Does Ryzen offer a smoother gaming experience?

[DarkKnightDude]

BF1 is definitely way smoother on the 1700X than my i7 3770K. Might be the extra cores helping do more background tasks, not 100% sure.

Might be placebo ofc, but I don't get any stutters like the occasional one on my 3770K.

I don't know why sites do tests of singleplayer BF1, I only play MP, and 64 man is certainly demanding.

 

[bystander36]

Everyone on this forum should see this

http://www.youtube.com/watch?v=ylvdSnEbL50<br /><br /><br />Ryzen-The tech presses loses the plot

I watched the video, and agree and disagree with a few of his ideas.

1) The reason people test low resolution is for gaming NOW, not how they age. Games now, bottleneck on the CPU with the fastest CPU's in existence. They just don't bottleneck all the time. Most benchmarks are in good GPU demanding areas, that are no where near bottlenecked by the CPU, but real world gaming is often different in other areas of the same game, and often time online.

2) Multithreading in gaming does not scale well at all, so while games going forward will improve more with good threading, it won't be the massive uplift he and many are expecting. (Joker redid those tests which were all GPU bound, at 720p, and Intel's CPU pulled ahead, despite what he thought would happen).

I agree with a lot of other stuff, such as waiting a year to really test them and optimizations may make it a lot stronger.

 

[Teizo]

While I agree, I do have to question your ability to be objective here, when you think Fast Sync makes things smooth. Nvidia themselves have made it clear that fast sync makes games choppier and my experience with it matches theirs. Fast Sync is good for low latency, not smoothness. At high enough FPS, fast sync isn't as bad, which is the main purpose for this type of sync (think CSGO).

Fast sync in my experience does not make games choppier. I say it improves the gaming experience because it eliminates tearing and on my system it is still butter smooth. Fast sync only works when your fps is above your refresh rate...My monitor is only 60hz. Not sure why you question my objectivity lol....I know for a damned fact that Fast Sync works well. It may, if you are a robot who can process millisecond differences, increase latency over vsync off....but only a data sheet elitist would complain in comparison to the latency of V-Sync off because it is so minuscule as stated only a robot could discern the difference. I've never used fast sync with fps below refresh rate....my 1080 doesn't perform poorly enough for that to ever happen.

I can't speak for others...but I'll say it again....the 7700K is butter smooth in gaming. Anyone saying otherwise is just wasting good oxygen. I am sure Ryzen 7 chips are butter smooth as well. And...I'll repeat it again....when I frame cap my 1080 to 65fps and using fast sync to eliminate the tearing that you get from it....game play (speaking directly from BF1 as that is the only game I play) is absolutely flawless. There is no discussion to be had and nothing further to debate. It is a fact. It is just as good with frames uncapped as well. I just use frame capping because my monitor is 60hz.

 

[bystander36]

Fast sync in my experience does not make games choppier. I say it improves the gaming experience because it eliminates tearing and on my system it is still butter smooth. Fast sync only works when your fps is above your refresh rate...My monitor is only 60hz. Not sure why you question my objectivity lol....I know for a damned fact that Fast Sync works well. It may, if you are a robot who can process millisecond differences, increase latency over vsync off....but only a data sheet elitist would complain in comparison to the latency of V-Sync off because it is so minuscule as stated only a robot could discern the difference. I've never used fast sync with fps below refresh rate....my 1080 doesn't perform poorly enough for that to ever happen.

I can't speak for others...but I'll say it again....the 7700K is butter smooth in gaming. Anyone saying otherwise is just wasting good oxygen. I am sure Ryzen 7 chips are butter smooth as well. And...I'll repeat it again....when I frame cap my 1080 to 65fps and using fast sync to eliminate the tearing that you get from it....game play (speaking directly from BF1 as that is the only game I play) is absolutely flawless. There is no discussion to be had and nothing further to debate. It is a fact.

Well, as someone who has used it for a while, to test it out (a few days), it showed very noticeable stuttering. Nvidia, when describing said it will cause stuttering. It's primary purpose is for players who play CSGO, and similar games where 300 FPS is possible, so that the game can have reduced latency.

The reason I question your objectiveness, or maybe I should say, I question your ability to compare, is because you do not see a difference and you should.

Edit: if you are only getting a little higher FPS than your refresh rate, Fast sync won't make much of a difference either way. That would probably be the only reason you don't see choppiness, or if it's super high above your refresh rate.

 

[Dante80]

 

[Teizo]

Well, as someone who has used it for a while, to test it out (a few days), it showed very noticeable stuttering. Nvidia, when describing said it will cause stuttering. It's primary purpose is for players who play CSGO, and similar games where 300 FPS is possible, so that the game can have reduced latency.

The reason I question your objectiveness, or maybe I should say, I question your ability to compare, is because you do not see a difference and you should.

Edit: if you are only getting a little higher FPS than your refresh rate, Fast sync won't make much of a difference either way. That would probably be the only reason you don't see choppiness, or if it's super high above your refresh rate.

As I said...I see no choppiness in BF1 with fast sync getting 125 fps or with frame capping at 65 fps. No stuttering at all either way. I use it explicitly...all the time. Not just a couple of days...but for the last two months (new pc build). Anyway, let me educate you on something. When you frame cap your fps via RTSS and AB...you get terrible frame tearing. I mean, really bad. Worse than V-Sync off. Same thing happens if you frame cap in game as well. So, out of experiment one day I decided to see if I frame capped BF1 via RTSS to 65 fps (I do this to reduce system power consumption and reduce temps) that if used fast sync would it eliminate the tearing. To my surprise...it worked...and worked really well. I got a consistent 65 fps with no variance whatsoever. Game play was butter smooth as well...just as smooth as it was with 125fps. Only by having a fps counter on would I know the difference. Only placebo effected gamers could think otherwise (except high refresh rate CSGO gamers perhaps). As mentioned. If you use the in game fps lock, fast sync doesn't work because fps has to be in excess of 60 fps. That is why I do it via RTSS. It was simply out of trial and error I discovered this. It doesn't matter if your frames are locked to 65 fps...the benefit is there and the benefit is good if you are trying to reduce power consumption and heat output. As mentioned earlier...I only do it because my monitor is 60hz and any frame over 60 fps is wasted anyway.

I had a friend who posts here test it out and he confirmed with me fast sync while frame capping works well. I really don't care what you have to say in return because I know for a fact it works..and works well...and he agreed. I saw the pc perspective video with Tom Petersen talking about fast sync and I don't recall him saying a thing about stuttering. He did mention latency and that fast sync has higher latency than v-sync off...but nothing any human could discern. But, the benefit was that you got no tearing...similar to G-Sync..just not with the higher refresh rate capabilities, etc.

 

[Joric]

I watched the video, and agree and disagree with a few of his ideas.

1) The reason people test low resolution is for gaming NOW, not how they age. Games now, bottleneck on the CPU with the fastest CPU's in existence. They just don't bottleneck all the time. Most benchmarks are in good GPU demanding areas, that are no where near bottlenecked by the CPU, but real world gaming is often different in other areas of the same game, and often time online.

2) Multithreading in gaming does not scale well at all, so while games going forward will improve more with good threading, it won't be the massive uplift he and many are expecting. (Joker redid those tests which were all GPU bound, at 720p, and Intel's CPU pulled ahead, despite what he thought would happen).

I agree with a lot of other stuff, such as waiting a year to really test them and optimizations may make it a lot stronger.

AdoredTV is almost always extremely optimistic towards AMD in his analysis/predictions and pessimistic towards intel/nvidia.
If you want relentless AMD cheerleading as tech analysis, he's your man.

 

[piesquared]

That says it all, he seems genuinely impressed. Especially when he rifling that mouse around in Watchdog 2 GTA V.

 

[tamz_msc]

That says it all, he seems genuinely impressed. Especially when he rifling that mouse around in Watchdog 2.

That was GTA V.

 

[tamz_msc]

Only by having a fps counter on would I know the difference. Only placebo effected gamers could think otherwise (except high refresh rate CSGO gamers perhaps).

You'd be wrong to assume that except those who play CSGO competitively at high refresh rates, other people cannot tell the difference between 60fps locked and 120fps locked.

16.67ms vs 8.33ms is very noticeable, even when you can hold them at all times.

 

[piesquared]

That was GTA V.

Yes it was, thanks. He sure seemed genuinely surprised though throughout at the smoothness.

 

[tential]

Judging ryzen on gaming before there is proper SMT support makes zero sense. Hyperthreading had similar issues with Windows at one point. Once Windows figures out how to manage SMT then I expect this to be a whole different story.

Otherwise, if windows can't use SMT, I can't use Ryzen.

 

[bystander36]

As I said...I see no choppiness in BF1 with fast sync getting 125 fps or with frame capping at 65 fps. No stuttering at all either way. I use it explicitly...all the time. Not just a couple of days...but for the last two months (new pc build). Anyway, let me educate you on something. When you frame cap your fps via RTSS and AB...you get terrible frame tearing. I mean, really bad. Worse than V-Sync off. Same thing happens if you frame cap in game as well. So, out of experiment one day I decided to see if I frame capped BF1 via RTSS to 65 fps (I do this to reduce system power consumption and reduce temps) that if used fast sync would it eliminate the tearing. To my surprise...it worked...and worked really well. I got a consistent 65 fps with no variance whatsoever. Game play was butter smooth as well...just as smooth as it was with 125fps. Only by having a fps counter on would I know the difference. Only placebo effected gamers could think otherwise (except high refresh rate CSGO gamers perhaps). As mentioned. If you use the in game fps lock, fast sync doesn't work because fps has to be in excess of 60 fps. That is why I do it via RTSS. It was simply out of trial and error I discovered this. It doesn't matter if your frames are locked to 65 fps...the benefit is there and the benefit is good if you are trying to reduce power consumption and heat output. As mentioned earlier...I only do it because my monitor is 60hz and any frame over 60 fps is wasted anyway.

I had a friend who posts here test it out and he confirmed with me fast sync while frame capping works well. I really don't care what you have to say in return because I know for a fact it works..and works well...and he agreed. I saw the pc perspective video with Tom Petersen talking about fast sync and I don't recall him saying a thing about stuttering. He did mention latency and that fast sync has higher latency than v-sync off...but nothing any human could discern. But, the benefit was that you got no tearing...similar to G-Sync..just not with the higher refresh rate capabilities, etc.

This is queued up at the point where he talks about it a bit after a question on it:
https://youtu.be/WpUX8ZNkn2U?t=14m7s

3rd party analysis
https://youtu.be/L07t_mY2LEU?t=13m19s

Anyways, I believe that you didn't see a difference, and your friend too, but it does put into question your sensitivity to smoothness in the first place.

 

[Zytryx]

I've noticed this too, playing BF1 my 6700k would stutter when zooming in and out of the spawn screen it wasn't fps as I'm locked at 75fps with vsync on a 1080.

With my 1700x same card, monitor and fps that animation is now butter smooth. I also see this in elite:dangerous. I don't know why but it's really there.

I also saw one YouTube vid where it was also noticeable and ppl commented on it in the comments

 

[guachi]

I don't read Hard OCP much. But one thing I like is that their GPU reviews do a "what settings can I choose and still have an enjoyable experience?" in the games they test. At some point, FPS differences just don't matter at all because the GPU is "good enough". That's why I think the 1080 tests are silly and I'd rather see 1440 or 4k tests. Every game tested at 1080 with a 1080 GPU is "good enough". I mean, people on this very board recommend a mid-level RX480/1060 for 1080 gaming and now I'm supposed to think I need a 1080 GPU and $300+ processor to play at 1080?

Exactly how fast does a CPU have to be to get max FPS at 4k settings? Because that's useful information. The reader can save hundreds of dollars on a CPU and shift that to his GPU instead.

As to results of testing and what matters - I mean, not all 50% fps differences are the same.

40 fps to 60 fps - big, noticeable difference.
60 fps to 90 fps - somewhat noticeable difference, especially if fps minimums drop enough
90 to 135 fps - just don't care.

Test at settings people would actually play at (and max/ultra probably isn't it. At least, the visual differences are negligible if even noticeable.)

Though in all of these cases you could just buy a *sync monitor and your gaming experience will improve measurably as your fps can get quite low before it matters.

 

[Glo.]

Can anyone explain to me, what affects the minimums in framerates? Is it core count? Or the microarchitecture itself?

 

[Markfw]

Can anyone explain to me, what affects the minimums in framerates? Is it core count? Or the microarchitecture itself?

I would say both, but microarchitecture is probably more to count after 4 cores.

 

[Teizo]

Anyways, I believe that you didn't see a difference, and your friend too, but it does put into question your sensitivity to smoothness in the first place.

I used to use 760 SLI and I definitely know micro-stuttering. It was atrocious in BF4. It got to the point I sold them and bought a 980. In fact, a single 760 delivered better game play with lower fps than SLI. Maybe enabling 'threaded optimization' in the Nvidia control panel helps...and maybe Kaby Lake helps as well. If it was there to any noticeable degree that affected game play...I would know. It may be there and can be measured with high sensitivity instruments...but I tell you game play is butter smooth. BF1 for me did have the occasional stutter with my 6600K before I switched over to the 7700K. Now it doesn't. BF1 does make good use of extra cores...so I imagine game play with any Ryzen 7 or any Ryzen 5 in the very near future will be very good too. One thing that may be helping this as well is my 960 Evo. I'm not sure. I don't see why it would...but maybe that is playing a role as well since data is being sent over PCIe X4 at a faster rate than Sata SSD and there is no data flushing happening. Or maybe the 3200mhz memory. Some how or another with my system...the stuttering is just not there to any noticeable degree worth mentioning and any discussion about 7700Ks stuttering is jibberish in my book and just faux nonsense to try to defend Ryzen for gaming use or make Ryzen seem more appealing for gaming.

