AMD Athlon XP 3000+ — AXDA3000DKV4E-WWW.NBCPU.COM
We specialized in CPU trade, contact us if you need bulk sale or purchase this kind of CPU,please do not hesitate to contact us :
Email :[email protected],Tel: +86 0755-82928539
| Type:CPU | Manufacturer:AMD |
| family:Athlon XP | microprocessors:AMD Athlon XP 3000+ — AXDA3000DKV4E |
| Type | CPU / Microprocessor |
| Market segment | Desktop |
| Family |
AMD Athlon XP |
| Model number | 3000+ |
| CPU part number |
|
| Stepping codes | AQXEA AQYHA |
| Frequency | 3000+ (rated) 2100 MHz (real) |
| Bus speed | 400 MHz |
| Clock multiplier | 10. |
| Package | 453-pin organic PGA 1.95″ x 1.95″ (4.95 cm x 4.95 cm) |
| AMD Package numbers | 27493, 28103 |
| Socket | Socket A (Socket 462) |
| Architecture / Microarchitecture | |
| Microarchitecture | K7 |
| Processor core | Barton (Model 10) |
| Manufacturing process | 0.13 micron copper process |
| Data width | 32 bit |
| The number of cores | 1 |
| The number of threads | 1 |
| Floating Point Unit | Integrated |
| Level 1 cache size | 64 KB code 64 KB data |
| Level 2 cache size | Full-speed exclusive on-die 512 KB |
| Multiprocessing | Uniprocessor |
| Features |
|
| Low power features |
|
| Integrated peripherals / components | |
| Integrated graphics | None |
| Electrical / Thermal parameters | |
| V core | 1. 65V |
| Minimum/Maximum operating temperature | 0°C — 85°C |
| Typical/Maximum power dissipation | 53.7 Watt / 68.3 Watt |
| Notes on AMD Athlon XP 3000+ | |
|
|
other microprocessors
|
AMD Low-power Athlon XP 3000+ — AXDL3000DLV4E |
AMD Athlon XP 3100+ — AXDC3100DKV3E |
|
AMD Athlon XP 3200+ — AXDA3200DKV4D |
AMD Athlon XP 3200+ — AXDA3200DKV4E |
|
AMD Athlon XP 1000 — AHD1000AMS3C |
AMD Athlon XP 1200 — AHD1200DMS3C |
|
AMD Athlon XP 3200+ — AXDA3200DUV3C |
AMD Mobile Athlon XP-M 1200+ — AXML1200GTS3B |
|
AMD Mobile Athlon XP-M 1300+ — AXML1300GCS3B |
AMD Mobile Athlon XP-M 1300+ — AXML1300GYS3B |
|
AMD Mobile Athlon XP-M 1400+ — AXMD1400FQQ3B |
AMD Mobile Athlon XP-M 1400+ — AXMD1400FQQ3C |
Tel:+86 0755-83247530 Email:raymond@nbcpu.
com
NBCPU All rights reserved..
AMD Athlon XP 3000+ 2.1GHz 400MHz 512KB 462 CPU Processor…
| General information | |
| Type | CPU / Microprocessor |
| Market segment | Desktop |
| Family | AMD Athlon XP |
| Model number | 3000+ |
| CPU part number |
|
| Stepping codes | AQXEA AQYHA |
| Frequency | 3000+ (rated) 2100 MHz (real) |
| Bus speed | 400 MHz |
| Clock multiplier | 10.5 |
| Package | 453-pin organic PGA 1.  95″ x 1.95″ (4.95 cm x 4.95 cm) |
| AMD Package numbers | 27493, 28103 |
| Socket | Socket A (Socket 462) |
| Architecture / Microarchitecture | |
| Microarchitecture | K7 |
| Processor core | Barton (Model 10) |
| Manufacturing process | 0.13 micron copper process |
| Data width | 32 bit |
| The number of cores | 1 |
| The number of threads | 1 |
| Floating Point Unit | Integrated |
| Level 1 cache size | 64 KB code 64 KB data |
| Level 2 cache size | Full-speed exclusive on-die 512 KB |
| Multiprocessing | Uniprocessor |
| Features |
|
| Low power features |
|
| Integrated peripherals / components | |
| Integrated graphics | None |
| Electrical / Thermal parameters | |
| V core | 1. 65V |
| Minimum/Maximum operating temperature | 0°C — 85°C |
| Typical/Maximum power dissipation | 53.7 Watt / 68.3 Watt |
| Notes on AMD Athlon XP 3000+ | |
|
|
| Price View | Price Range |
|---|---|
| Manufacturer | AMD |
| CPU Series | Athlon XP |
| L2 Cache | 512 KB |
| Cooling Component(s) Included | None/Processor Only |
| Condition | Certified Refurbished |
| Condition Comment | Full working order |
| Warranty | 30 Day RTB (Return to Base) Warranty |
| Important | It is your responsibility as a buyer to ensure this is compatible with your hardware or operating system before buying.
