Recently, we took an interesting look at how much progress Intel CPUs have made over the past six years, testing 5th generation cores versus 10th generation core processors. We covered two aspects: first, we isolated the gains made by increasing the operating frequency, and second, what they achieved by adding more cores. This gave us a clear picture of the architectural refinements that Intel has achieved with its 14nm process.
Of course, they have also achieved much higher operating frequencies over the years while expanding the core number at the same time. For this reason, we also looked at how increasing it to 6, 8, and even 10 cores improved gaming performance, although we ignored the frequency gains as this has been well researched at this point and the frequency scaling is more predictable. Ultimately, we found that Intel's development of the 14nm process brought limited improvements from 5th to 10th generation, representing a ~ 10% increase in gaming performance when comparing CPUs with the same core count and operating frequency, and this gain includes the transition from DDR3 to DDR4.
By far the biggest performance gains for Intel were made by increasing the L3 cache capacity, pushing the operating frequency close to its limit, and of course adding more cores. Essentially, Intel could have created its 10th generation 2020 range with Skylake five years earlier, but in the absence of competitive pressure, they didn't see the need, so they moved in smaller 122mm2 dies.
Speaking of competition: Today's test is all about AMD's own progress with the same test methodology. We're comparing flagship parts of each generation with only 4 cores activated at a fixed frequency, along with some parts of the current generation that run with all cores activated.
Without even looking at new numbers, we acknowledge that this could put AMD in a more positive light as we need to highlight the massive gains the company has made since the FX days and now on the latest generation of Ryzen Has. The fact is, however, that the company's CPUs lagged far behind Intel as AMD was largely irrelevant during the FX era.
So let's talk about what exactly we're looking at here …
We go back to 2012 when AMD released the FX-8350 based on the Piledriver core with the problematic 32nm process from GlobalFoundries. Tech-wise, the first FX-series processors arrived a year earlier and were codenamed Bulldozer, but we're going to skip the FX-8150 and the entire series, we really only need one FX processor and can just as easily do with the "best" go, "and I like to use this term loosely when talking about FX processors.
There was also the FX-8370, which came out two years later and had the same CPU, as well as the super-dumb FX-9370 and 9590, which were essentially the same parts again, but with an increased TDP of 220 watts or nearly 80% higher than the 8350 (!). These were just overclocked models, and in the case of the FX-9370 and FX-9590, the base and boost frequencies weren't even improved by 20%.
Since we are testing all parts at 4.2 GHz, we decided on the FX-8350 and blocked it at this frequency. We also test with only 4 cores activated, but since the FX series only lasted up to 4 cores with 8 threads, we didn't have to make any changes here. Some may still argue that the FX-8350 is an 8-core CPU, but it isn't, and it certainly doesn't comply with the California Consumers Legal Remedies Act. In 2015, AMD was successfully sued for allegedly misrepresenting the specifications of Bulldozer chips, and in August 2019, AMD agreed to settle the lawsuit for $ 12.1 million.
The latest BIOS revisions to the motherboard still claim that the FX-8350 is an 8-core processor while Windows states that it is a 4-core CPU with 8 logical processors, so basically a quad -Core with SMT support. We use the Asus M5A99FX Pro R2.0 with DDR3-2400 CL11-13-13-31 memory to test the FX-8350.
Then we have the Ryzen 7 1800X, 2700X, 3800X and 5800X for comparison. With the exception of the 1800X, we have limited all CPUs to 4 cores with 4.2 GHz. The 1800X was still running with only 4 cores activated, but they were clocked at 4.1 GHz as that was the highest stable frequency I could get with this part.
We also included the Ryzen 5 5600X and Ryzen 9 5950X for 6- to 16-core comparisons and these can be compared directly to the FX-8350, although they have a clock penalty as they are underclocked at just 4.2GHz.
The Ryzen 7 2700X up to the 5000 series were tested on the Gigabyte X570S Aorus Master, while the 1800X was tested on the MSI B450 Tomahawk Max.
All Ryzen CPUs were paired with DDR4-3200 CL14 dual-rank dual-channel memory, with all primary, secondary, and tertiary timings configured manually. Finally, all CPU configurations were tested with the Radeon RX 6900 XT. Let's look at the results …
Starting with Rainbow Six Siege, we see that the FX-8350 averaged good at 211 fps, which is obviously a lot of performance and this Vulkan title really doesn't require a lot of CPU power. That said, it was still miles slower than even the first-gen Ryzen part, with the 1800X being nearly 40% more powerful, and keep in mind that only half of the Ryzen CPU is enabled, despite being heavily overclocked.
The 2700X was just a small upgrade from the 1800X, here we see a 9% increase in performance. Then Zen 2 is known to provide a significant performance boost from Zen +, and in this example we're looking at a 23% increase in frame rate, which brings the Radeon 6900 XT to 390 fps, an incredible 84% increase over the FX-8350.
