AMD vs. Intel: The Evolution of CPU Gaming Efficiency

After looking at the last few generations of Intel and AMD CPUs separately, here is our final part in the gaming performance advancement series. In this article, we're going to directly compare a decade of AMD and Intel CPUs.

For those of you who missed the first few articles, it all started with a look at Intel's 10th Core CPUs. This gave us a clear and very interesting look at how L3 cache capacity affects gaming performance, and we have found that Intel often sees the biggest performance gains in today's games when the L3 cache capacity is increased rather than adding more cores.

By limiting all CPUs to only 4 active cores with 8 threads at the same clock frequency, we got a really good look at things that had never been tested before, and this led to comparisons with older Intel CPUs.

The next step, of course, was to do the same with a number of AMD processors while normalizing the core count and clock speed. This showed how weak the old FX series was and what big leap AMD has made with each Ryzen generation.

This CPU architecture benchmark was an interesting way to see the progress made by both companies. Now all we had to do was test a few generations of CPUs that were missing from previous articles like Sandy Bridge, Ivy Bridge, Haswell, and of course Rocket Lake, and throw all the data together.

For this test, all CPUs were clocked at 4.2 GHz, with the exception of the Ryzen 7 1800X, which only clocked at 4.1 GHz, but that is only 2.5% frequency deviation, which does not meaningfully affect the results. Models with DDR4 memory were paired with DDR4-3200 CL14, and DDR3 models used DD3-2400 CL11 memory. Otherwise everything is as good as possible, with each model running with 4 active cores.

Please note that no cores have been deactivated on the FX-8350. So if you think it's an 8 core CPU, then there were probably 8 cores active. Whatever the case, it's either a slow quad-core processor or a terrible 8-core processor, we'll leave that up to you. We used the Radeon RX 6900 XT during our gaming tests. Let's look at the dates …

Benchmarks

Based on the Rainbow Six Siege results, when we compare the 2012 options – AMD Piledriver and Intel Ivy Bridge architectures – we see that Intel delivered 21% better frame rates and a 29% stronger 1% lower frame rate, which is a significant benefit and as a result, AMD was forced to sell their parts at a much lower price. That really hurt when you consider that the FX dies were almost twice as large with 315 mm2 and were therefore significantly more power-hungry.

From Ivy Bridge to Haswell we're seeing a very sharp 23% increase in performance and I can't remember seeing such large margins in 2013, but of course games weren't as demanding back then and GPUs weren't nearly as high as they were powerful. In fact, it has been extremely difficult to highlight the benefits of SMT technology for gaming.

However, things got bad for AMD in 2013 as they competed with Haswell and Intel was almost 50% ahead, at least in these types of games. Broadwell doesn't offer much over Haswell with its eDRAM, a small 4% gain that is nothing special. In 2015 we also received Intel's Skylake architecture and at that point virtually no one spoke the name "AMD" when it came to CPUs. This is because, all other things being equal, Intel offered a 73% higher gaming performance in RSS.

Things are stagnating here for Intel. From Skylake to Kaby Lake, we're seeing the smallest performance gain in generations. Basically nothing was gained from the 6700K to the 7700K.

As we recently discovered, the main source of Intel’s improved gaming performance since Skylake has been to increase the L3 cache capacity. This starts with the 8700K jumping up from the 7700K's 8MB L3 cache to 12MB, resulting in a slight 4% performance increase in Rainbow Six Siege. Then Intel upgraded to a 16MB L3 cache with the 9900K and although the core number and clock frequency remained the same at 4.2 GHz, the frame rate was increased by a further 5%.

While Intel essentially stuck to Skylake, but with more cores and cache, AMD naturally saw room for its re-emergence. Ryzen was first released in 2017, five years after the FX-8350. The first generation architecture was almost 40% faster compared to the FX series in Rainbow Six Siege.

But as impressive as that win was, AMD was still miles behind Intel in gaming performance. In this example, they roughly caught up with Haswell, where Intel was back in 2013. Ryzen appealed to enthusiasts because it offered more cores and, as such, was a productivity beast. The gaming performance, while not exactly impressive, was decent, and there were other positives that helped make the series a success and a step in the right direction for AMD.

When Intel added two more cores with Coffee Lake, AMD had to push forward, and they did that with Zen +. Here the 2700X offered a decent improvement of 9% – still 17% slower than Intel's 2018 architecture – but the gap was slowly closing. Things got really hot in 2019 when AMD launched Zen 2 and they were no longer hot on Intel, with the 3800X only 5 to 10% behind the 9900K.