Assume that any photo is a library photo, not the actual item you will receive, unless expressly mentioned above. Item has been pulled from salvage machine so expect the item to be in a used condition with minor scuffs etc. Unless expressly mentioned in the description, there will be no other parts included with the item. This includes items such as drivers, cables, manuals, warranty cards etc. Hard drives, tape drives, caddies etc. will not include any screws, fixing rails unless expressly mentioned. |
We aims to deliver items ordered to you as soon possible. As a guide, we estimate that most items will be delivered according the following timescales. To help you plan for your delivery, we will send you an e-mail with tracking details, as soon as your order has been shipped.
UK Delivery
- Free Delivery: Monday-Friday (2-5 Working Days)
- Fast Delivery: Monday-Friday (1-2 Working Days)
- Next Working Day Delivery: if ordering before 1pm it is next day (excluding Saturday/Sunday), if you place order Saturday/Sunday we will dispatch your order Monday for delivery Tuesday.
Please note that Saturdays, Sundays, and Bank Holidays are not classed as working days.
Delivery services may take longer if you live in remote areas like the Scottish Highlands and Northern Ireland. As a result, Next Day service may not be guaranteed.
International Delivery (not including customs processing)
- You can choose UPS Courier at checkout, please fill in your address first to get a shipping quote. International orders are usually delivered within couple of days, speed depends on the services selected.
-
Goods are despatched Monday – Friday only.
-
Sending goods outside the United Kingdom may result in additional local customs clearance charges and taxes. Failure to pay these charges and taxes on arrival will result in the goods being returned to us. In such instances we may refund your order and delivery fees but retain a re-processing fee to cover the return costs and customs processing fees.
Changes from 1st January 2021 – Brexit
On 1st January 2021 the UK left the EU. A consequence of this is that goods purchased from the UK for delivery in the EU are now subject to additional duties and taxes as well as increased shipping costs. To our European customers we offering following delivery option:
Delivery Only Service – we will deliver your order excluding any customs clearance fees, import charges and taxes. When the goods arrive in your country you will be contacted by the delivery company to pay these additional costs directly to them before final delivery as the importer of record.
Delivery is processed under the ‘Delivery at Place’ (DAP) Incoterm.
We would advise you to check the cost of delivery before making purchase as the customs clearance fees, import charges and taxes can amount to quite a significant cost.
We are so confident in the quality of our work that we stand behind every MicroDream Certified Refurbished product with a 12 Month Warranty as standard and we are able to offer our 30 Day Money Back Guarantee with complete confidence.
We guarantee this product against all hardware faults for a period of 12 months from receipt of the product.
If there is a hardware fault within the first 30 days of receipt we will arrange for collection free of charge and replace or repair the product.
If there is a hardware fault after the first 30 days, you are responsible for safely returning the product to us. We can arrange collection for £19 including VAT. We will pay for the carriage of the replacement or repaired product back to you.
WHO IS COVERED? This warranty is non-transferable and covers only the original end purchaser. An original or copy of a sales invoice is required for warranty service.
WHAT IS COVERED? In the event of your PC hardware developing a fault you can return it for repair or replacement (or equivalent).
WHAT IS NOT COVERED?
- Normal wear and tear of Product use
- Misuse, lack of care, mishandling, accident, abuse or other abnormal use
- Use of the Product other than for its intended purpose
- Damage caused by improper or unauthorized repair or maintenance
- Warranty is void if repairing without our consent or warranty seal is broken
- Product that has been modified or altered
- Software issues (any problems associated with software – whether bundled or installed by user, file corruption, internet setup etc.
) - Accidental damage like breaking the screen or spilling liquid on it. You would need independent insurance to cover this.
- Used batteries, although we do guarantee that it will hold at least one hour’s charge. We will replace batteries which do not meet these criteria if reported within one week of receipt and the battery supplied is returned. Batteries are considered as consumable items — as they are expected to degrade over time. Some may require replacement.
