AMD Zen Discussion Thread

cherryghost

Lacklustre gaming benchmarks
http://www.kitguru.net/components/cpu/luke-hill/amd-ryzen-7-1800x-cpu-review/

K.O.Kiki

cherryghost wrote: »

Lacklustre gaming benchmarks
http://www.kitguru.net/components/cpu/luke-hill/amd-ryzen-7-1800x-cpu-review/

Balanced out by the phenomal productivity benchmarks though.

What a strange series of results.
Guess it really is more of a workstation CPU.

farna_boy

K.O.Kiki wrote: »

Balanced out by the phenomal productivity benchmarks though.

What a strange series of results.
Guess it really is more of a workstation CPU.

From what I've read, they seem poorly optimised at 1080p and seem to fair better at higher resolutions.

They also seem to fair much better in some games rather than others. Apparently Ashes of Singularity have released a statement that they need to release a patch for Ryzen and until then it shouldn't be used as a benchmark for Ryzen.

TerrorFirmer

cherryghost wrote: »

Lacklustre gaming benchmarks
http://www.kitguru.net/components/cpu/luke-hill/amd-ryzen-7-1800x-cpu-review/

Ah for **** sake, the 1800X, a €500 8 core, 16 thread CPU trading blows with the Ivy Bridge 2700K and the Sandy Bridge 3570K in games.

You can talk about Ashes being an anomaly, and being better in some games than others, but at the end of the day the flagship 1800X is lacklustre in every single gaming benchmark in that review, and miles behind the common and way cheaper i7 processors.

With the 8c/16t flagship CPU performing extremely lacklustre in games - what does that mean for the slower and lower clocked CPU'S - R3, R5, etc?

Massive disappointment. I do understand they are still good CPU's for productivity and from a workstation/professional prospective they're still viable, but from a gaming POV it is a titanic disappointment.

I was all set for a Ryzen R5 build but after looking like that it will be hands down an Intel build.

For gamers, it's the FX situation all over again.

Edit: Looking at some other reviews now, they're all saying the same thing - Ryzen is not for gamers.

deceit

By the looks of the kitguru review I got lucky with jumping the gun.
Perfectly suits me as I run an Ultrawide 3440x1440 monitor and would use my main rig for work first before I game so couldn't have picked a better cpu .

Now to convince the missus that a 1080ti is a good decision since I got such good value on this :P

The only bad point is the really awful overclocking . I'm hoping this is a bug in bios but I doubt it.

farna_boy

TerrorFirmer wrote: »

Ah for **** sake, the 1800X, a €500 8 core, 16 thread CPU trading blows with the Ivy Bridge 2700K and the Sandy Bridge 3570K in games.

You can talk about Ashes being an anomaly, and being better in some games than others, but at the end of the day the flagship 1800X is lacklustre in every single gaming benchmark in that review, and miles behind the common and way cheaper i7 processors.

With the 8c/16t flagship CPU performing extremely lacklustre in games - what does that mean for the slower and lower clocked CPU'S - R3, R5, etc?

Massive disappointment. I do understand they are still good CPU's for productivity and from a workstation/professional prospective they're still viable, but from a gaming POV it is a titanic disappointment.

I was all set for a Ryzen R5 build but after looking like that it will be hands down an Intel build.

For gamers, it's the FX situation all over again.

Edit: Looking at some other reviews now, they're all saying the same thing - Ryzen is not for gamers.

A different set of benchmarks tell a slightly different story:

http://www.guru3d.com/articles_pages/amd_ryzen_7_1800x_processor_review,16.html

K.O.Kiki

Probably should wait for retests with fixed drivers

TerrorFirmer

farna_boy wrote: »

A different set of benchmarks tell a slightly different story:

http://www.guru3d.com/articles_pages/amd_ryzen_7_1800x_processor_review,16.html

No they don't. That's GPU bottleneck @ 1080p. Obviously to a point they're gonna be OK - same as the FX series - but run a seriously CPU demanding title and see the clear separation.