With the release of Zen 3 in 2020, the time had finally come for AMD to outperform Intel in gaming and deliver the final fatal blow to the 14nm process. Clock for clock AMD was now 16% faster, and of course Zen 3 clocked very well, so AMD was now mostly faster for games. Intel tried to soften the blow with Rocket Lake, and while some gains were made, Intel took a step back in some cases like what we're seeing here.

That was a long look at Rainbow Six Siege results, so we're going to round up some of the games below …

Assassin's Creed Valhalla provides us with far less interesting data, as it is largely GPU-limited and does not use the CPU well. Intel hit the cap with the launch of Skylake and AMD never managed to catch up, but if we look back on 2012 when Intel was up to 56% faster, that margin is now just 7% in favor of Intel, at least with an adjusted clock for clock with the same core number.

Battlefield V was brutal on the AMD FX-8350. Here it only managed 47 fps on average, while the 3770K was good for 95 fps, a performance increase of 102% over Intel in 2012. If that wasn't bad enough, Intel achieved an increase of 37% with Haswell and jumped on average to 130 fps, almost an increase of 180% over the FX-8350. From Haswell to Broadwell we see a modest 8% increase and from Broadwell to Skylake a 4% increase, while Kaby Lake was nothing new.

After Ryzen arrives, we're looking at a Haswell-like performance, although that's not entirely accurate. While the average frame rates are very similar, Ryzen's 1% lower performance was almost 30% better. AMD then achieved an 11% increase from Zen to Zen +, while Intel increased clock-to-clock performance by 5% by increasing the L3 cache capacity of its Core i7 series.

Again it was Zen 2 that AMD put real pressure on and increased performance by 16%, which was enough to keep Intel competing in this game with the same core count and clock speed.

Intel recorded a performance increase with the 11th generation, which rose to an average of 177 fps, which corresponds to an increase of 10%. Despite this improvement, AMD is still ahead with Zen 3 and offers an average of 8% more performance. This means that Intel has seen a 20% increase in performance as of 2017, while AMD almost tripled that with a 57% improvement from Zen to Zen 3.

F1 2020 doesn't load the CPU as much as Battlefield V, but we still see strong gains in most CPU generations. The 2012 match-up sees Intel able to deliver 46% higher performance with Ivy Bridge compared to Piledriver. Then we see a modest 12% increase from Ivy Bridge to Haswell and this time Broadwell offered a 10% increase over Haswell, allowing Intel to make decent gains on the 14nm process for a few years without adding any more cores.

That stalled with Skylake, and we don't see any real change from the 6700K to the 8700K, which was a problem for Intel when AMD hit back with Ryzen.

Ryzen wasn't exactly that amazing from the start as the 1800X could only compete with the 5 year old 3770K. Zen + brings AMD up to speed with Haswell or something, and Zen 2 brings it up to par with Broadwell and not much behind Skylake to Coffee Lake. A massive push can be seen with Zen 3 to finally take the lead, even against the newer Rocket Lake architecture.

Hitman 2 shows an 11% increase in performance from Sandy Bridge to Ivy Bridge, which we don't see often, though I wonder how much of that is due to the difference in PCI Express bandwidth since we're comparing PCIe 2.0 to 3.0. In any case, it means that in 2012 the Intel architecture was 33% more powerful in that title.

Then Intel managed to improve performance by a whopping 38% from 2012 to 2013, and that meant their lead over AMD had faded to 84%. As we've seen several times now, Intel has made reasonable performance gains from Haswell to Broadwell to Skylake, and that's where things slowed down significantly for the Blue Team. Only with the 11th generation did Intel achieve a solid increase in performance and in this example it was enough to beat AMD's Zen 3.

But long before the arrival of Zen 3, AMD had to make do with first-generation Zen, and here they were still slower than Intel's 5-year-old Haswell architecture. Zen + put them on par with Haswell and Zen 2 got them to Skylake, which meant they were competitive with Coffee Lake.

The performance of Horizon Zero Dawn on Intel CPUs is very similar to Assassin's Creed Valhalla in that they were able to maximize it with Skylake. AMD, on the other hand, took up to Zen 3 to get there. AMD improved performance by just over 50% from 2017 to 2020, while Intel didn't move but was already ahead.

