log in |
Message boards : Number crunching : Nvidia Tesla P100 Problems
Previous · 1 · 2 · 3
Author | Message |
---|---|
was v31 compiled with explicit support for CC_6.0 in the NVCCFLAGS section of your makefile or build script? yes it was but with cc5 | |
ID: 9770 · Rating: 0 · rate: / Reply Quote | |
you need to add support for every cc level explicitly. --generate-code arch=compute_50,code=sm_50 --generate-code arch=compute_52,code=sm_52 --generate-code arch=compute_60,code=sm_60 --generate-code arch=compute_61,code=sm_61 --generate-code arch=compute_70,code=sm_70 --generate-code arch=compute_75,code=sm_75 --generate-code arch=compute_72,code=sm_72 --generate-code arch=compute_80,code=sm_80 --generate-code arch=compute_86,code=sm_86 --generate-code arch=compute_89,code=sm_89 plus whatever other flags you have there | |
ID: 9771 · Rating: 0 · rate: / Reply Quote | |
you need to add support for every cc level explicitly. ok, noticed, thx, we need to change that again | |
ID: 9772 · Rating: 0 · rate: / Reply Quote | |
v32 recompiled with the fix, I hope... | |
ID: 9773 · Rating: 0 · rate: / Reply Quote | |
I wanted to try v32, but even after resetting the project in the client, the server only sent me v28 cuda100 tasks. Oddly enough, the cuda100 tasks used to work on the P100 but now they don't. So I put this as an app_config.xml: <app_config>
<app_version>
<app_name>TF</app_name>
<plan_class>cuda120</plan_class>
<min_gpu_ram_mb>384</min_gpu_ram_mb>
<gpu_ram_used_mb>384</gpu_ram_used_mb>
<gpu_peak_flops_scale>0.22</gpu_peak_flops_scale>
<cpu_frac>0.01</cpu_frac>
</app_version>
</app_config> Now I can run the new v32 and it works fine, but the app_config has to be used at least on this particular computer. I have not tried the rest yet. EDIT: Actually, I spoke too soon. This app_config does NOT prevent the download of v28 cuda100 tasks. So I'm still not able to run all my GPUs. | |
ID: 9775 · Rating: 0 · rate: / Reply Quote | |
I saw the new version download. Works on both my PCs. | |
ID: 9776 · Rating: 0 · rate: / Reply Quote | |
I saw the new version download. Works on both my PCs. And you have P100s? | |
ID: 9777 · Rating: 0 · rate: / Reply Quote | |
You most likely know I don't. | |
ID: 9778 · Rating: 0 · rate: / Reply Quote | |
I actually had no idea, your computers are hidden. | |
ID: 9779 · Rating: 0 · rate: / Reply Quote | |
I wanted to try v32, but even after resetting the project in the client, the server only sent me v28 cuda100 tasks. Oddly enough, the cuda100 tasks used to work on the P100 but now they don't. So I put this as an app_config.xml: It should work without an app_config file or not? | |
ID: 9780 · Rating: 0 · rate: / Reply Quote | |
your cuda100 app P100s are now failing with this: ./mfaktc.exe: error while loading shared libraries: libcudart.so.10.1: cannot open shared object file: No such file or directory maybe you removed the cuda toolkit? | |
ID: 9781 · Rating: 0 · rate: / Reply Quote | |
your cuda100 app P100s are now failing with this: https://srbase.my-firewall.org/sr5/download/libcudart.so.10.1 I will update the app including this file. Update: The missing libfile is now in the zipfile of v29 cuda100 app | |
ID: 9782 · Rating: 0 · rate: / Reply Quote | |
Thank you for all the work you agreed to do just to get one old model of GPU running, I appreciate it greatly. | |
ID: 9784 · Rating: 0 · rate: / Reply Quote | |
All seems to be working well on P100s. I believe that in honor of this fix, I will put all my GPUs on TF for a while. | |
ID: 9856 · Rating: 0 · rate: / Reply Quote | |
Message boards :
Number crunching :
Nvidia Tesla P100 Problems