Message boards :
News :
New GPU OpenCL versions available
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · 6 · Next
Author | Message |
---|---|
Send message Joined: 30 Apr 18 Posts: 1 Credit: 2,017,908 RAC: 0 |
I try it with Adrenalin 20.2.2, but NumberFields still not working on my three RX 5700. :( |
Send message Joined: 24 May 19 Posts: 38 Credit: 1,358,375 RAC: 771 |
I also tried version 20.2.2 that does not work, but the processing progress percentage stops at 30.042% and not 10.020% as in previous drivers. OpenCL: AMD/ATI GPU 0: AMD Radeon RX 5700 XT 50th Anniversary (driver version 3004.8 (PAL,LC), device version OpenCL 2.0 AMD-APP (3004.8), 8176MB, 8176MB available, 4685 GFLOPS peak) Eric, did you report the problem to AMD? |
Send message Joined: 8 Jul 11 Posts: 1341 Credit: 494,164,631 RAC: 559,354 |
I also tried version 20.2.2 that does not work, but the processing progress percentage stops at 30.042% and not 10.020% as in previous drivers. No, I did not. Without an explicit example proving it it their driver, I doubt they would listen. If there was a way to see everything they changed between the last working driver version (19.10.1) and the next version, then maybe we could isolate the cause. |
Send message Joined: 14 Mar 19 Posts: 9 Credit: 102,990,631 RAC: 128,327 |
I also tried version 20.2.2 that does not work, but the processing progress percentage stops at 30.042% and not 10.020% as in previous drivers. Got same probelm with RX 5700 |
Send message Joined: 18 Mar 13 Posts: 2 Credit: 1,389,952 RAC: 0 |
I had one AMD WU go over 16 hours and showed 100% done but kept running... another doing the the same thing so aborted also. https://numberfields.asu.edu/NumberFields/workunit.php?wuid=72560796 Nvidia WU's seem to run fine using 442.50 with clean install. 8-) |
Send message Joined: 8 Jul 11 Posts: 1341 Credit: 494,164,631 RAC: 559,354 |
I had one AMD WU go over 16 hours and showed 100% done but kept running... another doing the the same thing so aborted also. It looks like that computer no longer has the AMD GPU attached (according to BOINC). Which GPU model was it and what driver version? |
Send message Joined: 24 May 19 Posts: 38 Credit: 1,358,375 RAC: 771 |
Hi guys I wanted to notify you that I sent AMD the report of the application problem with the drivers. I invite all those who have had the same problems to do the same. The more reports they receive and maybe first solve the problem. |
Send message Joined: 24 May 19 Posts: 38 Credit: 1,358,375 RAC: 771 |
Update. AMD released a new beta version of the drivers, version 20.3.1 and I tried it, but unfortunately it does not work, the processing progress percentage stops at 10.020% as in previous drivers. OpenCL: AMD/ATI GPU 0: AMD Radeon RX 5700 XT 50th Anniversary (driver version 3004.8 (PAL,LC), openCL 2.0 AMD-APP device version (3004.8), 8176MB, 8176MB available, 4685 GFLOPS peak). Reported issue to AMD and I invite anyone with the same problem, to do the same. Let's see if they solve the problem to the fury of reports. |
Send message Joined: 8 Jul 11 Posts: 1341 Credit: 494,164,631 RAC: 559,354 |
Update. Thanks for your dedication to this! |
Send message Joined: 24 May 19 Posts: 38 Credit: 1,358,375 RAC: 771 |
Update. AMD released a new beta version of the drivers, version 20.4.1 and I tried it, but unfortunately it does not work, the processing progress percentage stops at 9.386% as in previous drivers. OpenCL: AMD/ATI GPU 0: AMD Radeon RX 5700 XT 50th Anniversary (driver version 3004.8 (PAL,LC), openCL 2.0 AMD-APP device version (3004.8), 8176MB, 8176MB available, 4685 GFLOPS peak). Reported issue to AMD and I invite anyone with the same problem, to do the same. Let's see if they solve the problem to the fury of reports. |
Send message Joined: 24 May 19 Posts: 38 Credit: 1,358,375 RAC: 771 |
Update. AMD released a new beta version of the drivers, version 20.4.2 and I tried it, but unfortunately it does not work, the processing progress percentage stops at 28.129% as in previous drivers. OpenCL: AMD/ATI GPU 0: AMD Radeon RX 5700 XT 50th Anniversary (driver version 3004.8 (PAL,LC), openCL 2.0 AMD-APP device version (3004.8), 8176MB, 8176MB available, 4685 GFLOPS peak). Reported issue to AMD and I invite anyone with the same problem, to do the same. Let's see if they solve the problem to the fury of reports. |