Cyberpunk 2077's results aren't that different from other CPU-intensive games we've tested so far. In 2012, Intel enjoyed an architectural performance advantage of 48% and by 2015 that margin had doubled to 110% with Skylake. This is why the AMD FX series has been such a disaster for the company. Knowing that it was a catastrophic failure, they were forced to shift gears with a fundamental redesign, and that took time.

So, starting in 2012, it was five years before we saw another flagship CPU from AMD, and during that time they had to struggle with Intel, which in this example had an architectural advantage of 110% and used half the power, brutal stuff.

What we can also see here is Intel's steady refinement of their 14nm process all the way to Skylake. From Sandy Bridge to Skylake, Intel has seen nearly 50% more performance over a 4 year period, which is remarkable. However, from Skylake to Rocket Lake, which is a 6 year gap, they only found 13% more performance when no cores were added and the same clock speed was running.

Shadow of the Tomb Raider shows us how Intel took advantage of a much more efficient architecture almost 10 years ago and delivered 32% more power at the same frequency. Then from Ivy Bridge to Haswell we see another 23% increase for Intel and from Haswell to Broadwell a 13% increase. Then it gets slower, Skylake was only 5% faster than Broadwell, and that means that from 6th to 10th generation Intel only increased performance by 11%.

We also see AMD shadowing Haswell with Zen and then matching with Zen +. But when Zen 2 arrived, AMD was up to speed and then raced ahead with Zen 3, though Intel caught up with them a little with Rocket Lake.

Watch Dogs Legion provides us with another great example of Intel's stagnation over the past 6 years, essentially hitting a wall with Skylake, although they recently overcame it with Rocket Lake. As before, we can see how brutal it was for AMD to get a margin in the 95% margin with Haswell before 2017, and AMD wouldn't be able to achieve that level of performance until 2017 with Zen.

Performance summary

All of the data we have collected provides a real glimpse into comparing AMD and Intel CPU architectures over the past decade for gaming. We've seen AMD come out of nowhere to often beat Intel, while the latter has made smaller strides in terms of pure architecture in the past five to six years. Now let's look at the 9-game average …

In 2012, when comparing the FX-8350 and Core i7-3770K with the same clock frequency, Intel was ahead with an average of 43%. That's a huge difference, and what's more, AMD used significantly more power back then.

But if that were just a worst-case scenario for AMD. Just a year later, Intel expanded that margin to 77% and by 2015 it was 110% ahead.

Tech-wise, the 6700K wasn't even Intel's flagship desktop piece in 2015, that credit goes to the Core i7-5960X and its 20MB L3 cache, but that's another story. Given these massive performance gaps, AMD was forced to cut prices from $ 195 in 2012 to just $ 120 in 2015, with the chip dropping as low as $ 90 on some sales. That would be like having to sell a part like the 5800X for less than $ 200 today.

Obviously, AMD was able to reverse this situation years later. And while Zen wasn't a gaming monster in 2017, it was good enough to improve on those basics. Zen + continued this trend in 2018, and by 2019, AMD slumped Intel's desktop CPU sales.

What we learned

When we summarize this interesting look at the architectural performance differences for gaming between AMD and Intel over the past decade, this development clearly shows why our ratings as of 2018 were overwhelmingly positive about AMD Ryzen processors and why they were so pro-Intel before 2017.

In the FX days, it was extremely difficult to recommend AMD CPUs for PC gaming. The arrival of Ryzen marked a turning point, and while I wasn't entirely convinced of the first generation, at least for games, there were plenty of scenarios in which it could be recommended.

With this data, you can see why Zen + has started shifting things in favor of AMD. They still relied well on Intel for gaming when the CPU was limited, but a lot of games aren't CPU-limited, and not all gamers just want to play games. This made Ryzen a desirable and well-rounded solution that was also inexpensive.

With Intel and AMD now tied for gaming performance – AMD has a distinct efficiency edge, which is funny looking back at 2012 – we can't wait to see what the next generation of hardware brings.

There is a lot of promise behind Intel's 12th generation Alder Lake CPUs, which are arriving very soon and are expected to bring about major changes, that should be very exciting to potential buyers and the overall marketplace. Stick with it as we plan to review the new chips in due course.

Purchase abbreviation:

Leave a Reply

Your email address will not be published. Required fields are marked *