Write Your Own Review
You’re reviewing:AMD Athlon XP 3000+ 2.1GHz 400MHz 512KB 462 CPU Processor AXDA3000DKV4E
Your Rating
Price
1 star 2 stars 3 stars 4 stars 5 stars
Value
1 star 2 stars 3 stars 4 stars 5 stars
Quality
1 star 2 stars 3 stars 4 stars 5 stars
Nickname
Summary
Review
detailed tests of the new processor based on the Barton core — Ferra.
ru
On February 10, 2003, after a three-month lull, an important event took place in the desktop processor market for PCs, which at the same time became a holiday on the street of fans of AMD processors — this company finally released the new Athlon XP on the long-awaited Barton core with twice the L2 cache. We will consider the marketing background of this step and some details related to the operation of new processors a little later, but now we will focus on the purely technical side — the processors themselves and the study of their performance in a wide range of tasks.
The new Barton core (see Table 1) is manufactured using the same 0.13 micron manufacturing process with copper interconnects and at the same production facilities as the revision B Thoroughbred core. Actually, the difference between these two cores, by and large, is only that Barton added another 256 KB of L2 cache on the edge, due to which the new die stretched in length by about 2.5 mm — from 11.5 to 14 mm (see photo of crystals below and processor above).
At the same time, the die area increased by about 20%, the number of transistors — by 44%, and the total amount of cache memory (L1 + L2) — by 67%.
9000 9000
itself of the Complex of the processor. has not undergone (see pictures), that is, in fact, we have the same Thoroughbred revision B with added memory.
On February 10, the company released three new Athlon XP processors based on the Barton core — models 3000+, 2800+ and 2500+ (see Table 2). All of them are designed for a system bus with a frequency of 333 MHz, have the same supply voltage as the Thorobreds, but less typical power consumption.
Table 2.
Basic characteristics of modern models of AMD Athlon XP processors (the top three lines are the Barton core, the rest are the Thoroughbred-B core).
| Model of the AMD Athlon XP | Frequencies of the nucleus, MHz | Consistent Consum, KBIT | The maximum heat power power, W | |||
| 3000+ | 2167/333 | 128+ 512 | 1.65 | 58.4 | 74.3 | 2800+ 2800+ 2800+ 2800+ 2800+ 2800+ 2800+ 2800+ 2800+ 2800+ 2800H0022 | 2083/ 7 | 68.3 |
| 2700+ | 2167/333 | 128+ 256 | 1,65 | 62.0 | 68.3 | |
| 2600+ | 1.65 | 62.0 | 68. 3 |
|||
| 2600+ | 2133/266 | 128+ 256 | 1.65 | 62.0 | 68.3 | |
| 2400+ | 2000/266 | 128+ 256+ 256+ 256+ 256+ 256+ 256+ 256+ 256+ 256+ 256+ 256+ 256+ 256+ 256+ 256+ 256+ | 62.0 | 68.3 |
The larger cache allowed AMD marketers to move their processors to a new benchmark in the performance rating reflected by the model number. But if earlier this rating could be compared to the processor clock frequency according to a certain mathematical formula (within at least several models), then with the advent of the «Burtons» everything was completely mixed up, and the number of each model began to be based purely on the arbitrariness of the developers (it is alleged that this rating reflects the average performance of processors in real applications relative to a certain base level, which is taken either by the «ancient» Athlon 1400 processor, or you yourself guess that J).
Below we will try to figure out how adequately this model number (derived from the AMD test package, see figure) reflects the real performance of new processors. Indeed, the 3000+ processor has the same frequency as the old 2700+, and the new 2800+ is the same as the Thorobread 2600+ on a 333 MHz bus (note that the Athlon XP 2800+ announced in October with a frequency of 2250 MHz , see www.ferra.ru/online/system/21660, will not go into mass production, and Barton with a lower clock frequency will now act as a mass model 2800+). That is, an increase of 200 or 300 rating points should be provided only by increasing the L2 cache (by the way, based on the frequency of 2500+ and 2800+ models, it would be more logical to assume a rating of 29 for the older model00+, but, apparently, the round number was too attractive). Moreover, the old 2800+ model with a frequency of 2250 MHz in a number of tests, where 384 KB of cache will be enough, can be noticeably faster than the new 3000+ model with a lower clock frequency (and even more so — the new 2800+ with a frequency almost 200 MHz lower ).
