AMD's Ryzen Benchmarks Show Where It Wins, And Loses, To Intel

Image: Kotaku

Everyone is used to companies bragging about how their hardware is better than the competition, and AMD's launch event for Ryzen was no exception. But what you don't see often is a company showing the instances where their hardware doesn't win.

After Dr Lisa Su gave her initial presentation at AMD's Ryzen launch event in San Francisco, other speakers began to take the stage. But it wasn't until a breakout session afterwards where AMD delved a little deeper into gaming benchmarks for their 1800X 8-core CPU - as well as some figures for their 6-core offerings, Ryzen 5.

Ryzen is officially on sale now, and the benchmarks were under NDA so I couldn't share them with you earlier. But what you'll see is the Ryzen 7 1800X, 1700X and 1700 pitched against Intel's i7-6900K, i7-6800K and the more common i7-7700K, respectively, with Intel's CPUs coming out ahead for the most part.

Images: AMD

The CPU that's stands out to me is the Ryzen 7 1700, since it's priced at $469 locally (the kind of CPU people building a gaming PC for $2000 or under would consider). It's also a fraction cheaper, by about $20, than the i7-7700K used in the last slide. (The gap between the 1700 and the 7700K in GTA 5 at 1440p is pretty stark, mind you, considering both machines are running a NVIDIA GTX 1070 GPU.)

If you're interested, here was the settings AMD say they used for all six machines in their testing:

Image: AMD

It's a bit of an odd mix, but it's also missing a few titles that have favoured AMD over the last year. There's no Total War: WARHAMMER, which AMD was a big supporter of prior to release, and there's no Hitman or Rise of the Tomb Raider, two other games that have been staples in DirectX 12 testing.

But those games have also been good to AMD graphics cards, not AMD CPUs in particular. And one point AMD frequently hammered home in San Francisco was that they were still in the process of optimising their compilers and reaching out to developers, which should result in performance gains down the road. (Those performance gains will probably be marginal, as you often see with driver updates, but every bit adds up.)

Right now, AMD's pitch is all about the value play. As far as they're concerned, their CPUs are cheaper - significantly so, in the case of the 1800X - and the performance is comparable, if not marginally better in certain scenarios. And having more cores means they'll get more usage out of non-gaming applications, like streaming or video encoding, too.

But gamers will want to keep an eye on what those scenarios are. Right now, the price difference between the Ryzen 7 1700 and the i7-7700K is pretty negligible. But if you happen to be a streamer? Or you're a game developer who does a lot of encoding/rendering and you want an 8-core or 10-core CPU? That's a different equation.

Either way, it's always interesting to see a company come out with tests that don't always show their product winning. Gamers, and the tech world, is so often accustomed to the opposite, tests that are skewed just the right way so CEOs and PR can say "we have the fastest product EVER".

AMD's pitch is a little different. Third party benchmarks might not be quite as flattering - and we'll have our own in the next week or so - but it's a refreshing change of message nonetheless.


The author travelled to San Francisco as a guest of AMD.


Comments

    Looked at the prices for these new CPU's and they don't seem that much cheaper compared to the Intel CPU's. Unless if they just had a price drop?

      They haven't yet; I've been checking in with retailers to check just in case. The 6900K is still priced at $1400+ however, while the Ryzen 1800X is about $700. The difference between the 7700K and the Ryzen 1700 is pretty minimal in Australia though, which weakens AMD's messaging a tad.

        I probably wouldn't expect the 6900K to come down by much until the KBL-X CPUs hit in a few months. There's an opportunity here for AMD to pick up some of the HEDT market if they can convince people that it's better to buy now for cheap than wait to see how KBL-X looks.

        Are these the benchmarks you mentioned in your reply to me last article, Alex? You mentioned benchmarks that you had (or had seen) but weren't allowed to talk about yet.

          Basically. There were a couple of other graphs with the Ryzen 5 - but those weren't supplied to me, so unfortunately I can't republish it here.

    Does it mention anywhere what GPU they were using for these tests? Just curious how CPU-bound the results are or if none of them were being properly taxed. With a good GPU the CPU load usually isn't that stressful. On the settings table, systems C/D and E/F have the same settings so I'm guessing they had different GPUs in them.

    Last edited 03/03/17 12:34 pm

      Alex mentions a "NVIDIA GTX 1070" in passing about one of the comparisons, assume it's that one.

        TITAN X for the first two comparisons, GTX 1080 for the second, GTX 1070 for the last lot.

          Here is your instant problem. Ryzen destroys Nvidia and Intel when pared with AMD GPU hardware like 480s when running in DX12.

          The bottleneck is not hardware with Ryzen, it's Nvidia's dogshit API and awful Drivers. Ironically contrary to popular belief, AMD is miles, miles ahead of Nvidia in the Driver/API department when it comes to DirectX 12.

          Run these same Ryzen tests with 480s. You will see massive, massive DX12 improvements.

    So basically Intel is still king for gaming. Hopefully these AMD GPUs will drop in price enough to justify the performance and lack of overclockability.

      its because most dx11 games will perform better with higher clock speed than more cores. Which is why a $309.99 i7 6700K will perform better than a $1,109.00 i7-6900K.

        Yeah. The bet here is that games will start to take advantage of more cores/multi-threading and that the balance will tip to users with 6/8/10/12 cores in the next year, two years.

        (but that relies an awful lot on work from the developers, single core threaded performance is still a priority for developers because most PCs are 2 or 4 core systems, and Vulkan/DX12 is still in the early stages, so a lot of the benefits it provides isn't being taken advantage of)

          I hope you're right but we've been waiting 10 years since PS3, Intel q6600 days for developers to get off their asses haha.

    You can only truly stress thus test the CPU's at lower resolutions such as 1080p many of these benchmarks are completely invalid for comparing these CPU's unfortunately and people are easily fooled!

    Dear Alex,

    Thanks for the information regarding AMD Ryzen benchmarks and your thoughts, I am happy to see AMD has after many years come out with a product that can be competitive in workstation and multimedia environments. Unfortunately it seems that they have not been able to grasp Intel's leadership when it comes to gaming side as the performance is lesser. I plan on making a video on YouTube, hopefully you will be able to view it and provide feedback.

    https://www.youtube.com/channel/UC7jFuvAklYvYYIMXUUsixwA

    Thank you,

    Maximilian

Join the discussion!

Trending Stories Right Now