I'm getting really pissed off at a lot of review sites lately for running garbage benchmarks like GPU dependent titles at 1080P. What's that supposed to tell us!

This is a better example - heavily CPU dependent title at 1440P. Notice how the i5-6600K soundly beats it at stock speed at half the price, while the i7-7700 destroys it at stock even with SMT disabled which is apparently causing some issues.

With SMT, the results are even worse for the 1800X - almost loses to the bloody Kabylake i3 on raw CPU performance in a CPU dependent game!

Xenoronin

That honestly looks like a weirdly optimised game to be honest. I don't see how turning on SMT causes a FPS drop.

TerrorFirmer

Xenoronin wrote: »

That honestly looks like a weirdly optimised game to be honest. I don't see how turning on SMT causes a FPS drop.

It's not just that game, most reviewers are mentioning SMT as a factor.

It doesn't make an overall huge difference though, even without it and with increased performance, it's still lacklustre from a gaming POV.

Universal consensus across the board - Ryzen is a very good CPU for productivity, but weak for games.

Lukker-

FYI there is a massive discrepancy across different sites

Had it on par with 7700k

cherryghost

If you read the entire article, Guru3d are completely stumped with the 1080p benchmarks, assuming it could be a bug. Personally dont buy it, it just scales better at higher resolutions paired with a GPU that likely won't bottleneck the system.

TerrorFirmer

That's because a lot of review sites are stupidly using very GPU dependent titles where GPU hits bottleneck at 1080p.

Find review sites using games like Fallout 4, Total War: Warhammer etc and see a massive difference.

Quote from Legitreviews which sums it up:

Ryzen is a whole world better than the outgoing AMD FX CPU’s, but at times is only as fast as a ~6 year old Intel Core i7-2700K. The unfortunate part of that is the Intel Core i7-2500K and 2600K are often showing up as the “Minimum Specs” for some Triple A PC game titles. Hopefully AMD can figure out a way to improve 1080P gaming performance as they’d have an amazing part here. Is Ryzen an Intel killer? No, but it is putting Intel on notice.
Read more at http://www.legitreviews.com/amd-ryzen-7-1800x-1700x-and-1700-processor-review_191753/15#WX3TgpWe5yZrqs5j.99

It's still a threat to Intel as the productivity performance is very good, which is the market that really matters, but it's just incredibly disappointing for us gamers.

Lukker-

cherryghost wrote: »

If you read the entire article, Guru3d are completely stumped with the 1080p benchmarks, assuming it could be a bug. Personally dont buy it, it just scales better at higher resolutions paired with a GPU that likely won't bottleneck the system.

In the review above I just posted an R7 1700 at 3.9 Ghz pretty much trades blows with an i7 7700k at 5.0 Ghz all at 1080p, with the 7700k being behind in mutlithreaded games and slightly ahead in the others.

deceit

I wonder could the differing reviews be down to some reviewers are using gigabyte board which seem to get good reviews and the reviews I've see that where worse are running the Asus. I'm hoping this is the case as these reviews really look all over the place the more I look into them. Some really good, some really poor. Nothing in the middle.

L

So, slightly stupid question, what's causing the divergence between performance in productivity and gaming?

deceit

L wrote: »

So, slightly stupid question, what's causing the divergence between performance in productivity and gaming?

This is something I'm wondering myself/cant figure out.
The only thing I can think of is optimization if single threaded performance is close/similar in non gaming applications?
Edit:
I was reading a few reviews and a few of the reviews believe it could be bugs which will hopefully be ironed out over time. Here's hoping as I would love to have the 1800x and for intel being forced to drop their 6900k in price to match so could then pretty much get two cpu's for the price of one :P
When I get it, I will try compare it against a 5820k and see how it fairs before deciding if it will be my main / other rig.

deceit

quick check on linustechtips and AMD responded, looks like its simply optimization:
https://linustechtips.com/main/topic/746289-amd-responds-to-1080p-gaming-tests-on-ryzen/

TerrorFirmer

L wrote: »

So, slightly stupid question, what's causing the divergence between performance in productivity and gaming?

