risks — rebooting the PC due to excessive load intensity.
- Video card performance test starts.
- You will see the temperature graph — in the first minutes it will start to rise sharply, then it will stabilize.
- Download the application from Steam through the client from under the account and run it. Download
- Night Raid — DirectX 12 test for devices with low-power integrated graphics;
- FireStrike — for graphics cards with DirectX 11 support;
- Sky Driver — analogue of FireStrike for budget devices;
- Cloud Gate — DX 10 and DX 11 graphics test for non-productive computers;
- Ice Storm Extreme — portable PC testing.
- TimeSpy DirectX 12 performance test;
- FireStrike DirectX 11 performance test;
- Nigth Raid — DirectX 12 test for computers with integrated graphics;
- Sky Driver — DirectX 11 test for computers that can’t achieve high scores in the FireStrike test;
- Cloud Gate — DirectX 11 and DirectX 10 test for laptops and regular home PCs;
- Ice Storm Extreme — for mobile devices and tablets;
- Ice Storm — for budget smartphones.
- Ice Storm for mobile devices
- Cloud Gate for entry-level systems
- Sky Diver for budget gaming PCs
- Fire Strike for powerful PCs
- Time Spy for testing the system to work with DirectX 12
Maximum useful information
Before proceeding with the load tests, it would be nice to study the characteristics of the card. The free utility GPU-Z
will help you get quite complete and detailed information about it.0009
The «Lookup» button will take you to the GPU-Z developer’s site, more precisely, to the description and comparison page of your video card.
The second tab of GPU-Z — Sensors, takes sensor readings.
By default, it displays the current GPU and memory clock speeds, GPU temperature, cooling fan speed, percentage of GPU and memory controller usage, load on the data bus, GPU supply voltage, and reasons for its current performance degradation. If desired, you can change the setting so that the program displays not the real, but the maximum, minimum or average indicators of the sensors, and also enable the logging file.
Information similar to GPU-Z is provided by many other programs that analyze hardware devices, for example, the free HWiNFO utility and the paid Aida64. By the way, the latter allows not only to obtain information about the equipment, but also to test it. For a video card, Aida has a GPGPU benchmark and a load test to check for stability. However, the tests will be discussed further.
All at once: 30 current video cards in the 3DMark 11 test
As a rule, when testing new video cards, we select rivals based on their cost and market positioning. But sometimes you want to see the whole picture — how the performance of a full range of modern video cards compares, from the most budgetary to top-end solutions. This test is just dedicated to this issue.
We took 30 currently relevant video cards. In the photo below you can see what this magnificence looks like when you put them all together. For this group photo, the video cards had to be placed on the floor, since they simply did not fit on any office table.
Tested on a bench with the following configuration:
0004 | Catalyst 11.5 for AMD Radeon and Forcware 270.61 for NVIDIA GeForce |
the industry standard due to its simplicity, high repeatability and the use of the most advanced 3D visualization technologies.
Let’s see how the «ranking table» looks like for modern video cards that support DirectX 11. Let’s start with the 3DMark 11 Graphics Score results obtained in the Extreme preset, that is, the maximum load on the video subsystem of the computer.
Among the dual-core video cards, the Radeon HD 6990 looks the best, and in the single standings, NVIDIA’s single-chip flagship GeForce 580 GTX is the leader. Further, the results decrease almost linearly, and another qualitative leap occurs when switching to mid-range solutions using a 128-bit memory bus (Radeon HD 5770).
Of course, we also got the results in the Performance and Entry presets below. To avoid unnecessary confusion and to make it easier to compare the results of different presets (if anyone needs it), the sorting of the list of video cards has remained the same.
The picture remains the same, except that according to the results of NVIDIA video cards, they have grown somewhat in comparison with their closest neighbors from the red camp.
When switching to the Entry preset, NVIDIA accelerators again slightly improve their performance in relation to the closest neighbors manufactured by AMD. However, for top-end video cards, this does not make much difference. It is unlikely that a powerful video card is purchased for playing at a resolution of 1024×600 with low graphics quality settings. So the results in this mode may be interesting for graphics adapters of the middle and budget class.
The next test we paid close attention to was 3DMark 11 Combined. This test creates a load on both the central processor and the graphics accelerator. The CPU calculates the «physics» of rigid bodies, and the GPU calculates the movement of tissues using the Bullet engine, which runs through DirectCompute. We are interested in how video accelerators cope with the computational load, since the same central processor was used during all tests. As before, let’s start by reviewing the results in the «heaviest» mode — Extreme (the order of video cards on the diagram remains the same).
The overall picture remains the same as we saw in the Graphics test.
When switching to the Performance mode, the Radeon HD 6990 unexpectedly loses its leadership, and the single-chip GeForce GTX 580 overtakes the dual-chip Radeon HD 5970. , dual-chip video cards are simply not able to reveal their potential with the most lightweight graphics quality settings. In general, NVIDIA solutions in this mode look somewhat preferable to their closest competitors.
Some readers might be wondering why we don’t list the final 3DMark 11 results? The fact is that the final assessment includes the results of not only graphic tests, but also tests of «physics» carried out on the central processor. Therefore, it is difficult to directly compare the overall result of 3DMark 11 obtained on different systems. At the same time, the test of the graphics subsystem, especially in heavy mode, depends on the power of the central processor rather weakly, unless, of course, you are testing on a not very ancient CPU. For those who are interested, we inform you that the result of the 3DMark Physics Score of our test stand was 8750 points. And finally, below is a table with the results of all tests in various modes, including the overall score (Total).
Entry, 1024 × 600 | Performance, 1280 × 720 | Extreme, 1920 × 1080 | ||||||
3DMARK 11. Ver. 101 Prof. | Total | Graphics | Combined | Total | Graphics | Combined | Total | Graphics |