Send message Joined: 14 Mar 19 Posts: 9 Credit: 102,990,631 RAC: 128,327 |
Thank you for the update. |
Send message Joined: 22 Jul 19 Posts: 8 Credit: 1,467,691 RAC: 1,091 |
All the WUs of the “Get Decic Fields v3.05 (opencl_amd) windows_x86_64” application show the “Ramaining (estimated)” time is forever rising. I prefer to abort all the WUs and to wait for some news in this forum thread. GPU: Sapphire Radeon R7 240 (2GB) Radeon Software Adrenalin 2020 Edition 19.12.3 Driver Version 19.50.03.05 (Windows Driver Store Version 26.20.15003.5016) OpenCL: AMD/ATI GPU 0: AMD Radeon R7 200 Series (driver version 3004.7, device version OpenCL 1.2 AMD-APP (3004.7), 2048MB, 2048MB available, 499 GFLOPS peak). I know my GPU is quite old. If you think my GPU is the root cause of my issue, please tell me. |
Send message Joined: 24 May 19 Posts: 38 Credit: 1,358,375 RAC: 771 |
All the WUs of the “Get Decic Fields v3.05 (opencl_amd) windows_x86_64” application show the “Ramaining (estimated)” time is forever rising. Have you read the whole discussion? Have you tried installing version 19.10.1 and 19.9.2 OF AMD RADEON drivers to see if it works as it did to me? |
Send message Joined: 8 Jul 11 Posts: 1341 Credit: 494,164,631 RAC: 559,354 |
All the WUs of the “Get Decic Fields v3.05 (opencl_amd) windows_x86_64” application show the “Ramaining (estimated)” time is forever rising. Yes, I believe that card is too old for this app. |
Send message Joined: 22 Jul 19 Posts: 8 Credit: 1,467,691 RAC: 1,091 |
Have you tried installing version 19.10.1 and 19.9.2 OF AMD RADEON drivers to see if it works as it did to me? Thank you. I'll try. |
Send message Joined: 22 Jul 19 Posts: 8 Credit: 1,467,691 RAC: 1,091 |
Hi Eric Driver and mg13[HWU], no good news. The downgrade of the graphical adapter device driver was not useful. I installed Radeon Software Adrenalin 2019 Edition 19.10.1. OpenCL: AMD/ATI GPU 0: AMD Radeon R7 200 Series (driver version 2906.10, device version OpenCL 1.2 AMD-APP (2906.10), 2048MB, 2048MB available, 499 GFLOPS peak) After 19 hours of “Elapsed Time”, the “Progress” is frozen at 15% and the “Ramaining (estimated)” time is 4 days. The “Ramaing (estimated)” time is still rising. I agree to Eric Driver, my GPU is to old for “Get Decic Fields 3.05 (opencl_amd)” application. I had the same issue (I mean the “Ramaing (estimated)” time is rising limitless) with Einstain@home application. The response was: R7 240 has GCN 1.0 architecture. |
Send message Joined: 8 Jul 11 Posts: 1341 Credit: 494,164,631 RAC: 559,354 |
Hi Eric Driver and mg13[HWU], no good news. This highlights one of the advantages of Cuda over openCL. Cuda gets built one time before deployment instead of at run time. This allows the developer to see which architectures work - Cuda simply fails to build on the older architectures (i.e. older compute capability). So older cards for both AMD an Nvidia fail, but at least Nvidia Cuda tells me which ones are the problem... if only there was a way to do that with AMD cards. |
Send message Joined: 24 May 19 Posts: 38 Credit: 1,358,375 RAC: 771 |
Hi Eric Driver and mg13[HWU], no good news. I'm sorry it's not working. If you have time and want to try version 19.9.2, just for the correctness of the tests and my curiosity, thank you. |
Send message Joined: 22 Jul 19 Posts: 8 Credit: 1,467,691 RAC: 1,091 |
Hi Eric Driver and mg13[HWU], If you have time and want to try version 19.9.2, just for the correctness of the tests and my curiosity, thank you. Of course, I’m going to find some spare time to do it but I’d like Eric Driver reads what I’ve written below. The test makes sense if the development team tell me how to enable the debug mode and to collect the trace/log file. I explain me better. My environment could be affected by the obsolescence of the GPU HW, not by a device driver bug. Only the trace file could show us where the root cause was. |