Thread 'NVIDIA GPU beta — volunteers wanted'
Message boards : Solver development and optimisation : NVIDIA GPU beta — volunteers wanted
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5
| Author | Message |
|---|---|
|
New member Send message Joined: 14 Oct 26 Posts: 1 Credit: 1,105,794 RAC: 71,672  
|
Hello Benjamin, are other GPU then the named allowed? When, then: -NVIDIA GeForce GTX 1050 Ti; 4.00 GB RAM; Driver: 58278; OpenCL: 3.0 on Win10pro, ID 701; - NVIDIA GeForce RTX 4060 Ti; 7.74 GB RAM; Driver: 55099; OpenCL: 3.0 on Linux, ID 703. Will they go? Greets, vonboedefeldt. |
|
Send message Joined: 23 Sep 26 Posts: 126 Credit: 9,500,963 RAC: 562,489  |
@Freewill: thanks for correcting me, and sorry, I assumed MPS wrongly. So the spread comes only from what's inside each batch: some contain many more deep endgames than others, and those take much longer to compute, while the credit is the same because the ticket count is the same. Over many tasks it evens out. @vonboedefeldt: yes, both work, and both are now enabled! On our side the IDs are the other way round: host 701 is the Linux RTX 4060 Ti and host 703 the Windows GTX 1050 Ti (older cards like the 1050 Ti use our separate version for older GPUs). Just keep new tasks allowed and click Update in BOINC; no CUDA installation or reset is needed. |
|
Send message Joined: 27 Sep 26 Posts: 17 Credit: 7,714,804 RAC: 426,742
|
Yet another task is gone downhill... bw_runner: progress 59244/65552 (90%), restored runtime 4603 s bw_runner: progress 59999/65552 (91%), restored runtime 4651 s bw_runner: progress 60820/65552 (92%), restored runtime 4697 s bw_runner: progress 61628/65552 (94%), restored runtime 4743 s bw_runner: progress 62467/65552 (95%), restored runtime 4788 s bw_runner: progress 63359/65552 (96%), restored runtime 4835 s bw_runner: progress 64251/65552 (98%), restored runtime 4881 s bw_runner: progress 64712/65552 (98%), restored runtime 4925 s bw_runner: progress 64823/65552 (98%), restored runtime 4973 s bw_runner: progress 64848/65552 (98%), restored runtime 5020 s bw_runner: progress 64862/65552 (98%), restored runtime 5066 s bw_runner: progress 64870/65552 (98%), restored runtime 5112 s bw_runner: progress 64872/65552 (98%), restored runtime 5157 s bw_runner: progress 64873/65552 (98%), restored runtime 5203 s bw_runner: progress 64877/65552 (98%), restored runtime 5300 s bw_runner: progress 64879/65552 (98%), restored runtime 5348 s bw_runner: progress 64880/65552 (98%), restored runtime 5396 s bw_runner: progress 64881/65552 (98%), restored runtime 5445 s bw_runner: progress 64882/65552 (98%), restored runtime 5492 s bw_runner: progress 64883/65552 (98%), restored runtime 5538 s bw_runner: progress 64885/65552 (98%), restored runtime 5584 s bw_runner: progress 64886/65552 (98%), restored runtime 5725 s bw_runner: progress 64887/65552 (98%), restored runtime 5960 s bw_runner: progress 64888/65552 (98%), restored runtime 6006 s bw_runner: progress 64889/65552 (98%), restored runtime 6386 s bw_runner: progress 64890/65552 (98%), restored runtime 6814 s bw_runner: progress 64891/65552 (98%), restored runtime 7047 s bw_runner: progress 64892/65552 (98%), restored runtime 7652 s bw_runner: progress 64893/65552 (98%), restored runtime 9961 sThe time for each new ticket is gotten much higher, from 50 seconds (for 750 tickets) to over 2000 seconds per ticket. |
|
Send message Joined: 23 Sep 26 Posts: 126 Credit: 9,500,963 RAC: 562,489 |
In reply to Werinbert's message of 22 Oct 2026: Yet another task is gone downhill... Thanks Werinbert. This is an exceptional case: your task has 659 tickets left, just above our 1% cut (655 tickets), so it's grinding through a few very hard tickets before it can hand the rest over to CPU tasks. It only needs 4 more tickets to reach that point, so it should finish shortly. It happens rarely, but it can happen. Please let it run, no need to abort it. |