Inferior single thread performance. Productivity applications tend to scale excellently across cores/threads.

Games to a much lesser degree and single thread performance is still very important even when games do scale across cores.

Hence why AMD's FX series, even with eight cores, frequently lose to dual core i3 processors in games, even those that favor more cores.

Ryzen looks to be a major improvement on the FX line but still behind Intel for games, but poses a serious threat in general workstation setups.

Lukker-

TerrorFirmer wrote: »

Inferior single thread performance. Productivity applications tend to scale excellently across cores/threads.

Games to a much lesser degree and single thread performance is still very important even when games do scale across cores.

Hence why AMD's FX series, even with eight cores, frequently lose to dual core i3 processors in games, even those that favor more cores.

Ryzen looks to be a major improvement on the FX line but still behind Intel for games, but poses a serious threat in general workstation setups.

Single thread performance is actually very similar, it's higher clock speeds that ultimately make them lag behind. Some other stuff too seem to be affecting performance but I suspect we won't get the full story for a couple of weeks.

wotzgoingon

Well tbh we knew the single core performance was behind a Kaby lake processor before today.

According to Joker Productions the 1700 is on par with a 7700K in gaming. Credit to Lukker- for posting the Youtube link.

TerrorFirmer

wotzgoingon wrote: »

Well tbh we knew the single core performance was behind a Kaby lake processor before today.

According to Joker Productions the 1700 is on par with a 7700K in gaming. Credit to Lukker- for posting the Youtube link.

That's one source though. Across the board it's clear that it is not on par. It is not also only a matter of raw clock speed.

Maybe it is a bug that can be fixed and I'm not saying Ryzen won't turn out to be a great gaming chip, but to say right now it is on par/beating Intel is totally false.

For me at the moment Ryzen is a bit of a lottery and Skylake i7 is still a much safer bet for those who are gamers (not that 1700 or 1800X were ever geared towards gamers in any case but it doesn't bode well for R3 and R5 if the flagship is lacklustre)

Ars Technica wrote:

The Ryzen 7 1800X, then, is a tale of two chips. One is a brilliant, disruptive, market changing, eight-core workstation powerhouse; the other is a competent, if wholly unsatisfying gaming CPU.

Techspot wrote:

But as smooth as the experience was, it doesn't change the fact that gamers running a high refresh rate monitor may be better served by a higher clocked Core i7-6700K or 7700K.

While the gaming results might not be as strong as we had hoped for, they are highly competitive and that should hold particularly true for the Ryzen 5 and 3 series.

PC Gamer wrote:

As it stands, Ryzen looks like an awesome workstation processor, and the Ryzen 7 1700 delivers an incredibly potent CPU for non-gaming purposes, especially if you want to overclock it. But the gaming story is disappointing.

With the huge strides in performance relative to the archaic (in CPU terms) FX-8370 Vishera chip, I expected Ryzen to achieve parity with Intel's X99 processors. It gets there in the CPU-centric tests, but falls well short on gaming performance. And I don't really have a good explanation, other than the feeling I keep getting that Ryzen was pushed out the door before it was truly ready.

Gamespot wrote:

The bottom line is that the Ryzen 7 1800X makes the most sense for users that need more than a CPU for gaming, those who are considered “prosumers.” The performance for multicore applications is impressive for its $499 price tag, which increases accessibility for those who need more power for video production and image rendering.

The 1800X isn’t intended for those with only gaming on their mind, as most games benefit more from individual core speed. However, we’re curious as to why it substantially underperforms in our two game tests.

