![]() ![]() PS I love amdgpu in general, in terms of stability and ease of use it's a big step up from the old fglrx days, so thanks to everyone involved with its development and maintenance. Obviously this is not a tenable long-term position - is there anyone who has had similar experience and/or can assist with BOINC GPU processing under ROCr?Īppreciate any help or assistance. One might argue that individual BOINC projects will have to adapt to using ROCr, but I think it would be better to work out why ROCr breaks the processing in the first place and tackle the problem at the root.įor the time being I am stuck with staying on amdgpu-pro 20.40 and because of aforementioned point 1, this also means I am stuck on an out-of-date Linux kernel. However, not at the expense of breaking existing functionality. I understand the aim of unifying software stacks and in principle agree it's a good idea. ![]() I experienced this with both the November 2020 release of amdgpu-pro 20.45 (build 1164792) and December 2020 (build 1188099). ![]() Using OpenCL device "gfx900" by: Advanced Micro Devices, Inc.Ĭouldn't create OpenCL command queue (error: -6)!īoth of these error logs appear to be problems with using OpenCL devices under ROCr, and even though BOINC does detect OpenCL devices correctly apparently the project applications cannot work properly. Using OpenCL platform provided by: Advanced Micro Devices, Inc. I have confirmed this with two separate projects - keeping in mind that GPU processing works just fine with PAL OpenCL.Įxample 1: Error: Failed to compile opencl source (from CL or HIP source to LLVM IR).Įxample 2 (gfx900 = Vega 64): boinc_get_opencl_ids returned I have been running these GPUs with PAL OpenCL just fine for several months now, but with the switch to ROCr, all BOINC processing on GPUs fail. However, my problems lie with amdgpu-pro on Vega 10 and Vega 20 GPUs (Vega 64 and Radeon VII respectively). I have confirmed this by similarly running amdgpu-pro with 'legacy' OpenCL on Fiji GPUs successfully. eg My BOINC comrade is only running Polaris GPUs under amdgpu-pro and thus can get by with the 'legacy' OpenCL installation. amdgpu-pro 20.45 displaces PAL OpenCL in favour of a ROCr implementation.įor point 1, this was resolved with amdgpu-pro 20.45.įor point 2, this appears only to be an issue for GPUs based on Vega and later.amdgpu-pro 20.40 is incompatible with Ubuntu Linux kernels after 5.4.0-54.I made a search for existing posts that may be related to my specific issue, but I didn't find any.īetween the two of us we've established the following: It is not recommended to use HT for PRP/LLR, only for factoring on CPU.After corresponding with a fellow BOINC user I decided to make a post here to highlight some difficulties I've been having - in particular, I was advised may be able to assist. Prime95 will always force you to not run all threads on your cores for better performance even when you set it on local.txt file. Message 6916 - Posted:, 16:58:16 UTC - in response to Message 6915. I'm still looking for a means of getting SRBase to use my GPU without (significantly) interfering with Prime95. HWInfo shows my GPU utilization at 100% (though Task Manager seems to think the GPU is idle). I tried giving BOINC just 1 thread and 1% of available CPU time and I tried giving it all 8 threads and 100% of CPU time processor utilization varies, but rarely approaches the over 90% it does running Prime95 alone. I get how due to the vagaries of Windows, giving Prime95 a priority of 10 doesn't result in it getting any more cycles, but none of the available settings in Prime95 or BOINC seem to alter the fact that Prime95 won't utilize more than half of available processor time if BOINC is running. I spoke too soon - in Windows Task Manager, it appears that Prime95 never exceeds ~48% CPU usage if BOINC is running. Message 6915 - Posted:, 16:52:16 UTC - in response to Message 6909. ![]() I haven't rebooted - I'll try that and edit this or report back if computation switched to my GPU.Ĭheck also in BOINC client the prefs, Use GPU while computer is in use, also check the event log if your GPU is recognized. Having done so (and completed several work units since), should SRBase now detect my GPU and switch computation to it automatically, or do I need to do something else? My CPU is at full capacity and my GPU is idle. "TF"? There is a telegraphic reference to a path with regards to OpenCL - all I've done is install and run this project under BOINC, then install AMD's SDK 3.0 (from Softpedia, AMD reportedly having withdrawn it) while it was running. Those radio buttons (ATI and Nvidia) were already checked in the project's preferences when I looked. Message 6905 - Posted:, 7:05:32 UTC - in response to Message 6904. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |