AMD Ryzen 7 2700X 3.7 GHz

[Charcharo]

https://www.tweaktown.com/reviews/8602/amd-ryzen-7-2700x-5-2600x-review/index9.html

another review showing 8700k stock beating ryzen 2700x stock at 1440p by 20 fps in multiple games.

also GamersNexus shows multiple OC's in his review. its fine TPU doesn't as I have other sources that do, so meh. also like 90% of the chips hit 5ghz.

You wanted more games being tested? Sure thing, here you go, 8700k is 100-42% faster in many games, the cherry picked latest games its only 10 fps or 10% faster, sure...

I have 20 more games as well... you get the picture by now I think...

Include the source.

http://imgur.com/a/oTNrz

ANyways, the reason for some of these discrepencies is not Ryzen, but the Nvidia driver. How do I know this? I tested the exact same scenes that Jester/Boredgunner tested (we are friends) on a Ryzen 5 1500X and a Fury. It got higher fps than his 1700X OCed with a 1080 Ti. Let that sink it.

I honestly do not how this will be fixed or when, but Maxwell and Pascal + Ryzen is sometimes (not always!) a bad combination.

As for the person who used purepc.pl as an argument... those guys are definitely not testing games to find the most stresful parts. I do not know how or what they are doing, but I suspect it is not super competent.

As for the discussion about settings: Ultra settings increase CPU load but by how much depends on the game, scene, and engine. IMHO testing low resolution, Ultra settings is a valid CPU benchmark. Testing low resolution and low settings is valid only for E sports titles, and even then a difference between 400 and 500 fps is not going to be noticed by 99.99% of the people who read benchmarks. They are not good enough for it to matter to them.

 

[intrepid3d]

Something else about that slide ^^^^^ it is massively magnified, its a 6% difference made to look like its a massive more than 2x performance win for the 8700K.

I can't take people who magnify slides like that seriously.

 

[Charcharo]

Something else about that slide ^^^^^ it is massively magnified, its a 6% difference made to look like its a massive more than 2x performance win for the 8700K.

I can't take people who magnify slides like that seriously.

All the tests are serious. His slide skills are irrelevant . BTW, there are more slides than just one, just FYI.

 

[Vayra86]

@Vayra86 you're the one failing to grasp what we are talking about, you're talking about something entirely different to what the rest of us are, you're talking about resolution, no one is disputing using low resolution is a good idea because it helps stop the GPU from becoming the bottleneck, its actually really obvious what we are talking about higher stress complex scenes in games that cause low performance and sometimes stutter on the 4 core Intel CPU's.

ERRRN wrong. The games that stutter are limited mostly by a single core performance bottleneck, not a lack of cores. A single core can carry multiple game threads but you cannot split a single game thread across multiple cores ad infinitum.

So adding cores does only one thing, and that is providing more headroom for the heaviest game thread that is forced to that one core. And if a game hammers four cores with significant loads, then yes, anything you run in the background will influence ingame performance. But if all you run is an OS and some lightweight stuff alongside it, four cores are sufficient for the vast majority of games. Exceptions exist, but don't make the rule.

I will agree with you straight away that a hexacore is a sensible purchase for a gaming machine, but it is simply not true that a quad core by definition means you get stutter. The people experiencing stutter on quad cores today are mostly using DDR3-based rigs of Haswell or older - this included myself not too long ago. Most CPU-related stutter is actually attributable to RAM in equal or greater measure than it is to core count. Core counts are rather irrelevant, as long as you have performance left on those cores its just fine.

Another important bit of info you keep omitting is the fact that all of these tests are about > 60 FPS, not sub 60 FPS, ie the realm of Vsync, adaptive sync, Fast Sync or frame capping. Tons of ways to avoid stutter at this point without noticeably degrading the experience. When you speak of sub 60 FPS performance, the balance shifts rapidly towards clocks, and *sufficient* core counts, and show no scaling whatsoever beyond a 6 core CPU and only marginal scaling beyond 4 core CPUs.

 

[Charcharo]

ERRRN wrong. The games that stutter are limited mostly by a single core performance bottleneck, not a lack of cores. A single core can carry multiple game threads but you cannot split a single game thread across multiple cores ad infinitum.

So adding cores does only one thing, and that is providing more headroom for the heaviest game thread that is forced to that one core. And if a game hammers four cores with significant loads, then yes, anything you run in the background will influence ingame performance. But if all you run is an OS and some lightweight stuff alongside it, four cores are sufficient for the vast majority of games. Exceptions exist, but don't make the rule.