Guru3D wrote:

So here is where we need to write an entire paragraph. You have been able to see that the Ryzen 7 1800X performance is good, but not just yet 100% where it needs to be. During the past week we have been going back and forth to figure out what could be causing the relatively lower game performance. To date, we have no valid answer to that. The graphics card runs properly at PCI-Express x16 3.0. We know from the RAW and synthetic performance benchmarks that the cores are fast enough, in fact VERY fast. Somehow that does not relate to game performance.

HardOCP wrote:

Gaming benchmarks on Ryzen are a critical mess. However it is worth noting here again, that these are very much benchmarks, and not actual gameplay. What we have come to find is that AMD has an issue with Ryzen at 1080P and seemingly lower resolutions when gaming, which is odd. I have talked to Ryan over at PCPer.com a few times during testing, and he has covered real world desktop gaming in-depth. I highly suggest giving it a read. I am not going to say much more about gaming here since I have little experience on that, but assuredly, there are some Ryzen problems when it comes to "low resolution" gaming.

Legit Reviews wrote:

Ryzen is a whole world better than the outgoing AMD FX CPU’s, but at times is only as fast as a ~6 year old Intel Core i7-2700K. The unfortunate part of that is the Intel Core i7-2500K and 2600K are often showing up as the “Minimum Specs” for some Triple A PC game titles. Hopefully AMD can figure out a way to improve 1080P gaming performance as they’d have an amazing part here. Is Ryzen an Intel killer? No, but it is putting Intel on notice.
Read more at http://www.legitreviews.com/amd-ryze...pWe5yZrqs5j.99

deceit

TerrorFirmer wrote: »

Inferior single thread performance. Productivity applications tend to scale excellently across cores/threads.

Games to a much lesser degree and single thread performance is still very important even when games do scale across cores.

Hence why AMD's FX series, even with eight cores, frequently lose to dual core i3 processors in games, even those that favor more cores.

Ryzen looks to be a major improvement on the FX line but still behind Intel for games, but poses a serious threat in general workstation setups.

What explains the single threaded application benchmarks being on par with Intel then?
The cpu's single threaded performance is very good in everything bar gaming by the looks of it which means this wouldn't be the issue if I'm not mistaken.
FX series single threaded performance was awful and can't be compared to this. Its single threaded performance was poor equally across everything.

If it can be fixed it will be a very good option for gamers (I suspect this is the case).
If not then people who solely use their system for gaming should steer clear and hopefully benefit from the dropped Intel prices.

TerrorFirmer

I honestly don't have the know how to account for it, neither do a lot of reviewers. AMD say incoming optimizations from certain developers for Ryzen will help but broadly speaking, there's been nothing offered as a concrete reason for the sub-standard gaming performance.

To me, it's just disappointing, and while it may be fixed, for me it would be a gamble to buy a high-end Ryzen CPU now purely for gaming. Originally the R3 would have competed with lower end i5's, and the R5 with the i7, from preliminary benchmarks. However, if the €500 1800X falls behind €350 i7-7700 and sometimes even €250 i5-7600, I can only imagine the R3 will be very disappointing with half the cores and threads and the amazing value it was promised to offer evaporates completely.

Equally, if you're using your PC as a workstation, it's thumbs up all the way without question for Ryzen so far.

I sold my i7-6700 machine to fund a new Ryzen build. Now I will wait it out a few weeks/months and see what happens in that time before making any decision.

Lukker-

Lisa Tsu and the AMD CPU team are doing an AMA on Reddit for those interested

https://www.reddit.com/r/Amd/comments/5x4hxu/we_are_amd_creators_of_athlon_radeon_and_other/

Q's still being answered.

deceit

https://www.pcper.com/news/Processors/AMD-responds-1080p-gaming-tests-Ryzen?utm_source=dlvr.it&utm_medium=facebook
AMD specifically say its because of a lack of optimization.
looks like two devs area already working on fixes for it:
Ashes dev and Total war dev.
I wouldn't bother about Ashes as this always benefits AMD but I would be interested to see what happens with Total war after improvements.