Then we see the biggest step forward with the introduction of Zen 3 and here the 5800X was 135% faster than the FX-8350 with half the cores enabled and this underscores how bad the FX series was and the performance of the 6900 XT cut back by more than half.
For those wondering, the 5800X was tested in a 4 + 0 configuration as AMD switched to an 8-core CCX with Zen 3, while Zen 2, Zen +, and Zen only had 4-cores per CCX. Finally, when testing with Rainbow Six Siege, it's interesting to see only a 9% increase in performance when increasing the Zen 3 core count from 4 to 6, and then only seeing a 10% increase in performance when moving from 4 to 16 cores switch.
Obviously this title does not use the CPU too much and therefore the FX-8350 is able to deliver a very playable performance.
The experience of playing Assassin's Creed Valhalla is very different, at least with the FX-8350. While it stayed very fluid in Rainbow Six Siege, constant stuttering was very noticeable when playing Valhalla with a 1% low result of only 39 fps. The 1800X increased 1% low performance by almost 60% and this made for a significantly better gaming experience.
The 2700X was much faster again, increasing the 1% low power by another 21% and then with the 3800X we were very close to pushing the 6900 XT to its limit. In fact, the 4-core 5800X wasn't faster, while the full-fledged 5950X only improved the frame rate by 7%.
Battlefield V is a great example of how bad the AMD FX series was, and of course still is. It seems the more you use the FX processors, the more they fall apart. Low-level APIs like Vulkan, which help bypass the CPU resulting in significantly less CPU overhead, often work well with severely underpowered CPUs like the FX-8350, and we just saw that in Rainbow Six Siege.
But when you play Battlefield V with the preferred DX11 API for that title, the FX-8350 crumbles. With only 47 fps on average and 29 fps 1% low, the game was completely unplayable by my standards and completely useless even for semi-competitive gameplay. Remember, we're using a Radeon RX 6900 XT for this review.
The Ryzen 7 1800X, on the other hand, was 160% faster, which is a performance increase of 160% from one generation to the next for AMD, a failed generation to a reasonably successful generation. AMD was able to increase the performance with the Zen + update by a further 11% and with Zen 2 by 16%, since the 3800X allowed an average of 158 fps.
Then with Zen 3 we see a further increase in performance from 21% to 191 fps, an increase of 306% over the FX-8350.
Based on the 4-core configuration with the 5800X, the performance in this title was increased by 10% with 6-cores and by 27% with 16-cores. That means the Ryzen 9 5950X is 415% more powerful than the FX-8350, wow.
F1 2020 is not the most demanding title and also supports DirectX 12. Here the FX-8350 does reasonably well and certainly enables a pleasant and very playable gaming experience. Nevertheless, we still see a performance increase of 46% with the 1800X, in which again only half of the cores are activated in order to enable a more apple-to-apple comparison of the architectures.
Again, AMD took a small step with Ryzen of the 2nd generation, here the 2700X offered a performance increase of 9% and then with the 3rd generation we see a further increase of 15%. When we see a pretty substantial increase of 24% from the 3800X to the 5800X, it hits 266 fps.
If we then enable more cores for the Zen 3 architecture, we see an increase of 14-15% from just 4 cores to 16 cores. In this example, the 5950X was 162% faster than the FX-8350, and we're seeing a similar range for the 6-core version.
Hitman 2 is much more CPU intensive than F1 2020 and this causes problems with the FX-8350. The game was more playable than Battlefield, but the experience with a 6900 XT wasn't ideal to say the least. The 1800X offered a 67% increase in average frame rate and an 82% improvement in 1% low performance, making the 1st generation Ryzen processor miles faster.
We see similar performance improvements across all Ryzen generations, and by the time we hit Zen 3, AMD improved the FX-8350's performance by 147% when doing a normalized core and frequency comparison.
From the 4-core 5800X configuration, the 6-core 5600X was 10% faster and the 16-core 5950X 21% faster. We expect about a 200% increase in performance from the FX-8350 to the 5950X, with both CPUs running at 4.2 GHz.
Horizon Zero Dawn plays well on the FX-8350, and what we have here is a game with a low-level API that again isn't particularly CPU-heavy. Even so, the 1800X offered 35% more performance with half of its cores activated, and when we reach the 5800X, which still only has half of its cores activated, we see a 105% increase in performance over the FX processor.
Interestingly, adding more cores doesn't do much for this game. So if there was a quad core Zen 3 CPU it would be more than capable of getting the most out of the 6900 XT in this game.
Cyberpunk 2077 is another game that uses a low level API, but this game is very CPU demanding and as a result the FX-8350 tank gives a 1% low score of only 32 fps and let me tell you, im Compared to all other CPUs tested, the experience with the FX processor was terrible.
Even the 1800X with half its cores was able to improve 1% low performance by 94%, an incredible increase in performance. Then, from 1800X to 5800X, AMD was able to improve its core performance by 53% if you look at the average frame rate. With 2 more activated cores, the 1% low performance is improved the most, from 86 fps to 101 fps, so a nice performance increase of 17%.
