Post Your BaseMark GPU Scores!

[T4C Fantasy]

The bottom entry for Buzznoob is correct, but there is another one higher up the table that is wrong.

done thats a great way to find errors atleast

 

[johnspack]

Well, I'm going to throw a new rock into the pond... linux! First of all, had to fix permissions on the installed folder, or else it would lock up my os. Vulkan seems to suck compared to opengl on my maxwell card, or is it linux? Cpu is at 4.7ghz, gpu is at +55 which is maybe 1500mhz it's saying, and no ram oc, not available under linux. Arg. And cpu-z only partially works, and gpu-z not at all, so I've used linux equivalents.

Guess this will be added to my last post... I just figured out how to oc my card sorry hadn't really tried under linux yet... also didn't know you would only count the vulkan results:

Sorry again, realized you do score opengl, this is a more correct run..

 

[Tomgang]

Alright back with a new run after fine tunning GPU clocks a bit higher and oc CPU to 4.75 GHz

Vulkan. Just as i already knew, here GPU where already pretty much maxed out in the first runs. So only a few points better score. The limiting factor here is the GPU´s low power target at only 120 % max. So GPU hits power target often and force GPU to clock down. With a higher power target, i bet i cut get a few points more.

OpenGL. Again as i said in my early post, here CPU is the bottleneck. Just see score after a higher OC score clime significantly, but is still lower than other systems with stronger CPU´s so still a bottleneck here.

Hardware info

In the end X58 still handles a powerful GPU pretty well. Not as great as a new CPU, The OpenGL test clearly shows that. But for how old X58 is becoming, X58 and us owners has nothing to be a shame of up until now and properly a year or two more.

 

[JalleR]

OpenGL is the best for Mali GPU's

Exynos 9810 Mali-G72 MP18

 

[johnspack]

Sorry yet again!.... Yes you can oc vram under linux, and there is more headroom than I thought for both clocks, so final update for my gtx970:

 

[Nuckles56]

Well I tried to bench this morning in the -3 conditions outside (I managed to get the core temp of the GPU down to 2 degrees at desktop), but I couldn't manage to get a score higher than the 3993 point run I already had sadly. My RX 480 just runs out of overclocking potential under air at the 1450MHz core mark sadly and I can't push it any harder. Maybe under water it might manage a little more but I doubt it.

 

[johnspack]

What I would like to see is someone who is still dual booting linux and windows to post results from both versions with the same hardware.
I killed my windows installs a while back, so I can't do it....

 

[xkm1948]

After the fun of DDU'ing and reverting back to 18.5.1 drivers, finally got this running in vulkan on my Vega 64. Posting results from OpenGL and from Vulkan, but it's pretty clear something's very wrong with the OpenGL results. In Vulkan, my GPU utilization stays near-continuously pegged at 99%, and the clocks are pegged in the 1600s (see hwinfo screen snip from that session- Avg GPU Utz is 91% in spite of the dead period as the benchmark is loading, average clock is 1543, etc). In the OpenGL session, however, the utilization is terrible - average under 70%, and because it's not pushing the GPU the clocks stay low as well (avg just under 1200, and never even bumps into the 1500 range).

Once again nothing wrong with your OpenGL. GCN based cards sucks at OpenGL and excels at Vulkan.

https://www.overclock3d.net/reviews/gpu_displays/basemark_gpu_performance_review/3


As you can see both Vega and Polaris works better under Vulkan

 

[mouacyk]

5/4.7 8700K, 4266-17-18-18-38-2T, 1080Ti 2101/12627

 

[xkm1948]

Who is gonna break 10K first? I wonder.

 

[T4C Fantasy]

Who is gonna break 10K first? I wonder.

@NicklasAPJ

Im in bed atm not adding scores til morning

 

[defiiancecp]

Once again nothing wrong with your OpenGL. GCN based cards sucks at OpenGL and excels at Vulkan.

You keep saying that, and I'm going to keep knowing it's not correct - repetition won't change it so let's agree to disagree. This is clearly not a matter of poor GPU performance: The GPU is at extremely low utilization and clocks. If this was a result of the GPU being bad at OpenGL, it would show high utilization and still relatively low results. That's not what we're seeing. Your graphs aren't contradicting anything, they're just restating the problem: We already know OpenGL scores are low. I don't think it's the benchmark, I think something's wrong with the drivers (as supported by the fact that this is seen in other OpenGL apps, examples included in the AMD community issue discussion I linked before), but saying nothing's wrong simply ignores the low GPU utilization. I mean, Radeon *is* weaker at OpenGL than vulkan (as in the Doom example), but Doom also shows that OpenGL *CAN* exercise the GPU (because GPU Utilization is very high during the doom OpenGL run). So yes, something is wrong.

Still, it's not really important since basically nothing uses OpenGL anyway...

 

[biffzinker]

You keep saying that, and I'm going to keep knowing it's not correct - repetition won't change it so let's agree to disagree. This is clearly not a matter of poor GPU performance: The GPU is at extremely low utilization and clocks. If this was a result of the GPU being bad at OpenGL, it would show high utilization and still relatively low results. That's not what we're seeing. Your graphs aren't contradicting anything, they're just restating the problem: We already know OpenGL scores are low. I don't think it's the benchmark, I think something's wrong with the drivers (as supported by the fact that this is seen in other OpenGL apps, examples included in the AMD community issue discussion I linked before), but saying nothing's wrong simply ignores the low GPU utilization. I mean, Radeon *is* weaker at OpenGL than vulkan (as in the Doom example), but Doom also shows that OpenGL *CAN* exercise the GPU (because GPU Utilization is very high during the doom OpenGL run). So yes, something is wrong.

Still, it's not really important since basically nothing uses OpenGL anyway...

The problem is the OpenGL driver going back all the way to ATI before the merge. Nvidia's OpenGL ICD has always performed better on Windows.

 

[Zyll Goliat]

The problem is the OpenGL driver going back all the way to ATI before the merge. Nvidia's OpenGL ICD has always performed better on Windows.

Sure no doubt about that...however this is clearly such a HUGE gap in results that seems like something else causing some additional issue....Just look that Vega 64 result in Vulkan GPU scores 6333 but in Open Gl score is 3254.....I mean that´s like almost 100% difference....weird.....

 

[fray_bentos]

Sure no doubt about that...however this is clearly such a HUGE gap in results that seems like something else causing some additional issue....Just look that Vega 64 result in Vulkan GPU scores 6333 but in Open Gl score is 3254.....I mean that´s like almost 100% difference....weird.....

...which is not *that* different from the 100fps reported for Doom under Vulcan vs. 66fps with OpenGL quoted earlier in this thread.

 

[Zyll Goliat]

...which is not *that* different from the 100fps reported for Doom under Vulcan vs. 66fps with OpenGL quoted earlier in this thread.

Well...it is different...this is even + 20 %(on top of that) worst result then your DOOM example and thats quite a lot in my eyes........

 

[Hail the Brain Slug]

Who is gonna break 10K first? I wonder.

I did, but it was the first day the benchmark was out and I didn't take a screenshot with GPU-z/CPU-z. I'll have to rerun when I have a chance.

5960x at 4.7, DDR4 3200, 1080ti FTW3 at some settings I don't even remember, all in a custom loop. The GPU clocks were definitely unstable territory.

This is my daily/gaming/stable setting. Same CPU settings but GPU is at 2088 core, 12,410 mem.

 

[Sirillya]

Second !

 

[xkm1948]

Damn nice! How about some TitanXp or TitanV? lol

 

[T4C Fantasy]

Damn nice! How about some TitanXp or TitanV? lol

Nick has the Xp

 

[xkm1948]

You keep saying that, and I'm going to keep knowing it's not correct - repetition won't change it so let's agree to disagree. This is clearly not a matter of poor GPU performance: The GPU is at extremely low utilization and clocks. If this was a result of the GPU being bad at OpenGL, it would show high utilization and still relatively low results. That's not what we're seeing. Your graphs aren't contradicting anything, they're just restating the problem: We already know OpenGL scores are low. I don't think it's the benchmark, I think something's wrong with the drivers (as supported by the fact that this is seen in other OpenGL apps, examples included in the AMD community issue discussion I linked before), but saying nothing's wrong simply ignores the low GPU utilization. I mean, Radeon *is* weaker at OpenGL than vulkan (as in the Doom example), but Doom also shows that OpenGL *CAN* exercise the GPU (because GPU Utilization is very high during the doom OpenGL run). So yes, something is wrong.

Still, it's not really important since basically nothing uses OpenGL anyway...

Go download Unigine Superposition benchmark. Run 4K optimized first in DrectX11 then in OpenGL. Compare the score yourself.

OpenGL an DX11 requires A LOT OF driver optimization which RTG has little resources to devote to. DX12 and Vulkan on the other hand requires little driver optimization because they can saturate whatever GPU they run on.

Fiji/Vega has 4096 SPs. They don't lack the raw resources. What they lack is efficient utilization of those resources. GCN has always been heavy on the raw power instead of focusing on the overall balancing of the structure. That is why under Vulkan GCN based cards really shine.

 

[Anzlew]

My results, with slightly OC'ed GPU:

 

[defiiancecp]

Go download Unigine Superposition benchmark. Run 4K optimized first in DrectX11 then in OpenGL. Compare the score yourself.

OpenGL an DX11 requires A LOT OF driver optimization which RTG has little resources to devote to. DX12 and Vulkan on the other hand requires little driver optimization because they can saturate whatever GPU they run on.

Fiji/Vega has 4096 SPs. They don't lack the raw resources. What they lack is efficient utilization of those resources. GCN has always been heavy on the raw power instead of focusing on the overall balancing of the structure. That is why under Vulkan GCN based cards really shine.

Yet another example proving my point - Superposition 4k OpenGL, GPU utilization nearly100%, clocks consistently maxed, while this benchmark averages high 60's in utilization with clocks consistently extremely low.

YET AGAIN I REPEAT: I'm not arguing that OpenGL results shouldn't be lower. They should, and will. But right now that's being overstated, because the GPU *ISN'T BEING UTILIZED* by this benchmark.

 

[RealNeil]

CPU is running at 4.7GHz. under load.
Maybe tomorrow It's coming apart to swap an 8700K into this case.
Will post again then.

 

[T4C Fantasy]

View attachment 102989

You need to show cpu and gpuz