One parallel clearly suggests itself here — just over a year ago, Intel released the Northwood core for its Pentium 4 processors (see www.ferra.ru/online/system/15067). One of the main differences of this core compared to the previous Willamette core was the doubled (from 256 to 512 KB) size of the L2 cache memory with the computing core practically unchanged. That is, the situation is very similar to the current one. Then this step allowed processors to significantly increase in speed at the same frequency precisely due to a more capacious cache, and the increase «due to the cache» averaged 5-10% (see www.ferra.ru/online/system/15484). On the other hand, the Athlon XP processors have an exclusive cache, that is, data in L1 memory does not duplicate data in L2, and the actual total amount of cache memory has now increased not twice (as in Pentium 4 a year ago), but by two-thirds — from 384 up to 640 kb. By itself, 384 kB of memory was already large enough to effectively cache many «everyday» tasks, so increasing it to 640 kB may not provide as much performance gain in many applications as the Pentium 4 did.
To get to the bottom of this, we did a comprehensive comparison test of several AMD Athlon XP processors, including the 2700+ and 3000+ models with the same clock speed but different cache sizes. And for the «rating» comparison, we took Intel Pentium 4 processors 3.06 GHz and 2.80 GHz on platforms with different memory. We tested AMD Athlon XP processors on the fastest platform at the moment — on the NVIDIA nForce2 chipset (ASUS A7N8X Deluxe board, see www.ferra.ru/online/system/21270, with BIOS version 1002) with dual-channel DDR333 memory operating synchronously with system bus timings 2-2-2 (see www.ferra.ru/online/system/20800). As platforms for the Pentium 4 MF, we chose several of the most characteristic and high-performance ones from our recent review (see www.ferra.ru/online/system/22900): on the now most common Intel 845PE chipset (ASUS P4PE board) with DDR333 memory, on the progressive SiS648 (SiS board) with DDR400 memory, on the latest E7205 (ASUS P4G8X board) with synchronous dual-channel DDR266 (so far this is the fastest platform for Pentium 4) and, finally, on the well-deserved Intel 850E (ASUS P4T533-C board) with lightning-fast PC1066 RDRAM.
You can find a description of all these boards and configurations, for example, at www.ferra.ru/online/system/23586.
Otherwise, the platform configurations were identical: ASUS V8460 Ultra video accelerator (AGP 4X bus), 80 GB IBM Deskstar 120GXP hard drive (courtesy of the www.arkanoid.ru online store) and 512 MB system memory (for high-quality memory) Kingston ValueRAM PC3200, see www.ferra.ru/online/system/21352, we thank AK-Center Microsystems). 16-bit PC1066 modules manufactured by Samsung and 32-bit RIMM4200 modules from the same manufacturer (kindly provided by Nix) were used as RDRAM memory. In our tests, DDR memory worked in 2-2-2 timing (where it was possible), for RDRAM the «Performance mode» mode was activated in the BIOS Setup of ASUS boards. For Intel Pentium 4 3.06 GHz, Hyper-Threading technology has been activated.
For multilateral testing of new AMD processors, we used our standard set of tests and applications, described, for example, in articles at www. ferra.ru/online/system/22900 and www.ferra.ru/online/system/20800. All tests were run under the Microsoft Windows XP Professional Service Pack 1 operating system. A fresh copy of the operating system was installed on a blank system partition (8 GB) for each platform. For Nvidia microcircuits, the official (WHQL) Detonator 40.72 drivers were used (since we are only talking about comparing chipsets with each other, the use of a video driver that is not the most recent version is not fundamental, the compatibility of the accumulated results is much more important) and a set of drivers version 2.0 (WHQL) for the chipset nForce2. Between blocks of tests, the computer was rebooted. The tests were repeated several times and averaged. The test results are presented in the diagrams.
But before proceeding to discuss them, one more important point should be noted. In the presentation dedicated to the release of the new Athlon XP processors based on the Barton core, AMD demonstrates several slides where the Athlon XP 3000+ outperforms the competitor Intel Pentium 4 3. 06 GHz. According to AMD, their new brainchild is faster than its rival by an average of 15-20% in office and professional applications, as well as in numerous games (the average superiority according to AMD is 17%). Something told me that not everything is clean here: indeed, it turned out that for testing the Pentium 4, the company used far from the fastest platform — on the Intel board on the 845GE chipset with single-channel DDR333 (Intel boards have always been slow, and this chipset now not the fastest, see www.ferra.ru/online/system/22900), while the fastest platform was chosen for its processor — the ASUS A7N8X (nForce2) board with dual-channel DDR333. In addition, AMD in this comparison stuck to its traditional and rather outdated set of tests (see above), many of which are poorly adapted for the Pentium 4. For the sake of objectivity, we will use a more recent, diverse and voluminous set of tests and applications, and for the Pentium 4, we will choose not only the mainstream platform based on the Intel 845PE chipset with DDR333, but also the fastest platforms at the moment with dual-channel DDR266 and RDRAM PC1066.