The 5600X and 5950X delivered virtually the same level of performance in this title, and that meant the 5600X was 216% faster than the FX-8350 when comparing 1% lower performance.
Shadow of the Tomb Raider is another DX12 title that is very CPU heavy. For testing, we don't use the built-in benchmark, as it's more of a GPU test, but test in a large village, which increases the CPU load dramatically.
Here the FX-8350 delivers playable performance depending on the standard. For a single player game I'd say it's playable but not desirable considering we're using a very fast GPU like the Radeon RX 6900 XT. This meant the Ryzen 7 1800X was able to improve 1% low performance by 51%, a massive generational improvement made possible by the fact that the FX series was a dumpster fire that took AMD 6 years to complete to put it out, and even then the stench hung around for a year or so.
AMD was really gone with the arrival of Zen 2, which increased performance by almost 30% over the 1800X for the same core and clock frequency, which it never did as TSMC's 7nm process clocked much better. When Zen 3 was released, AMD was 126% more powerful than the FX-8350 when it came to core count and clock speed.
But with 6 cores activated we see another 30% increase and then a 38% increase with the 5950X. So the 5950X is 213% faster than the FX-8350 when you compare the average frame rate and 250% faster when you compare the 1% low result. So yeah, it's a little faster.
Last we have Watch Dogs Legion, which is also CPU-heavy and has brought the FX-8350 so far that it was not quite playable with a 1% low of only 28 fps. The 1800X with half of its activated cores stormed forward with an average of 76 fps, which corresponds to an increase in performance of 81%.
After that we see a massive increase in performance of 126% for the 3800X and 152% for the 5800X, again with only 4 cores activated. The jump from Quad Zen 3 cores to 6 brings us an improvement of 18% if you look at the average frame rate, or 29% for the 1% low. The jump from 6 to 16 cores doesn't do much in this title, and we are a long way off from this kind of processing power being a requirement for games.
As expected from AMD's FX series dark days, the company made serious performance improvements with Ryzen and eventually managed to overtake Intel with Zen 3, although Intel is expected to hit back with Alder Lake (12th generation Core series) soon.
Before we continue with this discussion, let's take a look at the average performance of the 9 games tested.
On average, the FX-8350 was good for 70 fps with a 1% low of 47 fps. In comparison, the 1800X's Zen architecture, with half the cores disabled but a fairly significant frequency boost, improved the clock performance by a whopping 60%, or 70% if we look at the 1% low results. That's a mega-generation performance improvement and while the 1800X wasn't clocked at 4.1 GHz and only the best silicon reached that frequency, it remains a massive architectural improvement when adjusted at the same clock speed.
Even if we turned the 1800X down to 3.6 GHz, which is a frequency reduction of 12%, you would only expect a performance drop of about 10%. Either way you cut it, 1st Gen Ryzen was a monumental upgrade for AMD.
Then we see another 10% increase from Zen to Zen +. This time the frequency of 4.2 GHz was more realistic from the point of view of the out-of-the-box product. AMD achieved another 15% increase with Zen 2 and 17% more with Zen 3.
An impressive run
It's crazy to see how AMD has made progress over the past decade and how much it had to catch up to finally knock on Intel's door with 3rd generation Ryzen. At that time, Intel demolished Coffee Lake, which architecturally was only Skylake 2 years earlier.
As early as 2016, a year before the first Ryzen processors were released, AMD relied on parts like the FX-8350 to fight Skylake; With CPUs like the 4C / 4T 6600K and 4C / 8T 6700K, even with the Core i3 dual cores of the time, AMD struggled to keep up, as most games only needed an Intel 2C / 4T CPU.
Roughly speaking, the i7-6700K corresponds to the 3800X with the same frequency and number of cores. That said, there was a time when Intel processors would offer ~ 100% more performance when you removed the GPU limits. Because of this, the FX-8350, which was realistically the flagship of the series, was cheaper than Intel's fastest Core i5 and still a terrible buy at the price.
Since then, AMD has made serious strides with Ryzen, increasing performance by ~ 50% from the 1000 series to the 5000 series at the same frequency. Since Zen 3 is clocked higher and has 12- and 16-core models, the real increase in performance is much greater.
In case you are wondering "Where are the electricity usage numbers?" We prefer not to include them in these "for science" benchmarks because they are terribly inaccurate and no real conclusions can be drawn from them. This is because parts like the 1800X are grossly overclocked and therefore received a significant amount of voltage to achieve stability, while the 5800X was essentially underclocked. Power measurements are the best power in stock so we can do a power-per-watt comparison, and you can find this information in the early reviews for these parts.
Granted, you can't compare Intel's 10% architecture gains over the past six years to AMD's 50% increase over the past four years because they started from two very different starting points and it took a long time to get back to the top .