• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Leaked AMD Ryzen Benchmarks?

When we approached AMD with these results pre-publication, the company defended its product by suggesting that intentionally creating a GPU bottleneck (read: no longer benchmarking the CPU’s performance) would serve as a great equalizer. AMD asked that we consider 4K benchmarks to more heavily load the GPU, thus reducing workload on the CPU and leveling the playing field.
Click to expand...

Again from gamer's nexus. Another page I need to bookmark for future discussions about AMD and it's benevolence and exemplary track record of marketing honesty yadiyadayada
 
Ieldra said:
Again from gamer's nexus. Another page I need to bookmark for future discussions about AMD and it's benevolence and exemplary track record of marketing honesty yadiyadayada
Click to expand...

Yep, funny how AMD is never the evil company ;)

Same method as always.

116a.jpg
 
Anyway, all in all good showing from AMD. Seems to perform very competitively for productivity specifically video editing and CPU render but question of why performance is all over the place remains (not games)
 
razor1 said:
Max oveclock on a Noctuna 14, what was ball's saying, he was going to get 4.2-4.4 ghz lol, they got 3.9!
Click to expand...
Something else is up with that tbh.
It seems to need unusually high voltage for the 4GHz.
Too early to say whether this is immature firmware/code or more inherent to the CPU.
Worth noting the 6900K needs around 1.2V for 4GHz, while 1800X needs anywhere from 1.35V to 1.425V it seems.

But yeah I guess until the issue is resolved (and if it is simple for AMD/motherboards to rectify) even 4.2GHz is not going to happen even on water IMO (sustainable and usable context), let alone air.
Cheers
 
I'm slightly confused about the clocks to be honest, how are some reviews stuck at 3.95ghz when XFR is supposed to go to 4.1

Is XFR single core only? Does it only work for short bursts?
 
Ieldra said:
I'm slightly confused about the clocks to be honest, how are some reviews stuck at 3.95ghz when XFR is supposed to go to 4.1

Is XFR single core only? Does it only work for short bursts?
Click to expand...
They are being hit by the high voltage and so impacting XFR.
Needs to be seen if that voltage-frequency-core behaviour is correct or to do with the immature state of the whole platform, going by Broadwell-E one would think something is not right for now with the launched Ryzen from a HW-platform perspective IMO, needs to seen if this can be corrected.
Cheers
 
Ieldra said:
I'm slightly confused about the clocks to be honest, how are some reviews stuck at 3.95ghz when XFR is supposed to go to 4.1

Is XFR single core only? Does it only work for short bursts?
Click to expand...


Its all variable, XFR is just like auto boosting in GPU's depending on the applications, depending on exterior conditions.....
 
razor1 said:
Its all variable, XFR is just like auto boosting in GPU's depending on the applications, depending on exterior conditions.....
Click to expand...
Fair enough, as for the voltages it was my understanding that they are being misrepresented by the bios. Are the chips actually running 1.4v out the box?
 
don't know, have to wait and see, but as CSI PC stated, something is up with the voltage, Voltage is upped to counter act leakage, and using that much at the frequencies they are getting looks like something is off.

Shouldn't say counter act, should say overcome.
 
Not saying it is related (as we do not have enough info yet) just as an example that highlights how it can influence.
Remember the very 1st z270 boards that automatically increased the voltage too much and caused temperature problems for Kaby Lake in its very early days and out of the box.
Cheers
 
And put that TJMax 'throttling' into context is the need for 1.22V to 1.24V for just 3.7GHz with Blender and critically minimum of 1.395V for 3.7GHz when Cinebench 1T, also as reported by GamersNexus in the Ryzen vid.

In comparison Broadwell-E can do around 1.2V for 4GHz.

21-6900K-Voltage.png



21-6950X-Voltage.png


Der8aeur shows the same for his 6950X all cores stressed at 4GHz using 1.2V.

So the Ryzen TJmax 'throttling' is probably only indirectly part of the issue, comes back to voltage behaviour-spec relative to performance-clocks.
Cheers
 
Last edited:
razor1 said:
There are multiple reviews showing the same thing, yeah you can remove an outlier here and there, but still that doesn't change the facts.
Click to expand...

You refuse to deal with the impossibility of the FX 8370 being better in these games than Ryzen which has 52% IPC advanatage and SMT advantage over FX8370. Until you explain this you all are full of crap as far as I am concerned.
 