Let’s see what happens as a result of such a more correct comparison of processors.
Since the SYSmark 2002/2001 tests had a lot of complaints, in order to assess the «office» performance of processors and platforms, we will limit ourselves to PC Magazine Business Winstone 2002 version 1.0.1, which has not been stained yet. Here, AMD processors are certainly in favor, and Burton’s large cache allows you to get a net increase of more than 7% and bypass the fastest Pentium 4 by more than 5%. However, the «office» speed of all these processors is so high that an extra 7-9% gap between Pentium 4 2.80 and Athlon XP 3000+ are unlikely to be useful to anyone J. But in the new PC Magazine Multimedia Content Creation Winstone 2003 (with patch wm312125.exe for Windows Media Encoder 7.1), which comprehensively assesses the «professional» performance of platforms when working with video, audio and graphics, Intel processors take their toll, and Athlon XP 3000+ lags even behind the higher-frequency Athlon XP 2800+ on the Thoroughbred core (for definiteness, we designate it everywhere in this article as 2800T, so as not to be confused with the serial 2800+ on the Barton core) . The increase from Burton’s cache in this test barely exceeded one and a half percent — sparsely!
Another comprehensive test, PassMark PerformanceTest 4.0, demonstrates the complete superiority of Intel processors, and the impact of Barton’s cache is limited to 2.8%. But in the synthetic test of simple processor calculations CPUmark 99, which still works well for comparing modern processors, Athlones make up for lost time, although the 3000+ model turned out to be at the level of the 2800+(T) model, and the larger cache gave only 3% gain.
Archiving in WinRAR 3.00 demonstrates a large speed spread between platforms that are similar in many other tests — more than 22%, and here Barton clearly gives an increase (7.5%), although he is not able to catch up with the fastest platforms with Pentium 4.
On the other hand, ScienceMark 2.0, a complex test of modern scientific calculations, clearly favors AMD processors, which lead by a wide margin, although Barton 3000+ is sometimes inferior to the higher-frequency Thorobread 2800+.
To summarize this block of computational tests, we geometrically averaged the performance of all platforms, taking Pentium 4 2.80 on 845PE with DDR333 as a unit. Due to the ScienceMark 2.0 block, AMD processors outperformed older Intel systems, but the gap between the 3000+ model and its predecessors is not so great and corresponds more to a rating of 2900+.
Let’s move on to the block of video, audio and graphics encoding tests. JPEG transcoding by the popular ACDSee 5.0 program reveals the superiority of Intel solutions, while the Athlon XP 3000+ is almost as slow as its single-frequency Thorobread 2700+ (and noticeably inferior to the 2800T). Audio encoding is a little better for Athlones, although the «three-thousander» here clearly does not correspond to its rating.
We see a similar picture with complex video encoding to MPEG4 format (with simultaneous encoding of an audio track to MP3 format): in FlasK 0.78, the efficient operation of Hyper-Threading on the older Pentium 4 allows it to noticeably break away from all AMD processors (here 3000+ is again slower 2800T, and the increase in speed due to the larger cache on streaming tasks is negligible), and when transcoding MPEG2 in VirtualDub 1. 4. + reaches 22%).
Video encoding in the new Windows Media Encoder 9, which has all sorts of optimizations, only confirms the already seen picture — in streaming tasks, AMD processors are clearly too tough to compete with Pentium 4, and in many respects — thanks to the successful work of Hyper-Threading technology! Again, the growth of «Burton-Thorobread» is minimal, and often the «three-thousander» is slower than Thorobread by 2800+.
The (geometrically) averaged system performance index when encoding draws a bold line: the Athlon XP 3000+ turned out to be slightly worse than the Thorobread 2800+ (the speed increase from a larger cache in these tasks is only 1%), and the Athlons could come closer to the older Pentium 4, if not for the excellent working Hyper-Threading, which provides the latter with an average of about 20% performance increase.
Now some tests in professional 3D modeling packages. The final rendering of frames in Maxon Cinema4D brings us back to the conclusions of the previous paragraph: hyper-threading makes the Pentium 4 invulnerable to Athlons, and the «3000» Athlon XP obviously does not justify its rating, working at the speed of the 2700+ model.
