Folding Forum

Sometimes a driver crash successfully resets the display portion but needs a full system reboot to restore compute capabilities.

This is an old Linux box that I had resurrected to fold, with the idea of it sitting quietly in a corner and not being used for anything else other than looking for a possible cure. Based on the above, even if my client returns complete WUs, it is likely that someone else has already finished the s...

I don't believe this is a GROMACS issue. The parameters FAHclient passes to mdrun results in PME being used. It allows for better utilization of high thread counts. PME could be disabled by passing -npme = 0 but would cause this problem to occur more often . The current procedures that FAH uses hav...

… and that (bruce's post) just goes to show I'm not a GPU folder :) It does surprise me though that the Tesla K40ms are considered rather weak … I know their clocks are down in comparison to the 1050s but with the significantly larger shader count (x4 ish) and higher FLOPs performance (x2 ish) I'd ...

Welcome to the forum glennsuys! Since you were receiving work units before, that's a good sign your drivers and client are setup correctly. The time between checks for new work increases each time to reduce load on the servers. If a slot gets to an hour between attempts, pause the slot for 30 second...

Humm... In that case, maybe _r2w_ben can explain what this means has they have been awesome at documenting the results (viewtopic.php?f=72&t=34350): 4 particles communicated to PME rank 10 are more than 2/3 times the cut-off out of the domain decomposition cell of their charge group in dimensio...

FAH uses GROMACS 5.0.4, which was released in 2014 and predates Zen. The message about AVX_128_FMA is relevant to Bulldozer/Piledriver that were available at the time. For those architectures, AVX_128_FMA > AVX_256 > SSE2. With Zen 2 the message doesn't appear to be correct. The code probably checks...

Goodmorning folders, Is a few days that I noticed that my CPU (AMD A12-9720 RADEON R7 COMPUTE CORE 4C+8G) is stucked on low usage (seeing from task manager of Windows 10), infact when I fold I can see that is only using almost 12-20% of my computing power, and is working only about 0.7-1.0 GHz freq...

GPUs.txt from Apr 7: 0x1002:0x6898:1:5:Cypress [Radeon HD 5800/6800] 0x1002:0x6899:1:5:Cypress Pro [Radeon HD 5800/6850] 0x1002:0x689b:1:5:EG Cypress [Radeon HD 6800 Series] A version I downloaded earlier today: 0x1002:0x6898:1:2:Cypress [Radeon HD 5800/6800] 0x1002:0x6899:1:5:Cypress PRO [Radeon HD...

HD 5850 is listed as species 5 in GPUs.txt and should be capable of folding: 0x1002:0x6899:1:5:Cypress PRO [Radeon HD 5850] An older version I had locally listed it differently: 0x1002:0x6899:1:5:Cypress Pro [Radeon HD 5800/6850] HD 6850 is a Barts Pro and shouldn't have been combined in the label. ...

This looks like a bad work unit. Pause the slot, go to /var/lib/fahclient/work, delete /00/ and resume the slot.

Instead of running two slots with 4 CPUs, you might want to run a single slot with 8 CPUs. Faster completions are preferred and rewarded by the Quick Return Bonus.

Based on it being a HD 6850 or similar, your GPU likely does not support double precision. Can you look back in your logs and confirm whether the last work unit it received was FahCore_21? There are still a few of those projects around (listed as OPENMM_21 ) but most new GPU projects are using FahCo...

Upload failure here as well. 01:39:12:WU02:FS01:Sending unit results: id:02 state:SEND error:NO_ERROR project:11743 run:0 clone:7560 gen:80 core:0x22 unit:0x0000006f8ca304f15e6bc48b14cb7895 01:39:12:WU02:FS01:Uploading 12.56MiB to 140.163.4.241 01:39:12:WU02:FS01:Connecting to 140.163.4.241:8080 01:...

Data for a 7x7x6. It's similar to larger projects with the exception of excluding 11, 99, 102, and 108 threads. p13840 - max 7x7x6 - PME load 0.19 2 = 2x1x1 3 = 3x1x1 4 = 4x1x1 5 = 5x1x1 6 = 6x1x1 7 = 7x1x1 8 = 4x2x1 9 = 3x3x1 10 = 5x2x1 12 = 6x2x1 15 = 5x3x1 16 = 4x4x1 18 = 6x3x1 20 = 4x4x1 16 + 4 ...

Scaling is a challenge for parallel processing. As the number of threads increases, so does the time spent communicating and waiting to synchronize. If the operating system is scheduling well, 6 threads would each end up on a physical core. Beyond that, threads will be sharing resources on physical ...