TerrorFirmer

AMD CEO says this about gaming performance in the AMA:

Ryzen is doing really well in 1440p and 4K gaming when the applications are more graphics bound. And we do exceptionally well in rendering and workstation applications where more cores are really useful. In 1080p, we have tested over 100+ titles in the labs…. And depending on the test conditions, we do better in some games and worse in others. We hear people on wanting to see improved 1080p performance and we fully expect that Ryzen performance in 1080p will only get better as developers get more time with “Zen”. We have over 300+ developers now working with "Zen" and several of the developers for Ashes of Singularity and Total Warhammer are actively optimizing now

Fine, but doing really well in 1440P and 4K gaming is grand, but....only because the GPU is taking the bottleneck pressure not the CPU.

Will be interesting to see Total Warhammer optimizations though as that's an excellent benchmark for raw CPU ability and one where the 1800X did particularly badly!

Inviere

Wow, only catching up on this now - very disappointing. Looks like I might have to go back to a 7600k build, rather than the 1600x 1080p is still where its at for me in gaming, so have no real interest in going above it to get the benefit of Ryzen - assuming it even is a benefit, and not just less of a bottleneck at that level

BloodBath

Confusing and disappointing stuff. I expected the 7600k/7700k overclocked to still be the best gaming cpu's, I didn't expect the 1800x to fall so far behind intel's 8/16 cpu's in gaming benchmarks at the same clock speeds. But then they aren't in other benchmarks from other reviewers.

There was supposed to have an issue with the memory controller which AMD patched but not on time for a lot of the reviews. I have no idea if this impacts these benchmarks but it could explain some of the differences in results.

Also how is the max overclock from 4Ghz only 4.1Ghz? There seems to be an issue with exhausting the heat from the cpu which is usually the case with high core count CPU's and SMT is causing lower performance in games so x models seem to be not worth the price premium at all.

Did AMD release their cpu overclocking software? Wasn't there a feature to disable cores to potentially achieve higher overclocks by using a lower amount of cores? If you could set up a profile to disable 4 cores maybe you can use it as a 4/8 at higher clocks for games.

Venom

TerrorFirmer wrote: »

AMD CEO says this about gaming performance in the AMA:



Fine, but doing really well in 1440P and 4K gaming is grand, but....only because the GPU is taking the bottleneck pressure not the CPU.

Will be interesting to see Total Warhammer optimizations though as that's an excellent benchmark for raw CPU ability and one where the 1800X did particularly badly!

The 4k results tying with Intel chips are down to GPU bottlenecking so will be interesting to see how Ryzen does with 1080Ti's in crossfire.

BloodBath wrote: »

Confusing and disappointing stuff. I expected the 7600k/7700k overclocked to still be the best gaming cpu's, I didn't expect the 1800x to fall so far behind intel's 8/16 cpu's in gaming benchmarks at the same clock speeds. But then they aren't in other benchmarks from other reviewers.

There was supposed to have an issue with the memory controller which AMD patched but not on time for a lot of the reviews. I have no idea if this impacts these benchmarks but it could explain some of the differences in results.

Also how is the max overclock from 4Ghz only 4.1Ghz? There seems to be an issue with exhausting the heat from the cpu which is usually the case with high core count CPU's and SMT is causing lower performance in games so x models seem to be not worth the price premium at all.

Did AMD release their cpu overclocking software? Wasn't there a feature to disable cores to potentially achieve higher overclocks by using a lower amount of cores? If you could set up a profile to disable 4 cores maybe you can use it as a 4/8 at higher clocks for games.

The memory issue fix is weeks away as far as I heard. The overclocking headroom range on Ryzen could just be low. What worries me the most, is that so many reviews are saying serious top tier cooling is needed for these new chips which offsets there cheaper prices.