15% is acceptable it puts it around Broadwell levels. I assume that is mostly for single core games. If a game tested favors Kaby Lake when more than 8 threads are involved it was not tested properly period. If you don't think some of these reviews you clowns are posting are not skewed by reviewers getting favors from Intel you are grossly naive. I guess razor 1 is only a gamester although he sounds a little to mature to use his compouter only for games. Where are the reviews with productivity apps. Don't throw those shit ptograms like winrar and seven zip intothe equation . They are antiquated . Give me a winzip test. You clowns only show the few pathetic apps that Intel does better in you are fighting a losing battle. 8 core
cores are the future of computer use and gaming. As soon as more Vulcan-base games appear like Bethesdas inventory the other game developers will fall in line. Then you can beat your breasts ragged and you will whistling in the wind. Pathetic. By the way how much Intel stock do you own????
 
os2wiz said:
You refuse to deal with the impossibility of the FX 8370 being better in these games than Ryzen which has 52% IPC advanatage and SMT advantage over FX8370.
Click to expand...
Where? Not a single review i has seen put FX anywhere above dumpster.
os2wiz said:
he sounds a little to mature to use his compouter only for games.
Click to expand...
From what you can read, he has a separate dual socket rig for actual work. Ryzen does not exactly constitute a good workstation for people who are not that budget constrained.
os2wiz said:
15% is acceptable it puts it around Broadwell levels. I assume that is mostly for single core games. If a game tested favors Kaby Lake when more than 8 threads are involved it was not tested properly period. If you don't think some of these reviews you clowns are posting are not skewed by reviewers getting favors from Intel you are grossly naive.
Click to expand...
The only one who is grossly naive and thinking AMD have not done more spinning so far is you. With lack of CS knowledge too.
 
Shintai said:
Ryzen trhottles at 75C it seems.
Click to expand...

What does that prove. I never ran my FX9590 over 67 celsius under load. I water cool. Your pathertic Intel chips burn too much energy especially Kaby Lake. You proved how badly designed Intel Kaby lake is
So clown explain how an lowly FX8370 did better in the games and apps you displayed than a Ryzen 1800X. Until you give an acceptable explanation, I say some of these tests you posted are flawed in methodology and possibly fraudulent. Not on your part on the part of the person who conducted the test.
Post all system configuration data if you want credibility.
 
Last edited:
os2wiz said:
So clown explain how an lowlu FX870 id better in the games and apps you displayed than a Ryzen 1800X. Until you give an acceptable explanation, I say the tests are flawed in methodology.
Click to expand...
Before throwing eggs, you gotta pick one up, so where is the screen you are talking about.
 
lolfail9001 said:
Where? Not a single review i has seen put FX anywhere above dumpster.

From what you can read, he has a separate dual socket rig for actual work. Ryzen does not exactly constitute a good workstation for people who are not that budget constrained.

The only one who is grossly naive and thinking AMD have not done more spinning so far is you. With lack of CS knowledge too.
Click to expand...

You are blind then. Earlier today in the morning Shintai posted a bunch of tests showing various cous. In a couple of these tests the FX 8370 out performed Ryzen 1800X. That is sn impossibility and yet the clown is yet to exlain why.
 
os2wiz said:
Earlier today in the morning Shintai posted a bunch of tests showing various cous.
Click to expand...
I was in class at the time, report them. Besides, it is actually possible to create a scenario in which Ryzen will perform worse than Bulldozer, by abusing CCX interconnect.
os2wiz said:
In a couple of these tests the FX 8370 out performed Ryzen 1800X. That is sn impossibility and yet the clown is yet to exlain why.
Click to expand...
How far was 7700k ahead?
 
os2wiz said:
You refuse to deal with the impossibility of the FX 8370 being better in these games than Ryzen which has 52% IPC advanatage and SMT advantage over FX8370. Until you explain this you all are full of crap as far as I am concerned.
Click to expand...


I just said remove the outliers, that's ok, no big deal there are many more reviews that show what Ryzen is truly at.
 
os2wiz said:
15% is acceptable it puts it around Broadwell levels. I assume that is mostly for single core games. If a game tested favors Kaby Lake when more than 8 threads are involved it was not tested properly period. If you don't think some of these reviews you clowns are posting are not skewed by reviewers getting favors from Intel you are grossly naive. I guess razor 1 is only a gamester although he sounds a little to mature to use his compouter only for games. Where are the reviews with productivity apps. Don't throw those shit ptograms like winrar and seven zip intothe equation . They are antiquated . Give me a winzip test. You clowns only show the few pathetic apps that Intel does better in you are fighting a losing battle. 8 core
cores are the future of computer use and gaming. As soon as more Vulcan-base games appear like Bethesdas inventory the other game developers will fall in line. Then you can beat your breasts ragged and you will whistling in the wind. Pathetic. By the way how much Intel stock do you own????
Click to expand...


