Folding Forum

muziqaz wrote: ↑Tue Sep 10, 2024 5:36 am

braiam wrote: ↑Tue Sep 10, 2024 4:41 am This still happens with core 23. Core 24 however seems to not have issues.

Screenshots or it didn't happen

Is there any way to force a core 23 work? Or for me to do a dummy job? I haven't found anything on the forums that would achieve this.

This still happens with core 23. Core 24 however seems to not have issues.

lafrad wrote:Yea, and the Control launches the workers too, so it'll have to interpret that too.... hmm

Are you sure?

numactl runs processes with a specific NUMA scheduling or memory placement policy. The policy is set for command and inherited by all of its children.

PantherX wrote:

schapman1978 wrote:...Maybe there's a shortage of more complex units? 2 of my GPUS's are idle this morning.

That's correct. Work is being done to meet the high demand for GPU WUs.

Any suggestion for my issue?

The rocm drivers are hit or miss from reports here on the forum. Some have got them to work, but it may depend on exact versions used. The AMD proprietary drivers and OpenCL support are still the recommended packages to use. In my case, I'm only using the OpenCL libraries. I've successfully folded ...

These projects seems to not being able to run at all on my GPU using the rocm libraries, meanwhile projects 16435, 11742-11745, 14549 have been able to. 16435 running right now. I'm not sure what's wrong with these projects, except: 01:27:17:WU01:FS00:0x21:*********************** Log Started 2020-05...

I got this project and is failing on my GPU, with rocm opencl drivers.

I did the same. But instead I edited the init file so that it uses root user instead. viewtopic.php?f=106&t=34231

I solved this problem installing ROCm libaries as described here viewtopic.php?f=106&t=34231&p=324652#p324652

There's a qualitative advantage of using the ROCm drivers rather than the AMDGPU(-PRO) packages. It has better support for newer cards, extremely modular (no more having to have vdpau and va drivers from the AMD drivers mixed with mesa) and it's developed with focus on computing. Another advantage i...

In the configuration, you can only set up one or the other will it fold. I recommend using mainly GPU since it is one of the most lacking resources for the project, as every system has a GPU. But no, there's no way to set only one at a time (unless you micromanage).

I've seen several issues that MIGHT be associated with AMD GPU drivers but no proof. Just on a lark, go to /var/lib/fahclient/work and confirm that the account that's running FAHCore_22 has r/w permissions. What files being created there? This log is showing WU01 is active so the files should be in...

I've seen several issues that MIGHT be associated with AMD GPU drivers but no proof. Just on a lark, go to /var/lib/fahclient/work and confirm that the account that's running FAHCore_22 has r/w permissions. What files being created there? This log is showing WU01 is active so the files should be in...

Well, found some but ended with BAD_WORK_UNIT: Bad platformId size 21:25:32:WU01:FS00:Download complete 21:25:32:WU01:FS00:Received Unit: id:01 state:DOWNLOAD error:NO_ERROR project:11776 run:0 clone:33079 gen:0 core:0x22 unit:0x00000000287234c95e7931379a1cd6d4 21:25:32:WU01:FS00:Starting 21:25:32:W...

Hm, I have seen 11777 folding successfully on other AMD GPU, though not in Linux. Have you tried restarting the slot? Also during the wait period, have you noticed if fahcore_22 is using 1.5/2 threads of CPU? I've tried removing the work unit from the /work directory, still the same. I scrapped eve...