I will agree with you straight away that a hexacore is a sensible purchase for a gaming machine, but it is simply not true that a quad core by definition means you get stutter. The people experiencing stutter on quad cores today are mostly using DDR3-based rigs of Haswell or older - this included myself not too long ago. Most CPU-related stutter is actually attributable to RAM in equal or greater measure than it is to core count. Core counts are rather irrelevant, as long as you have performance left on those cores its just fine.

Another important bit of info you keep omitting is the fact that all of these tests are about > 60 FPS, not sub 60 FPS, ie the realm of Vsync, adaptive sync, Fast Sync or frame capping. Tons of ways to avoid stutter at this point without noticeably degrading the experience. When you speak of sub 60 FPS performance, the balance shifts rapidly towards clocks, and *sufficient* core counts, and show no scaling whatsoever beyond a 6 core CPU and only marginal scaling beyond 4 core CPUs.

Funny thing is, that some engines like the 4A Engine or id Tech 6 (and likely 7) can saturate 12+ threads under the right condition. That is how I tested my old i5 4570 vs my Ryzen 5 1500X. In some places, 1500X was almost 50% faster than the 4 core 4 thread i5.

If only we had PC Gamers demanding more simulation aspects from their titles . Imagine A-Life 2.0 on a multi-core engine with improved ballistic calculations and armor simulation... that is what lets me sleep at night ...

 

[intrepid3d]

ERRRN wrong. The games that stutter are limited mostly by a single core performance bottleneck, not a lack of cores. A single core can carry multiple game threads but you cannot split a single game thread across multiple cores ad infinitum.

So adding cores does only one thing, and that is providing more headroom for the heaviest game thread that is forced to that one core. And if a game hammers four cores with significant loads, then yes, anything you run in the background will influence ingame performance. But if all you run is an OS and some lightweight stuff alongside it, four cores are sufficient for the vast majority of games. Exceptions exist, but don't make the rule.

I will agree with you straight away that a hexacore is a sensible purchase for a gaming machine, but it is simply not true that a quad core by definition means you get stutter. The people experiencing stutter on quad cores today are mostly using DDR3-based rigs of Haswell or older - this included myself not too long ago. Most CPU-related stutter is actually attributable to RAM in equal or greater measure than it is to core count. Core counts are rather irrelevant, as long as you have performance left on those cores its just fine.

Another important bit of info you keep omitting is the fact that all of these tests are about > 60 FPS, not sub 60 FPS, ie the realm of Vsync, adaptive sync, Fast Sync or frame capping. Tons of ways to avoid stutter at this point without noticeably degrading the experience. When you speak of sub 60 FPS performance, the balance shifts rapidly towards clocks, and *sufficient* core counts, and show no scaling whatsoever beyond a 6 core CPU and only marginal scaling beyond 4 core CPUs.

You say this right after ignoring 4 core Skylake DDR4 stutter examples, it contradicts your argument, ignore it, pretend it doesn't exist and make the argument anyway.

Your V-Sync arguments are also completely bizarre, again your making blanket statements that fly in the face of the facts and basic common sense, these tests are not about 60 FPS, good grief don't you think that if these tests would have been about V-Sync they would have actually used V-Sync? with that what is the point in CPU performance testing with V-Sync on? that's complete and utter madness.

A 4 core CPU is good as long as you cap your Frame Rates to 60 FPS? i have a better one for you, if you're going to cap your Frame Rates to 60 don't bother with a GTX 1080/TI get a GTX 1050 instead you will have exactly the same performance, if that is your answer then why are you so defensive of this review, V-Sync was off in all of them, as it should be.

 

[trparky]

Um... @intrepid3d sounds like your typical Intel fanboy.

 

[intrepid3d]

A couple more examples of bad Intel 4 core performance that you would never know about just looking at slides.

 

[trparky]

Oy... this guy is going on my ignore list because he's spamming my alerts list.

 

[cucker tarlson]

A couple more examples of bad Intel 4 core performance that you would never know about just looking at slides.

View attachment 100146
View attachment 100147
View attachment 100148

View attachment 100149

Are you kidding me ? Those are momentary screen shots. Show us full video - if there's stutter we should be able to noticeat least some of it.
You're proving a point which is valid, but I gotta say the case you're making for it is pretty bad, you're not even basing it on whole benchmark runs, you're basing it on print screens. No wonder people ignore you. You don't read and make your case on passive agressive tone more than anything.

I admit my perspective is not very different from yours, but I ran 3570K with 980Ti OC and 4790K with 1080 OC, spent hundreds if not thousands of hours playing, I speak from my expereince. I don't even know whether watching a video showing a CPU bottleneck is able to really capture what you experience live and how the controls feel in real time, probably not, what I can tell you is that I sure as shit can't see stutter in a print screen image.

 

[Vayra86]

Oy... this guy is going on my ignore list because he's spamming my alerts list.

That's not my style, but I get ya. The plank of wood is too thick to get through here and I'm done. Here we have a guy ignoring half my posts, misinterpreting the rest and then convincing himself he's right. Meanwhile, all of the screenshots above show > 60 FPS game performance which supports my statements fully, running uncapped FPS while no one in any sort of scenario would ever use an i3 OR a Ryzen CPU because both fall off in high refresh scenario's where it counts, which is what the argument started with. It would be hilarious if it wasn't such a sad display of lacking intelligence.

All has been said on the topic, I'm done. When in doubt, re-read and think hard

One last hint: look at the last screenshot, where a 12 core CPU 'gets 12 thread action' but only manages to push out ~40 more FPS with 84% utilization, while the quad core CPU is 100% loaded. Both CPUs would offer the exact same gaming performance (remarkably, even on the fastest monitors money could buy right now, everything is over 240FPS), with the 12 thread CPU being grossly inefficient doing so.

Or take the screenshot of Ryzen 1700 vs 7700K with IDENTICAL FPS regardless of the additional core count.

Circling back to what I started with: reading comprehension.

You say this right after ignoring 4 core Skylake DDR4 stutter examples, it contradicts your argument, ignore it, pretend it doesn't exist and make the argument anyway.

Your V-Sync arguments are also completely bizarre, again your making blanket statements that fly in the face of the facts and basic common sense, these tests are not about 60 FPS, good grief don't you think that if these tests would have been about V-Sync they would have actually used V-Sync? with that what is the point in CPU performance testing with V-Sync on? that's complete and utter madness.

A 4 core CPU is good as long as you cap your Frame Rates to 60 FPS? i have a better one for you, if you're going to cap your Frame Rates to 60 don't bother with a GTX 1080/TI get a GTX 1050 instead you will have exactly the same performance, if that is your answer then why are you so defensive of this review, V-Sync was off in all of them, as it should be.

Stutters can occur due to high frame rate variance, so in real world scenario's, a frame cap serves as an effective way to remove stutter caused by huge variations in frame time. Other options are available, and almost every gamer would use one of them to improve the ingame experience. For high refresh rate gaming, such a method is Fast Sync + uncapped FPS, which is also extremely effective at removing stutter while maintaining optimal button-to-pixel latency. The other cause of stutter in games is when RAM or VRAM get saturated either in capacity or bandwidth, which is a type of stutter that manifests mostly below 60 FPS and what Haswell and older quad cores fall prey to.

About the GPU choice: you nearly got the point there, but then you went off in the wrong direction once more. Nobody is saying 'pair an i3 with a 1080ti'. You're right if 60 fps is your goal, which it is for a good 75-85% of the gaming crowd because they have a 60hz monitor, all of these CPUs perform 100% the same, and ONLY the CPUs with the highest clock speeds are a guarantee that you will never, or almost never (outliers still exist, even with a 5Ghz 8700K, speaking from experience not a Youtube video) dive below 60 FPS.

So, there is a good CPU for each user and use case. For the vast majority gaming at 60 FPS/hz, Ryzen or an i3 will net similar results, with the Ryzen 6c/12t offering more headroom than the i3 at the *same clock* but the i3-K CPU will provide more headroom once you OC it, which is where Ryzen can still fall short in more games than the i3 would due to the lack of cores. And for the high-end, Ryzen is again a fine choice, but now we're talking about 'what is the optimal choice and that is where Intel wins the day due to clocks on their 6-core offerings.

So - that REALLY was my last attempt.

 

[cucker tarlson]

That's not my style, but I get ya. The plank of wood is too thick to get through here and I'm done. Here we have a guy ignoring half my posts, misinterpreting the rest and then convincing himself he's right. Meanwhile, all of the screenshots above show > 60 FPS game performance which supports my statements fully, running uncapped FPS while no one in any sort of scenario would ever use an i3 OR a Ryzen CPU because both fall off in high refresh scenario's where it counts, which is what the argument started with. It would be hilarious if it wasn't such a sad display of lacking intelligence.