Its not at Broadwell levels with IPC, at least not what we have seen thus far, need to see whats up with specific apps/games to get a better picture. Is it better than before, hell ya, but to sit here and deny that all the reviews got it wrong and only accept what AMD has shown us in their controlled tests, is far from the truth.

I do a ton of 3d rendering and special effects, but I already have a dual Broadwell system, so yeah Ryzen if I was in the market for a dev or video editing or systems like that for cheap I would be all over it if it had dual cpu's (will never go back to single cpu's for the things I do), but I'm not.
 
razor1 said:
Its not at Broadwell levels with IPC, at least not what we have seen thus far, need to see whats up with specific apps/games to get a better picture. Is it better than before, hell ya, but to sit here and deny that all the reviews got it wrong and only accept what AMD has shown us in their controlled tests, is far from the truth.

I do a ton of 3d rendering and special effects, but I already have a dual Broadwell system, so yeah Ryzen if I was in the market for a dev or video editing or systems like that for cheap I would be all over it if it had dual cpu's (will never go back to single cpu's for the things I do), but I'm not.
Click to expand...
What I am saying is there were bios updates as late as yesterday that would have improved memory performance. Did any of these reviewers update their bios to give a more accurate evaluation of Ryzen? I don't think so.
 
os2wiz said:
What I am saying is there were bios updates as late as yesterday that would have improved memory performance. Did any of these reviewers update their bios to give a more accurate evaluation of Ryzen? I don't think so.
Click to expand...

That is a only part of the problem, that is not where the problem lies with Ryzen's gaming performance at all.

Bios, Microcode, all that stuff is not going to fix this problem from the looks of it. There seems to be specific architectural problems that is the underlying cause we have benchmarks with motherboards that did have higher mem bandwidth still exhibiting the same problems.
 
Might make more of a difference than you think. I'm not expecting miraculous improvements but then again you don't know until you try. From the other thread:

WhoBeDaPlaya said:
Sure! Pretty informative read for me :
https://forums.anandtech.com/threads/ryzen-strictly-technical.2500572/
Click to expand...

:wideyed:
The northbridge of Zeppelin is officially called as the data fabric (DF). The DF frequency is always linked to the operating frequency of the memory controller with a ratio of 1:2 (e.g. DDR4-2667 MEMCLK = 1333MHz DFICLK). This means that the memory speed will directly affect the data fabric performance as well. In some cases, it may appear that the performance of Zeppelin scales extremely well with the increased memory speed, however that is necessarily not the case.

In many of these cases the abnormally good scaling is caused by the higher data fabric clock (DFICLK) resulting from the higher memory speed, rather than the increased performance of the memory itself.
Click to expand...

So slow RAM makes the L3 latency between CCXs go up. Starting to see a pattern here...
 
ecmaster76 said:
Might make more of a difference than you think. I'm not expecting miraculous improvements but then again you don't know until you try. From the other thread:



:wideyed:


So slow RAM makes the L3 latency between CCXs go up. Starting to see a pattern here...
Click to expand...


Well that is why we should take those reviews out of the mix cause yeah those reviews are the ones that really show weird results. But even without those reviews, you still have a massive gap in performance that was kinda of unexpected at least to the degree that we are seeing.
 
Last edited:
ecmaster76 said:
Might make more of a difference than you think. I'm not expecting miraculous improvements but then again you don't know until you try. From the other thread:



:wideyed:


So slow RAM makes the L3 latency between CCXs go up. Starting to see a pattern here...
Click to expand...

I also saw a review showing very significant gains in L3 bandwidth when using higher speed memory, can't for the life of me remember which though
 
razor1 said:
That is a only part of the problem, that is not where the problem lies with Ryzen's gaming performance at all.

Bios, Microcode, all that stuff is not going to fix this problem from the looks of it. There seems to be specific architectural problems that is the underlying cause we have benchmarks with motherboards that did have higher mem bandwidth still exhibiting the same problems.
Click to expand...