 

[Glo.]

I would say both, but microarchitecture is probably more to count after 4 cores.

Presumably its all done to number of instructions dispatched and executed by the CPU and its scheduler?

 

[Markfw]

Presumably its all done to number of instructions dispatched and executed by the CPU and its scheduler?

Now you are outside of my knowledge base. I was guessing based on the little I understand about microarchitecture.

 

[Glo.]

Now you are outside of my knowledge base. I was guessing based on the little I understand about microarchitecture.

Fair enough. Thanks anyway

 

[piesquared]

I've noticed this too, playing BF1 my 6700k would stutter when zooming in and out of the spawn screen it wasn't fps as I'm locked at 75fps with vsync on a 1080.

With my 1700x same card, monitor and fps that animation is now butter smooth. I also see this in elite:dangerous. I don't know why but it's really there.

I also saw one YouTube vid where it was also noticeable and ppl commented on it in the comments

So to summarize, for whatever reason, Ryzen seems to give gamers a smoother, stutter free gaming experience. If we remember when frame pacing was all the rage, that should seal the deal as far as which CPU is better for gamers. That was the sole reason some websites (PCPer, Techreport, etc) recommended NV cards over AMD cards at the time. I'm sure that hasn't changed (or has it??) and judging from that video from Jayz3cents it is quite noticeable. Like Zytryx reports others are reporting the same results. Seems like a no brainer then for gamers.

 

[bystander36]

I used to use 760 SLI and I definitely know micro-stuttering. It was atrocious in BF4. It got to the point I sold them and bought a 980. In fact, a single 760 delivered better game play with lower fps than SLI. Maybe enabling 'threaded optimization' in the Nvidia control panel helps...and maybe Kaby Lake helps as well. If it was there to any noticeable degree that affected game play...I would know. It may be there and can be measured with high sensitivity instruments...but I tell you game play is butter smooth. BF1 for me did have the occasional stutter with my 6600K before I switched over to the 7700K. Now it doesn't. BF1 does make good use of extra cores...so I imagine game play with any Ryzen 7 or any Ryzen 5 in the very near future will be very good too. One thing that may be helping this as well is my 960 Evo. I'm not sure. I don't see why it would...but maybe that is playing a role as well since data is being sent over PCIe X4 at a faster rate than Sata SSD and there is no data flushing happening. Or maybe the 3200mhz memory. Some how or another with my system...the stuttering is just not there to any noticeable degree worth mentioning and any discussion about 7700Ks stuttering is jibberish in my book and just faux nonsense to try to defend Ryzen for gaming use or make Ryzen seem more appealing for gaming.

Like I said, I believe you don't notice it. It is still there. The technique fast sync uses inherently has stutter. There is no way around it. It might not be as bad as being below your refresh rate with v-sync on stutter, but Fast sync technique will always have a degree of stuttering. It's how it is designed.

Either you are not sensitive to it, or maybe you are just liking the low latency aspect of it, and ignore the stuttering it adds. This isn't a bad thing for you. You simply don't notice mild stuttering.

Edit: go back to the Q&A video on the subject. He explains why it will due to the science of it. The 2nd video tests it, and noticed it in practice. Fast sync, unless fundamentally changed, will always add some stutter.

 

[Paul98]

it shouldn't be any surprise, Ryzen will have far more free resources to handle other tasks. Obviously it won't be as fast on single thread performance, but when you start throwing more and more at it it will handle it much better, will help avoid the stalls and stutter that you might get otherwise.

If you take a very much simplified example. If you have a single task and a fast single core you should be able to finish that up very quickly. Now take that same single task and add in a couple more smaller tasks all trying to be done on the same single core. You now have to stall that single task when those others are done. Where as if you have more cores you don't need to stall that main task even if the cores are a little slower. Sure that main task might take a little longer due to slower single core, but it won't stutter due to the extra tasks. Now obviously this is very much simplified and modern cores can do a lot, but you get the idea

 

[Rifter]

Anyone else find it funny that we have a ryzen fails at gaming thread right beside the ryzen has smoother gaming thread?

What a crazy world we live in.