Amd phenom cpu list: AMD Phenom II X4 Processors List

AMD Phenom and Phenom II x2, x3, x4, x6 Processors: Complete List

Phenom and Phenom II are AMD’s line of 64-bit desktop processors. The processors in this line are based on the K10 architecture, which AMD refers to as the 10h processor family. In this article, we will talk about the Phenom processor line, as well as provide a complete list of processors that have been released since the existence of this line.

Table of Contents

Phenom 9 at a Glance0007

The Phenom line appeared in 2007 and existed until 2011. The first Phenom processors were released for socket AM2+, later Phenom II appeared, which were released mainly for AM3 and AM3+.

When the first four core models of this line were released, AMD emphasized that only these processors can be considered truly quad-core, since then the Intel Core 2 Quad series is a multi-chip assembly, and in the case of Phenom, all four cores are on a single silicon chip .

The initial launch of the Phenom processors was not without noticeable problems. Just before the release of these processors, a serious bug was discovered in them, which could potentially lead to a system lock. To resolve this issue, a BIOS update was released that programmatically disabled the failed module. But, such a solution led to a decrease in performance by 10% or more. And due to the fact that this BIOS did not appear immediately, buyers received slower processors than those provided to the specialized media before the release. In later releases of Phenom processors (beginning with the B3 stepping), this error was fixed by hardware.

The early Phenom processor architecture included a 128-bit DDR2 memory controller at up to 1066 MHz (DDR3 support will be implemented later), 128 KB L1 cache per core, 512 KB L2 cache per core core, as well as a 2 MB L3 cache common to all cores.

One of the important features of this architecture was the emergence of new technologies to save energy. For example, Cool’n’Quiet 2.0 technology allows you to reduce the supply voltage, the clock frequency of the entire processor, as well as the clock frequency of individual cores, AMD CoolCore technology allows you to disable inactive core modules, and Dual Dynamic Power Management technology allows you to set different voltages for different processor cores and memory controller.

Phenom X3 processors

Stepping	Frequency	L2 cache	L3 cache	Hyper Transport Frequency	HT base frequency multiplier	Voltage	TDP	Socket
AMD Phenom X3 8250e	B3	1. 9 GHz	3x 512 KB	2 MB	1.6 GHz	9.5x	1.125 — 1.20	65W	AM2+
AMD Phenom X3 8400	B2	2.1GHz	3x 512 KB	2 MB	1.8 GHz	10.5x	1.20 — 1.25	95W	AM2+
AMD Phenom X3 8450	B3	2.1GHz	3x 512 KB	2 MB	1.8 GHz	10.5x	1.20 — 1.25	95W	AM2+
AMD Phenom X3 8450e	B3	2.1GHz	3x 512 KB	2 MB	1.8 GHz	10.5x	1.125 — 1.20	65 W	AM2+
AMD Phenom X3 8550	B3	2.2 GHz	3x 512 KB	2 MB	1.8 GHz	11x	1.20 — 1.25	95W	AM2+
AMD Phenom X3 8600	B2	2. 3GHz	3x 512 KB	2 MB	1.8 GHz	11.5x	1.20 — 1.25	95W	AM2+
AMD Phenom X3 8650	B3	2.3GHz	3x 512 KB	2 MB	1.8 GHz	11.5x	1.20 — 1.25	95W	AM2+
AMD Phenom X3 8750	B3	2.4 GHz	3x 512 KB	2 MB	1.8 GHz	12x	1.20 — 1.25	95W	AM2+
AMD Phenom X3 8750 Black Edition	B3	2.4 GHz	3x 512 KB	2 MB	1.8 GHz	12x	1.20 — 1.25	95 W	AM2+
AMD Phenom X3 8850	B3	2.5 GHz	3x 512 KB	2 MB	1.8 GHz	12.5x	1.20 — 1.25	95W	AM2+

