Report your benchmark results here! #8
Comments
|
I'd love to add to the benchmarks list. I've got two questions:
Cheers! |
|
Hi ibonito1, OpenCL support on EPYC CPUs is a bit difficult as these are not officially supported by AMD. Being x86-64, they should work with the Intel OpenCL CPU Runtime though, or alternatively with POCL. Fingers crossed! Regards, |
|
AMD Radeon RX 580: |
|
C-Dub2022 thank you very much for the RX 580 benchmark! If you can post the FP16S and FP16C benchmarks as well, I'll add them to the readme! |
|
Hopefully this is helpful. Let me know if there is anything else I can do.
|
|
RTX 3060 12GB - v1.1 FP32-FP16C FP32-FP16S FP32-FP32 |
|
MarcoAurelioFerrari thank you! |
|
Could you please tell me how to open the visualized interface of the flow domain as you said in the readme file? You said input the 2 can turn on the velocity field, but it does not work in the benchmark case. How can I generate pictures like you prensent on twitter ? |
|
Hi dongwang22, thanks for the benchmark! For the visual interface, uncomment Regards, |
|
AMD 5700 XT |----------------.------------------------------------------------------------| |----------------.------------------------------------------------------------| |----------------.------------------------------------------------------------| |
|
FP32/FP16C FP32/FP16S FP32/FP32 |
|
RTX 2080 SUPER |
|
I tried a 6900XT, but the score is lower than anticipated. The max bandwidth seems to be limited to 300GB/s, although GPUZ says it's connected via PCIe 4.0 16x and should top out at 512GB/s. The GPU clock is at 2540MHz and the memory clock at 2000MHz. GPU and memory controller loads are at 100%.
With the 3D Taylor-Green model and FP32/FP16S, the MLUPs/s and the bandwidth go through the roof. I'll try some other models, too. FP32/FP32 goes up to 2400 MLUPs/s and 370GB/s, with FP32/FP16C it's 9000 MLUPs/s and 700GB/s. |
|
Hi gittigittibangbang, thanks for the benchmarks! Efficiency is ~60% which is typical for the AMD GPUs. Performance is limited by VRAM bandwidth only, and the RX 6800 would presumably perform exactly the same. The benchmark setup is a 256³ box, that fills 1.5GB (FP32) or 0.9GB (FP16) of VRAM. The large infinity cache (128MB) is only an insignificant fraction of that so does not significantly boost performance. |
|
Vega 8 in R7 4750G |----------------.------------------------------------------------------------| |----------------.------------------------------------------------------------| |
|
Is it possible to add ready-to-run benchmark for MacOS so we can get more result on Mac? |
|
RTX3060 Laptop GPU with 12700H on ASUS ROG M16 Turbo mode (120W GPU TDP) and external laptop fan PS C:\Software\FluidX3D> .\FluidX3D-Benchmark-FP32-FP16C-Windows.exe PS C:\Software\FluidX3D> .\FluidX3D-Benchmark-FP32-FP16S-Windows.exe PS C:\Software\FluidX3D> |
|
@HAL9000COM thanks for the Vega 8 benchmarks! Quick question: Is your RAM is 2x16GB DDR4-3200MT/s? And do you have an idea why the GPU shows up 3 times? |
@edmond1992 unfortunately I don't have a Mac, so I can't compile add the executables for MacOS. But the code should work as-is; just compile it as-is with the third line in |
|
Cross compile?
…Sent from my iPhone
On 23 Oct 2022, at 16:06, Moritz Lehmann ***@***.***> wrote:
Is it possible to add ready-to-run benchmark for MacOS so we can get more result on Mac? Especially the test is bandwidth limited and Apple silicon should be good at this. Not to mention relatively cheap 64GB+ VRAM as they share the same main memory.
@edmond1992<https://github.com/edmond1992> unfortunately I don't have a Mac, so I can't compile add the executables for MacOS. But the code should work as-is; just compile it as-is with the third line in make.sh and you'll get the FP32 benchmark. Uncomment FP16S/FP16C in src/defines.hpp and recompile to get the other 2 benchmarks.
—
Reply to this email directly, view it on GitHub<#8 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ALNCZ732JE6QCL7RKB37QG3WETWXJANCNFSM6AAAAAAQTGZVRY>.
You are receiving this because you were mentioned.Message ID: ***@***.***>
[https://www.polyu.edu.hk/emaildisclaimer/85A-PolyU_Email_Signature.jpg]
Disclaimer:
This message (including any attachments) contains confidential information intended for a specific individual and purpose. If you are not the intended recipient, you should delete this message and notify the sender and The Hong Kong Polytechnic University (the University) immediately. Any disclosure, copying, or distribution of this message, or the taking of any action based on it, is strictly prohibited and may be unlawful.