For the 3ds max 5.0 package, we chose (based on a detailed analysis, see www.ferra.ru/online/system/21665 ) one test from each category (animation playback in the projection window, final rendering of one complex frame and final rendering of the video), which would most adequately reflect the average situation among the platforms in each of the tasks. Burton’s larger cache adds 2% animation speed to him and allows him to overcome Thorobread 2800+, but does not allow him to reach even older Pentium 4 2.80 GHz systems, let alone HT systems.
When rendering one frame, the “three thousandth” Athlone again “disgraced” his rating, losing to the 2800th. And the lack of an analogue of Hyper-Theading does not leave AMD processors any chance to compete with the three-GHz Pentium 4. But when rendering a video, where, as we remember, HT gives up slack (and even reduces overall performance), older Athlones have a chance to compete for prizes, although Barton again does not shine with his big cash.
The final rating for modern CAD-work is again disappointing neither for Athlon XP, nor for the Barton core: the «3000» noticeably loses to the higher-frequency Thorobread 2800+, and Hyper-Threading shields Intel processors like armor from rival pursuers, providing them with an average about 10% increase in speed.
Let’s move on to games. First — games under DirectX. Here, finally, Barton has a chance to take his own: in the comprehensive 3Dmark 2001SE test, he even outperformed Pentium 4 3.06 on the 845PE/DDR333 platform, although he lost to Intel platforms with faster memory. The increase from the larger cache is not large (2%), but it already allows it to outperform the 2800+(T) model and partly “justify” its three thousandth J rating.
Comanch 4, although the Intel microarchitecture is clearly more profitable here. But in the complex realistic 3D graphics of the Codecreatures Benchmark Pro test (200,000-500,000 polygon scenes), on the contrary, the powerful Athlone coprocessor is more profitable.
Based on a detailed analysis of the Unreal Tournament 2003 test (see www.ferra.ru/online/system/21634 ), we chose the three most typical botmatch scenes: in one (antalus) Athlone «hegemonizes» and gain for Barton reaches 5%, in the other (asbestos) — on the contrary, Intel processors are in favor (Barton provides up to 6.5% speed increase), and in the third (anubis), reflecting the average picture for UT2003 as a whole, Barton again gives +6, 5% and more than justifies the rating of 3000+.
On average, according to tests under DirectX, we see a three percent gain from switching to the new Athlon XP core, the senior Athlon was finally able to confirm his claims for a «round» rating, but the Thorobreads 2700+ and 2800+ look clearly better than «their» rating (if it is counted about the competitor’s processors).
Now — 3D game tests under OpenGL. If Quake III doesn’t leave any chances for Athlones, then RTCW on the same engine doesn’t, because the 3000er managed to get ahead of P4 3. 06 on 845PE/DDR333 (but lagged behind older Intel systems). Again, it can be noted that the 3000+ model is barely faster than the 2800+(T) and 2700+ models, which is clearly not enough to justify such a high rating.
In the alpha version of DOOM III, Athlones feel more confident, as in Serious Sam SE — at the level of the oldest Pentium 4 3.06. Here, even the 2700+ model can be assigned the number «3000» if desired, although the larger Burton cache provides minimal benefits.
But in the DroneZ test with a modest image quality, it is possible to achieve a fairly large increase in speed for Barton — up to 11%. However, this exception only confirms the rule noted many times before.
For example, for the Vulpine GLmark test, even at a low resolution, a 640-kilobyte cache adds less than 3% to the performance of an nForce2 system (although this increase may be higher on slower chipsets and memory), which is also confirmed by the average index for gaming OpedGL applications . Here, the 3000+ is up to par again, although it cannot outperform the fastest Intel platforms.
As a conclusion, we will give the final average performance index of all these systems based on all our numerous tests from this review. For clarity, we brought the geometric average to the level of 2800 «parrots» for the Pentium 4 2.80 platform on 845PE/DDR333 (as the most typical one at the moment). Firstly, it can be noted that the Thoroughbred 2700+ looks clearly better than its «rating» and confidently pulls on 2800 «parrots», and even with a colorful tail. The brand new Barton 3000+ is a little short of the declared three thousand, and the price is more likely to be somewhere around 2900 «parrots», and he was ahead of the higher-frequency Thorobread by only a miserable third of a percent, and the average speed gain from increasing the size of the cache was only 3%. Systems based on Pentium 4 3.06 fully justify their «poogers», and they do it mainly due to the excellent work of Hyper-Threading technology in streaming coding and in CAD applications, although without HT their success would be at the level of the oldest Athlon XP.