Phenom X4 processors

Phenom II X2 Processors Processor name Stepping Frequency L2 cache L3 cache Hyper Transport Frequency HT base frequency multiplier Voltage TDP Socket AMD Phenom II X2 545 C2 3.0 GHz 2x 512 KB 6 MB 2 GHz 15x 0. 875 — 1.425 80W AM3 AMD Phenom II X2 545 C3 3.0 GHz 2x 512 KB 6 MB 2 GHz 15x 0.900 — 1.40 80 W AM3 AMD Phenom II X2 550 C3 3.1 GHz 2x 512 KB 6 MB 2 GHz 15.5x 0.875 — 1.40 80 W AM3 AMD Phenom II X2 550 Black Edition C2 3.1 GHz 2x 512 KB 6 MB 2 GHz 15.5x 0.850 — 1.425 80 W AM3 AMD Phenom II X2 555 Black Edition C3 3.2 GHz 2x 512 KB 6 MB 2 GHz 16x 0.875 — 1.40 80 W AM3 AMD Phenom II X2 560 Black Edition C3 3.3GHz 2x 512 KB 6 MB 2 GHz 16. 5x 0.875 — 1.40 80 W AM3 AMD Phenom II X2 565 Black Edition C3 3.4GHz 2x 512 KB 6 MB 2 GHz 17x 0.875 — 1.40 80 W AM3

Phenom II X3 processors

Processor name	Stepping	Frequency	L2 cache	L3 cache	Hyper Transport Frequency	HT base frequency multiplier	Voltage	TDP	Socket
AMD Phenom II X3 700e	C2	2.4 GHz	3x 512 KB	6 MB	2 GHz	12x	0.825 — 1.25	65W	AM3
AMD Phenom II X3 705e	C2	2. 5 GHz	3x 512 KB	6 MB	2 GHz	12.5x	0.800 — 1.25	65 W	AM3
AMD Phenom II X3 710	C2	2.6 GHz	3x 512 KB	6 MB	2 GHz	13x	0.875 — 1.425	95W	AM3
AMD Phenom II X3 715	C2	2.8 GHz	3x 512 KB	6 MB	2 GHz	14x	0.875 — 1.425	95W	AM2+
AMD Phenom II X3 720	C2	2.8 GHz	3x 512 KB	6 MB	2 GHz	14x	0.875 — 1.425	95W	AM3
AMD Phenom II X3 720 Black Edition	C2	2.8 GHz	3x 512 KB	6 MB	2 GHz	14x	0.850 — 1.425	95W	AM3
AMD Phenom II X3 740 Black Edition	C2	3. 0 GHz	3x 512 KB	6 MB	2 GHz	15x	0.850 — 1.425	95W	AM3

Phenom II X4 9 processors0007

Processor name	Stepping	Frequency	L2 cache	L3 cache	Hyper Transport Frequency	HT base frequency multiplier	Voltage	TDP	Socket
AMD Phenom II X4 805	C2	2.5 GHz	4x 512 KB	4 MB	2 GHz	12.5x	0.875 — 1.425	95W	AM3
AMD Phenom II X4 810	C2	2.6 GHz	4x 512 KB	4 MB	2 GHz	13x	0.875 — 1.425	95W	AM3
AMD Phenom II X4 820	C2	2. 8 GHz	4x 512 KB	4 MB	2 GHz	14x	0.9 — 1.425	95W	AM3
AMD Phenom II X4 900e	C2	2.4 GHz	4x 512 KB	6 MB	2 GHz	12x	0.850 — 1.250	65 W	AM3
AMD Phenom II X4 905e	C2	2.5 GHz	4x 512 KB	6 MB	2 GHz	12.5x	0.825 — 1.250	65 W	AM3
AMD Phenom II X4 910	C2	2.6 GHz	4x 512 KB	6 MB	2 GHz	13x	0.875 — 1.425	95W	AM3
AMD Phenom II X4 910e	C3	2.6 GHz	4x 512 KB	6 MB	2 GHz	13x	0.850 — 1.250	65 W	AM3
AMD Phenom II X4 920	C2	2. 8 GHz	4x 512 KB	6 MB	1.8 GHz	14x	0.875 — 1.425	125 W	AM2+
AMD Phenom II X4 925	C2	2.8 GHz	4x 512 KB	6 MB	2 GHz	14x	0.850 — 1.425	95W	AM3
AMD Phenom II X4 925	C3	2.8 GHz	4x 512 KB	6 MB	2 GHz	14x	0.9 — 1.400	95W	AM3
AMD Phenom II X4 940 Black Edition	C2	3.0 GHz	4x 512 KB	6 MB	1.8 GHz	15x	0.875 — 1.425	125 W	AM2+
AMD Phenom II X4 945	C2	3.0 GHz	4x 512 KB	6 MB	2 GHz	15x	0.875 — 1.425	125 W	AM3
AMD Phenom II X4 945	C3	3. 0 GHz	4x 512 KB	6 MB	2 GHz	15x	0.850 — 1.400	95W	AM3
AMD Phenom II X4 955	C2	3.2 GHz	4x 512 KB	6 MB	2 GHz	16x	0.850 — 1.400	125 W	AM3
AMD Phenom II X4 955	C3	3.2 GHz	4x 512 KB	6 MB	2 GHz	16x	0.850 — 1.400	125 W	AM3
AMD Phenom II X4 955 Black Edition	C2	3.2 GHz	4x 512 KB	6 MB	2 GHz	16x	0.875 — 1.425	125 W	AM3
AMD Phenom II X4 955 Black Edition	C3	3.2 GHz	4x 512 KB	6 MB	2 GHz	16x	0.825 — 1.425	125 W	AM3
AMD Phenom II X4 965 Black Edition	C2	3. 4GHz	4x 512 KB	6 MB	2 GHz	17x	0.850 — 1.425	140 W	AM3
AMD Phenom II X4 965 Black Edition	C3	3.4GHz	4x 512 KB	6 MB	2 GHz	17x	0.825 — 1.400	125 W	AM3
AMD Phenom II X4 970 Black Edition	C3	3.5 GHz	4x 512 KB	6MB	2 GHz	17.5x	0.825 — 1.400	125 W	AM3
AMD Phenom II X4 975 Black Edition	C3	3.6 GHz	4x 512 KB	6 MB	2 GHz	18x	0.825 — 1.400	125 W	AM3
AMD Phenom II X4 980 Black Edition	C3	3.7GHz	4x 512 KB	6MB	2GHz	18.5x	0.825 — 1.400	125W	AM3