The University specifically denies any responsibility for the accuracy or quality of information obtained through University E-mail Facilities. Any views and opinions expressed are only those of the author(s) and do not necessarily represent those of the University and the University accepts no liability whatsoever for any losses or damages incurred or caused to any party as a result of the use of such information.
|
2x32GB DDR4-3200 OC to 3533. No idea why GPU shows up multiple times. After some reboot, it now shows up as two devices. |
|
Nvidia Quadro RTX 5000 |
|
GTX 1050 on an old gaming laptop. It's amazing I figured out how to even run this and get a benchmark. Now I'm going to try and figure out how to run the simulation on an stl (or similar) file. I know how to use Blender quite well, but this is my first time with visial studio or command line stuff. I'm so out of my depth here 😟
|
|
@ProjectPhysX have now added the FP16 benchmarks RTX 3080 TiUpdated FP32 (was concurrently baking a fluid in Blender when I ran the last one): FP16S: FP16C: |
|
Hi @SirWixy, thank you so much for the benchmarks! Can you post the FP16S and FP16C results too? |
|
Quadro RTX 4000 below. I also tried two Xeon Gold 5218 (2x16 cores), with the FP32/FP32 benchmark they top out at 126MLUPs/s, 20GB/s and 8 steps/s. I did not have the patience to run it to the end. The speedup with GPUs is really dramatic, damn.
|
|
@gittigittibangbang thanks for the benchmarks! For the CPU you can just stop it with Ctrl+C after it has leveled at constant performance, and take the last MLUPs/s reading. Can you post the program header with the Xeon Gold for the specs, and performance values for FP16S and FP16C too for the Xeon? Thanks! |
|
|----------------.------------------------------------------------------------| FP32/FP32: 132MLUPs/s, 20GB/s bandwidth, 8 steps/s |
|
Results for the ASUS 4070 Ti Super 16GB card, non overclocked.
|
|
RTX 3080 12GB edition - FP16S RTX 3080 12GB edition - FP16C RTX 3080 12GB edition - FP32 |
|
Here's a multi GPU (technically) result for a Tesla K80 (2 core) GPU. There's a single core K80 (12GB) result in the benchmarks, but now that we have multi GPU functionality here's the 2 core K80 (24GB) result! |
|
|
Results on AMD Radeon RX590 8GB (Running on Clover-Mesa OpenCL 1.2) FP32 FP16C FP16S |
|
Results on AMD Radeon RX590 8GB (Running on Rusticl-Mesa OpenCL 1.2) FP32 FP16C FP16S So if you want to use an OpenSource OpenCL implementation (Clover or Rusticl) use Clover until Rusticl become better. Clover by default is OpenCL 1.1 conformant, but you can export:
to use OpenCL 1.2 |
|
RoG Strix Laptop:
Interesting how my Laptop 3080 Ti beats the other Laptops RTX 4080 ! |
|
Hi @gitcnd, thanks a lot! Can you please add the FP16S and FP16C benchmarks too? |
|
Sorry about that - here they are: |
|
And just for giggles... (the slowest benchmark here so far :-) |
|
@gitcnd Are both DIMM slots on the laptop populated for the Intel iGPU benchmark? If not, the results would be even slower... 😄 |
|
Yes - everything is populated and replaced for max performance (including special low-latency RAM: I replaced the originals).
This was the fastest laptop in the world when I finished upgrading it :-) |
|
RTX6000ADA // Without-ECC
and
|
|
Ryzen 7 7800X3D, FP16S Ryzen 7 7800X3D, FP16C Ryzen 7 7800X3D, FP32 |
|
Hi @roktmanskip, thanks a lot! That's the AMD Radeon Graphics iGPU. What memory speed are you running, and is it 2x 8GB dual channel? Can you please test the CPU itself as well? I'm curious how it performs. For this, install the Intel CPU Runtime for OpenCL, and then starting the executables from within CMD with the device ID:
Thanks! |
|
2x16Gb, 6400MHz
|
|
Orange Pi 5, with Rockchip RK3588S, ARM based SOC.
|
I don't know how to enable running on the CPU cores, instead of the default Mali GPU. |
|
NVIDIA RTX A1000 6GB Laptop GPU FP32-FP16C FP32-FP32 FP32-FP16S |
|
@gurkanctn can you please post Orange Pi benchmarks for FP16S and FP16C too? (enable those in src/defined.hpp) Thanks! |
|
NVIDIA RTX 3060 Laptop FP32-16S FP32-16C |
|
@ProjectPhysX , I'll try to do a fair benchmark (the previous one had some background applications running) for the cases and report. Thanks for the software. I started playing with it, and liked it a lot. It would be great if pressure visualization on the walls could be easier. |
|
Orange Pi 5 benchmark results for FP32, FP16C, and FP16S.
|
|
NVIDIA GeForce RTX 2070 SUPER FP16C: FP16S: FP32 |
|
AMD Radeon RX 5700 FP16S: FP16C: FP32: |
|
AMD Radeon RX 580: |
|
AMD Ryzen 7 5700X3D: |
|
@aervnu thanks! What memory speed are you running? 3200 MT/s? |
Yeah, I originally wanted to buy a 3600 kit but realized they were a tad bit more expensive than 3200 and since I figured that I wouldn't need 3600 I ended up with this thing here. I can definitely try to overclock it to 3600 still, if the 5700X3D's IMC blesses me. |
and others
You are welcome to report your benchmark results for the FP32/FP16S/FP16C accuracy levels here.
Especially numbers for AMD GPUs are desired for GCN/RDNA/RDNA2 architectures.
Thank you!
The text was updated successfully, but these errors were encountered: