Message boards :
Number crunching :
OpenCL errors with RTX 5xxxx series GPUs !?
Message board moderation
Author | Message |
---|---|
![]() Send message Joined: 15 Jan 18 Posts: 4 Credit: 134,459,756 RAC: 520,004 ![]() |
Hi, I recently got an RTX 5060 Ti, which continuously gives an error, with the 1.05 app (opencl_nvidia_101), exactly after 12 seconds of WUs processing. I tried to reset, delete and reinstall the project, but it does not download 1.03 (cuda31) WUs. All this with Windows 10 (64) professional, with latest update. P.S. I looked in the statistics, but no RTX 5xxx appears |
![]() Send message Joined: 15 Jan 18 Posts: 4 Credit: 134,459,756 RAC: 520,004 ![]() |
I managed to download WUs cuda, but it still gives error after 12~17 seconds. Same situation with older video drivers. |
![]() Send message Joined: 15 Jan 18 Posts: 4 Credit: 134,459,756 RAC: 520,004 ![]() |
The reason revealed... https://nvidia.custhelp.com/app/answers/detail/a_id/5615/~/support-plan-for-32-bit-cuda "CUDA Driver will not support 32-bit CUDA applications on GeForce RTX 50 series (Blackwell) and newer architectures. Support for running x86 32-bit applications on x86_64 Windows is limited to use with: CUDA Driver. CUDA Runtime (cudart)" |
Send message Joined: 24 Jun 23 Posts: 1 Credit: 436,322,466 RAC: 486 |
Hello, here is the same problem: [Jun 22 22:33:53 UTC] Automatic processor type detection did not recognize the processor (tag: "NVIDIA GeForce RTX 5070 ") I hope the programmers will find a solution to run MOO Wrapper on RTX 5xxx and newer cards. Thanks Jochen |
![]() Send message Joined: 15 Jan 18 Posts: 4 Credit: 134,459,756 RAC: 520,004 ![]() |
Hi, hmmm... I doubt that the project administrator can do anything, since the distributed.net client does not support 64-bit GPU applications on Windows OS: https://www.distributed.net/Download_clients And for a long time, no one has done anything about it. There are other projects that uses this platform, but they are 64 bit only for CPU (e.g. yoyo). A solution could be a dual boot (or a independent PC) with a Linux distribution. Even virtualizing does not solve the problem because you cannot assign large resources to such a machine. |