Phenom II X6 processors

Processor name	Stepping	Frequency	L2 cache	L3 cache	Hyper Transport Frequency	HT base frequency multiplier	Voltage	TDP	Socket
AMD Phenom II X6 1035T	E0	2. 6 GHz	6x 512 KB	6MB	2.0 GHz	13x	1.075 — 1.375	95W	AM3
AMD Phenom II X6 1045T	E0	2.7GHz	6x 512 KB	6 MB	2.0 GHz	13.5x	1.075 — 1.375	95W	AM3
AMD Phenom II X6 1055T	E0	2.8 GHz	6x 512 KB	6 MB	2.0 GHz	14x	1.125 — 1.400	125W	AM3
AMD Phenom II X6 1075T	E0	3.0 GHz	6x 512 KB	6 MB	2.0 GHz	15x	1.125 — 1.400	125W	AM3
AMD Phenom II X6 1090T Black Edition	E0	3.2 GHz	6x 512 KB	6 MB	2.0 GHz	16x	1.125 — 1.400	125W	AM3
AMD Phenom II X6 1100T Black Edition	E0	3. 3GHz	6x 512 KB	6 MB	2.0 GHz	16.5x	1.125 — 1.400	125W	AM3

Mobile Phenom II processors 7

Leave a comment

AMD Athlon II X2 260, Phenom II X6 1075T and FX-8350 processors

Time-tested models for the AM3+ platform in a modern software environment

Our recent testing of different processors with a discrete video card opened the way for us to study platforms that do not have integrated graphics: firstly, now we have something to compare the results with, and secondly , we found that it generally makes sense to do this — anyway, in non-gaming applications, the influence of the video card is insignificant. And we decided to start doing this with AM3 +: firstly, no changes have been taking place within its framework for a long time, and secondly, and it looks like it will never happen again. However, the platform continues to be of interest to many users due to the presence of productive (albeit in specific areas) budget processors within its framework, and very many have its representatives on hand. Actually, our testing today will be primarily useful for them, since two of the three processors found at hand generally belong to models for AM3, however, this is suitable for determining the approximate correspondence between this platform and modern devices, and expanding the list of tested processors (if any request from readers) it will not be too difficult for us later.

Test bench configuration

900 58

Processor	AMD Athlon II X2 260	AMD Phenom II X6 1075T	AMD FX-8350
Core name	Regor	Thuban	Vishera
Technology	45 nm	45 nm	32 nm
Core frequency, GHz	3. 2	3.0/3.5	4.0/4.2
Number of cores (modules) / threads	2/2	6/6	4/8
L1 cache (total), I/D, KB	128/128	384/384	256/128
L2 cache, KB	2×1024	6×512	4×2048
L3 cache, MiB	—	6	8
RAM	2×DDR3-1066	2×DDR3-1333	2×DDR3-1866
TDP, W	65	125	125
Average price 900 39	T-6215781	T-6411041	T-8493626

basically an estimate — we just wanted to evaluate what we can expect from the AM3 + qualitatively, so we did not actively search for various processor models, limiting ourselves, as already mentioned above, to three very different ones that literally turned out to be «at hand». In the future, the list of models, if there is interest from readers, can be expanded, since there are still no difficulties with their “extraction”. Especially the newest models of the FX line, which have long been similar to APUs for FM2 / FM2 +, having a larger number of modules and a third-level cache. Today we have only one such model — not the fastest in general (since there are both FX-8370 and «steroid» FX-9 in the line000 with an amazing heat pack, but also with standard frequencies up to 5 GHz), but well and repeatedly studied. As a result, and well suited for assessing what you can generally expect from a four-module FX, and what is not worth it.

Two more processors, much older. However, Phenom II X6 1075T at the time of its release in 2010 could claim high performance, overtaking, for example, all Core i5 of that time in the overall standings. And Athlon II X2 did not aim at such heights, competing with Celeron and Pentium. But the old Celeron and Pentium, which, for obvious reasons, are far behind modern ones. In this regard, we are interested in the Athlon X2 260 — this is a slow dual-core processor almost five years ago, but many people still have such and similar devices in their hands. Therefore, how they look against the background of a more modern range of AMD and Intel processors is very interesting.

9005 8

90 037

As for comparative benchmarks, we didn’t have much choice — even though in most tests (with the exception of gaming, of course) a specific video card does not matter, it’s better to do without rough estimates when it is possible to give accurate ones. Moreover, the set of processors tested with the Radeon R7 260X last time is quite representative, but only the results of the Core i7-479 are redundant.0K. We definitely need the A10-7850K, and fast Pentiums, Core i3s and Core i5s won’t hurt either.

As for other test conditions, they were equal, but not the same: the frequency of RAM operation was the maximum supported according to the specifications for all, except Athlon II X2, which also works smoothly with DDR3-1333, although no one promises this officially. But its volume (8 GB) and system drive (Toshiba THNSNh356GMCT, 256 GB) were the same for all subjects.

Test methodology

To evaluate performance, we used our performance measurement methodology using iXBT Application Benchmark 2015 and iXBT Game Benchmark 2015. We normalized all test results in the first benchmark against the results of the reference system, which this year will be the same for laptops , and for all other computers, which is designed to make it easier for readers to compare and choose:

Processor	Intel Core i5-3317U
Chipset	Intel HM77 Express
Memory	4GB DDR3-16 00 (Dual Channel)
Graphics	Intel HD Graphics 4000
Drive	SSD 128 GB Crucial M4-CT128M4SSD1
Operating system	Windows 8 (64-bit) Intel 9 cores0039	9. 18.10.3186

iXBT Application Benchmark 2015

Applications of this type (actively loading all the processor cores they can find) are the best application for multi-module AMD processors. As a result, the FX-8350 easily outperforms even any Core i5, and, of course, it is twice as fast as AMD’s best solutions for FM2+, where the architecture is the same, but there are half as many modules. The difference with the previous generation is also visible — the FX-8350 is one and a half times faster than the Phenom II X6 1075T. Yes — the last one is not the fastest Phenom II X6, however, the FX-8350 is at least nominally not the fastest FX for a long time.

The difference between the Athlon II X2 260 and the Pentium G3470 is also very indicative. As you can see, the statement “dual-core processors do not cope well with multi-threaded software” is not entirely true, since dual-core processors … are also different. Their performance can easily vary by a couple of times. Of course, with the same architecture, the more cores, the better — the fastest Pentium is twice as slow as the Core i5. But the six-core Phenom II X6 1075T is not three times slower (as one might assume from the number of cores), but a little more than one and a half. And technologies such as SMT (Hyper-Threading or AMD “x86-cores”) also work and affect the overall picture: after all, the same FX-8350 has as many “real” cores as the Core i5 and fewer than Phenom II X6, however, running two threads at the same time each allows you to get a better result.

But in these applications, the benefits of multi-core and multi-threading are less. In addition, the FX-8350 fell into the same trap as the recently Core i7-4790K: in the multi-threaded subtest of Adobe After Effects CC 2014.1.1, it lacked 8 GB of memory, which is why the whole result turned out to be lower than it could would: only at the level of Core i3. However, Phenom II X6 and A10 are equivalent to each other here (which is also very revealing) and both are Pentiums, so the older FX are a step forward anyway. Let already (unlike the previous case) we are not talking about direct competition with the Core i5 and i7, but other AMD models are even worse under such loads.

Working with raster images (private, but the most common representative of which are photographs) in modern software is not a strong point of AMD’s modular architecture. On the other hand, the old architecture of the company’s processors was even worse for this: despite the support for multithreading, six-core Phenoms look only slightly better than Pentiums and worse than two-module APUs, while three-four-module processors are still better. To some extent, even Core i3 competitors, though there are models with a TDP of 35 W, so such a competition is initially losing for FX. That is, the general verdict is not so bad, but not the tasks for which it makes sense to choose these processors.

When a couple of threads are enough and you can generally talk about optimization «for Core 2 Duo», you get a total defeat. However, no one expected anything else, and the Core also doesn’t look the best here, only slightly outperforming Celeron and Pentium.

Audition, as has already been established, can use several computational threads, but at the same time it increases performance rather sluggishly (this, by the way, is a good hint to those who consider total support for multi-core processors a sufficient condition for strong software acceleration — it is not always such) . It is noteworthy that some 9 independent cores are more useful for the program than «threads», which is why the result of Phenom II X6 is relatively good. But in general, for all AMD processors (both architectures) this is not the most “convenient” application — it’s hard to overtake only Pentium, and even Core i3 is already faster.

Multi-threaded integer, where modules work well — «semi-kernels» are not worse than kernels. In any case, the old cores of AMD itself — since then, Intel processors have stepped far ahead, so the two modules are only competitors of two single-threaded Haswell cores, but not spiced with Hyper-Threading. But the competitors of the four modules should already be looked for among the Core i7 (at least low-power ones), but not in the Core i5 company. Considering the prices — the alignment is good.

In archivers, the effect is heavily blurred by single-threaded unpacking. As in real life in general: as you can see, among the application software there are not so many applications that can fully use a large number of at least cores, at least modules, at least some green men walled up with accounts in the processor 🙂

A small «platform» loss, but just that it is small, and as usual, little depends on the processors themselves. Provided that they do not belong to too different classes (tablet and extreme for example), although in this case, with the same drive, the difference will be quite small.

In general and average, it turned out that the FX-8350 is a little better than the modern Core i3, and the Phenom II X6 1075T outperformed only the older Pentium. The result is funny, shocking to some users, but more than understandable. After all, what is the performance of a multi-core processor? The product of the performance of a single thread multiplied by the number of threads used by the software. With the first, everything is clear — the higher, the higher the performance in any programs, since there will always be one stream of computational code. But with the second one… The more computation threads are simultaneously supported by the processor, the greater the number of programs that will not load all of them. The ideal case is dual-core processors with “single-threaded” cores: they are most often loaded at 100%. And right away, by the way, it is clear that the performance of the «Athlon» cores is about one and a half times less than that of the «Hazwell» cores (we get another 10 percent due to the clock frequency). But the more cores there are, the more likely it is that only a part of them will be fully used, since they will turn out to be “superfluous” for a specific application. That is why the extensive path (increasing the number of cores) is no worse than the intensive one (increasing performance per thread) only under ideal conditions. In real — intense is much more preferable.

Another way to intensively increase performance is not to increase the number of cores, but to add computation threads. Under ideal conditions, it is obviously worse than increasing the number of cores, in real conditions it is probably also worse, but much cheaper. Both companies use it now, but in general, we can conclude that AMD’s modular architecture is more efficient than Intel’s Hyper-Threading, but also more costly in terms of transistors. In principle, the company mostly fulfilled its promises, but direct competition still did not work out — precisely because Intel has higher performance per thread. Therefore, dual-module processors are able to lose on average to dual-core processors without HT, while quad-module processors look decent only against the background of Core i5, but not i7. Alas, this is the harsh truth of life.

Gaming applications

For obvious reasons, when using the Radeon R7 260X, we limit ourselves to the minimum quality mode (this video card by itself is not enough for maximum settings), but in full Full HD resolution (with this, it, unlike many integrated solutions, does a great job).