Baloney, it is NOT architectural at all. The games need optimization for Ryzen the same way developers optimize their code for Intel. It can be done and it will be done through Vulcan. Bethesda one of the biggest developers announced collaboration with AMD to optimize the entirety of their large stock of games. Other developers will follow. It is in their interest to do so. Once that ball starts rolling you will see that there will be little to complain about Ryzen. The only real critique of Ryzen is that it has limited overclocking. That will improve in the next stepping release. The power of the Ryzen cores is a great achievement and the cache is superb. There will be no need to turn off cache for performance improvements once optimization occurs.
 
razor1 said:
Well that is why we should take those reviews out of the mix cause yeah those reviews are the ones that really show weird results. But even without those reviews, you still have a massive gap in performance that was kinda of unexpected at least to the degree that we are seeing.
Click to expand...


This is strictly a bios issue that will be resolved in the next 3 to 4 weeks. Then faster ram speeds for 4 modules will work and game scores will increase by 10 to 15%. The remaining deficits will be erased with optimization with Vulcan. AMD is lining up additional developers to collaborate on this venture. Bethesda was the first, it will not be the last.
 
Ieldra said:
Anyway, all in all good showing from AMD. Seems to perform very competitively for productivity specifically video editing and CPU render but question of why performance is all over the place remains (not games)
Click to expand...

That is, RyZen shines in workloads where a GPU is better.

Ieldra said:
I'm slightly confused about the clocks to be honest, how are some reviews stuck at 3.95ghz when XFR is supposed to go to 4.1

Is XFR single core only? Does it only work for short bursts?
Click to expand...

4.1GHz is the single-core XFR boost for the 1800X.
3.9GHz is the single-core XFR boost for the 1700X.
 
razor1 said:
That is a only part of the problem, that is not where the problem lies with Ryzen's gaming performance at all.

Bios, Microcode, all that stuff is not going to fix this problem from the looks of it. There seems to be specific architectural problems that is the underlying cause we have benchmarks with motherboards that did have higher mem bandwidth still exhibiting the same problems.
Click to expand...

We have known Ryzen would perform like an i5 in games since CanardPC tested the engineering sample

http://www.guru3d.com/news-story/fr...ng-sample-amd-ryzen-processor-benchmarks.html

Just before launch CanardPC got access to a qualification sample for the 1800X and got the same conclusion: bad performance in games and latency-sensitive workloads. They trace the real problem to the memory subsystem


and the rumor is that this problems is expected to be solved for Zen+. Meanwhile AMD will work closely with game developers to optimize the titles to extract all performance possible to minimize the real issue or will 'suggest' reviewers to test games only with GPU-bound settings to hide the real problem

http://www.gamersnexus.net/hwreview...review-premiere-blender-fps-benchmarks/page-7
 
os2wiz said:
This is strictly a bios issue that will be resolved in the next 3 to 4 weeks. Then faster ram speeds for 4 modules will work and game scores will increase by 10 to 15%. The remaining deficits will be erased with optimization with Vulcan. AMD is lining up additional developers to collaborate on this venture. Bethesda was the first, it will not be the last.
Click to expand...

os2wiz said:
Baloney, it is NOT architectural at all. The games need optimization for Ryzen the same way developers optimize their code for Intel. It can be done and it will be done through Vulcan. Bethesda one of the biggest developers announced collaboration with AMD to optimize the entirety of their large stock of games. Other developers will follow. It is in their interest to do so. Once that ball starts rolling you will see that there will be little to complain about Ryzen. The only real critique of Ryzen is that it has limited overclocking. That will improve in the next stepping release. The power of the Ryzen cores is a great achievement and the cache is superb. There will be no need to turn off cache for performance improvements once optimization occurs.
Click to expand...

The memory speeds yes but not the CCX issue.

I can sit here and go through it with you but it will be above you head, so far you are still in denial in everything about Ryzen, so its not even worth the time or effort.
 
The CCX complex is easily fixed by optimizing the windows scheduler. There is no doubt that AMD and Microsoft will collaborate to fix this in the next couple of months. You are trying to say these probllems are fixed in stone and they are not. New architectures inevitably face problems like these and kinks like this one are relatively easily fixable. Microsft has close working relationship with AMD, especially with XBox and Scorpio. This will be resolved and faster than you think.


razor1 said:
The memory speeds yes but not the CCX issue.

I can sit here and go through it with you but it will be above you head, so far you are still in denial in everything about Ryzen, so its not even worth the time or effort.
Click to expand...
 
You must log in or register to reply here.
Back
Top