I feel sad for him. He actually has a hunch about cpu bottleneck, but he's struggling to prove it sooooo badly. I've been trying to help him but I doubt more than 10% of what I write gets through.

As for the person who used purepc.pl as an argument... those guys are definitely not testing games to find the most stresful parts. I do not know how or what they are doing, but I suspect it is not super competent.

Care to explain how you came to that conclusion ? You don't know but suspect what they're doing is not actually what they say they're doing but in fact you think they rather don't fully know what they're doing. Doesn't seem like a solid case of disproving anything to me. You're not coming across as super competent either with remarks like this.

 

[Charcharo]

Percent utilization is not a perfect representation of CPU resources guys. You are forgetting that. All of you !

 

[intrepid3d]

Are you kidding me ? Those are momentary screen shots. Show us full video - if there's stutter we should be able to noticeat least some of it.
You're proving a point which is valid, but I gotta say the case you're making for it is pretty bad, you're not even basing it on whole benchmark runs, you're basing it on print screens. No wonder people ignore you. You don't read and make your case on passive agressive tone more than anything.

I admit my perspective is not very different from yours, but I ran 3570K with 980Ti OC and 4790K with 1080 OC, spent hundreds if not thousands of hours playing, I speak from my expereince. I don't even know whether watching a video showing a CPU bottleneck is able to really capture what you experience live and how the controls feel in real time, probably not, what I can tell you is that I sure as shit can't see stutter in a print screen image.

Its not so much about stutter, the 4 core CPU is very heavily loaded and as a result the performance is simply better on the Ryzen 1600X. can you not see that in the screenshots? Metro LL for example the performance is around 20 to 40% higher on the 1600X

 

[cucker tarlson]

Sorry, but this is joker. The guy who claimed +20% performance on Vega with the fall win 10 update. And he MEASURED it. I mean how can a youtube video lie about numbers, says right in the title he's unbiased

 

[intrepid3d]

Sorry, but this is joker. The guy who claimed +20% performance on Vega with the fall win 10 update. And he MEASURED it. I mean how can a youtube video lie about numbers, says right in the title he's unbiased

No he didn't, You've just made that up.

You've just had a go at me for not providing the proof, now its your turn.

 

[cucker tarlson]

Lol don't make me pull out a thread I made based on his video and make me look exactly like the sort of a fool that you're making of yourself now.

Okay, challeange accepted.


https://www.techpowerup.com/forums/...-performance-improvement-in-game-mode.238018/

He pulled the video after being called on his crap. But he's a sensationalist more than anything. He likes clicks more than he likes work ethics, that's not hard to tell.

 

[trparky]

Sorry, but this is joker.

He's acting like your typical hard-headed Intel fanboy. Nothing you say will get through his concrete skull.

 

[Vayra86]

No he didn't, You've just made that up.

You've just had a go at me for not providing the proof, now its your turn.

You need to realize this is not a discussion, its us trying to teach you something about gaming performance, reviews, and reading the numbers right. You need to change your stance and as long as you don't you will never gain the understanding that's required here.

Also this is not something that works in the opposite direction as well, you don't need to show or tell me anything. I've seen it all. Not from a Youtube idiot, but from first hand experience with the actual hardware.

Percent utilization is not a perfect representation of CPU resources guys. You are forgetting that. All of you !

BINGO. +10 pts

 

[cucker tarlson]

He's acting like your typical hard-headed Intel fanboy.

I don't know what he is at this point but he certainly stole the whole show from any CPU in the discussion, be it intel or amd.

 

[intrepid3d]

Lol don't make me pull out a thread I made based on his video and make me look exactly like the sort of a fool that you're making yourself.

Okay, challeange accepted.


https://www.techpowerup.com/forums/...-performance-improvement-in-game-mode.238018/

All i see in that link is you making a claim about 20% higher performance in games with nVidia and AMD GPU's, with an unknown removed video link..

There is no proof there.

 

[trparky]

You need to realize this is not a discussion

Who? Me?

 

[cucker tarlson]

All i see in that link is you making a claim about 20% higher performance in games, with an unknown removed video link..

There is no proof there.

Read my whole post God dammit ! And read that thread.

 

[Vayra86]

Who? Me?

Nope, the user you ignored and what is now causing you to be confused This is why ignoring is quite the nuisance

 

[intrepid3d]

Read my whole post God dammit ! And read that thread.

I read your whole post, there are about 3 lines in it and none of it says anything at all about JP.

So now you want me to find the proof in a 4 page thread myself?

You made it up, the burden of proof is not with me, its with you, you proof it to me..