The CPU requirements for this game are minimal and the graphics adapter requirements are high, but we have only one last one today. As a result, performance indicators are almost the same for everyone 🙂

You can’t say the same about «Tanks». Let’s leave aside Intel processors, which are practically unattainable in terms of single-threaded performance (and it is required here), and let’s see what we have «in the group». And in it, the most interesting thing is the approximate parity between the A10-7850K and the Phenom II X6 1075T. And, as it turned out, for the A10 in this game, the discrete only gives a reduction in the frame rate compared to the integrated core, but Phenom II did not have such. That is, such is the progress in a certain class. Nobody doubted that the FX-8350 was faster than this pair, but there were doubts that the Athlon II X2 260 would do the job at all. Still too old and weak solution. And, indeed, the performance is low, but you can play it;)

Due to the fact that the benchmark for this game requires support for AVX instructions, there are no results for half of the subjects. However, they are not so interesting — Grid2 is quite efficient on processors for ultrabooks with integrated graphics, i.e., in general, it works everywhere.

How to illustrate the expression «not enough processor»? But something like this — pay attention to the results of Athlon II X2 260. But it is old and initially not aimed at high results, so based on this you should not expand the statement to «there are not enough dual-core processors» (which some are trying to do) — compare the results with the Pentium G3460, which is about twice as fast. In terms of overall performance outside of games, it is also about twice as fast, which directly hints at the fact that there are “not enough” usually slow processors. Regardless of the number of cores. It’s just that in game engines that support multithreading, ceteris paribus, a multi-core processor is faster, but ceteris paribus.

And here is just an illustration of this case. As we already wrote, Metro 2033 is more demanding on processors than the next games in the series, and the engine supports multi-threading. As a result, the Athlon II X2 260 cannot be played at all, but the Phenom II X6 1075T is about three times faster than it. Ideal scaling is actually three times the cores and three times the performance. But this is when comparing processors of the same architecture. And when they are different, it suddenly turns out that the same X6 1075T cannot catch up with the banal Core i3-4170, and the FX-8350 is capable of this, but it is already lagging behind the Core i5.

In Hitman, this is not the first time that the alignment is already similar to Metro 2033, so this is not some kind of «twist» of a certain game, but an objective reality that manifests itself under a certain set of circumstances.

A10 and Phenom II are barely enough for the game, but they are enough — unlike the old Athlon II. The rest are faster, but as we already know, you won’t be able to get high results here when using this video card.

Amazing unanimity not for the first time — the game has very low CPU requirements, so give it a decent video card and everything will fly.

The game is susceptible to single-threaded performance, so all Intel solutions are on a par, while AMD has inconsistencies due to the fact that processors belong to different architectures and, in general, are designed differently.

Everything is on a par — processor dependence is low. At least in the minimum settings mode. Differences between processors may appear higher, but it makes sense to check this on a faster video card — the requirements for it are quite large, which is noticeable when you try to increase the picture quality.

Total

In principle, we did not expect any discoveries from the FX-8350 — it is architecturally similar to the APU for FM2, but it has «more than everything»: modules, cache memory, frequencies. As a result, when what it has “more” matters (and the frequency, for example, always has it), you get a good increase in performance and the ability to overtake the Core i5. And when it doesn’t work out, it turns out that FX and APU are almost the same, and at the level of older Pentiums just «heavy» loads, the buyer has the opportunity to save a lot by buying a four-module FX, and not a more expensive Intel processor from one of the older families. Looking at prices, this statement can be extended to three-module models, but two-module models have serious competitors in AMD’s assortment, but this, in general, is not news to anyone.

The time of the “old” AMD architecture is gone forever. From which one should not conclude that it’s time to run to change the old Phenom (if they already exist) for something new — as long as it works and meets the requirements, let it work. However, it’s not worth being surprised that as many as six cores, even at full load, can hardly always and with difficulty overtake only two. The architecture is old, but progress does not stand still.

But the results of gaming tests, let’s say, even amused us. First of all, this concerns the Athlon II X2 260, which easily confirmed the thesis — for games, you need a video card first, and then everything else. Yes, of course, the frame rate in his case is lower than when using other subjects, and some games are completely impossible to play even on the Radeon R7 260X (note that the card is inexpensive, but it already lacks such a processor).