CPU Only

Message boards : Frequently Asked Questions (FAQ) : CPU Only

Author	Message
SACPOP Send message Joined: 18 Aug 17 Posts: 1 Credit: 72,163 RAC: 0 Level Scientific publications	Message 60931 - Posted: 4 Jan 2024 \| 18:22:47 UTC
	Is it possible to crunch for GPUGRID with only the CPU? I understand it is primarily GPU-oriented, and I'm ok with irregular tasks, but is it possible to participate with only a CPU?
	ID: 60931 \| Rating: 0 \| rate: / Reply Quote

Ian&Steve C. Send message Joined: 21 Feb 20 Posts: 1079 Credit: 40,231,533,983 RAC: 16 Level Scientific publications	Message 60932 - Posted: 4 Jan 2024 \| 18:40:58 UTC - in response to Message 60931.
	no ____________
	ID: 60932 \| Rating: 0 \| rate: / Reply Quote

Endasil Send message Joined: 22 May 15 Posts: 4 Credit: 6,210,000 RAC: 0 Level Scientific publications	Message 61520 - Posted: 21 May 2024 \| 20:23:32 UTC
	I'm unable to get any work, perhaps someone can look at this log and tell me what is wrong? 2024-05-21 22:19:08 \| GPUGRID \| [prio] recent est credit: 0.00G in 60.68 sec, 37024.311710 + -1.802412 ->37022.509298 2024-05-21 22:19:08 \| GPUGRID \| [prio] -0.307684 rsf 0.500000 rt 37024.311710 rs 240664.535469 2024-05-21 22:19:08 \| GPUGRID \| [prio] -0.295703 rsf 0.500000 rt 35582.612329 rs 240664.535469 2024-05-21 22:19:23 \| \| choose_project(): 1716322763.563406 2024-05-21 22:19:23 \| GPUGRID \| [prio] -0.295703 rsf 0.500000 rt 35582.612329 rs 240664.535469 2024-05-21 22:19:23 \| GPUGRID \| [prio] -0.530563 rsf 0.250000 rt 37022.509298 rs 279118.895497 2024-05-21 22:19:23 \| \| [work_fetch] ------- start work fetch state ------- 2024-05-21 22:19:23 \| \| [work_fetch] target work buffer: 17280.00 + 25920.00 sec 2024-05-21 22:19:23 \| \| [work_fetch] --- project states --- 2024-05-21 22:19:23 \| GPUGRID \| [work_fetch] REC 37022.509 prio -0.531 can request work 2024-05-21 22:19:23 \| \| [work_fetch] --- state for CPU --- 2024-05-21 22:19:23 \| \| [work_fetch] shortfall 183580.35 nidle 0.00 saturated 24600.20 busy 0.00 2024-05-21 22:19:23 \| GPUGRID \| [work_fetch] share 0.000 no applications 2024-05-21 22:19:23 \| \| [work_fetch] --- state for NVIDIA GPU --- 2024-05-21 22:19:23 \| \| [work_fetch] shortfall 0.00 nidle 0.00 saturated 45044.92 busy 0.00 2024-05-21 22:19:23 \| GPUGRID \| [work_fetch] share 0.000 project is backed off (resource backoff: 28608.37, inc 38400.00) 2024-05-21 22:19:23 \| \| [work_fetch] --- state for Intel GPU --- 2024-05-21 22:19:23 \| \| [work_fetch] shortfall 43200.00 nidle 1.00 saturated 0.00 busy 0.00 2024-05-21 22:19:23 \| GPUGRID \| [work_fetch] share 0.000 project is backed off (resource backoff: 72401.44, inc 86400.00) 2024-05-21 22:19:23 \| \| [work_fetch] --- state for NVIDIA GeForce RTX 4080 Laptop GPU --- 2024-05-21 22:19:23 \| \| [work_fetch] shortfall 43200.00 nidle 1.00 saturated 0.00 busy 0.00 2024-05-21 22:19:23 \| GPUGRID \| [work_fetch] share 0.000 project is backed off (resource backoff: 75876.75, inc 86400.00) 2024-05-21 22:19:23 \| \| [work_fetch] --- state for Intel(R) UHD Graphics --- 2024-05-21 22:19:23 \| \| [work_fetch] shortfall 43200.00 nidle 1.00 saturated 0.00 busy 0.00 2024-05-21 22:19:23 \| GPUGRID \| [work_fetch] share 0.000 project is backed off (resource backoff: 5073.41, inc 86400.00) 2024-05-21 22:19:23 \| \| [work_fetch] ------- end work fetch state ------- 2024-05-21 22:19:23 \| GPUGRID \| choose_project: scanning 2024-05-21 22:19:23 \| GPUGRID \| can't fetch CPU: no applications 2024-05-21 22:19:23 \| GPUGRID \| can't fetch NVIDIA GPU: project is backed off 2024-05-21 22:19:23 \| GPUGRID \| can't fetch Intel GPU: project is backed off 2024-05-21 22:19:23 \| GPUGRID \| can't fetch NVIDIA GeForce RTX 4080 Laptop GPU: project is backed off 2024-05-21 22:19:23 \| GPUGRID \| can't fetch Intel(R) UHD Graphics: project is backed off 2024-05-21 22:19:23 \| \| [work_fetch] No project chosen for work fetch 2024-05-21 22:20:09 \| GPUGRID \| [prio] recent est credit: 0.00G in 60.61 sec, 37022.509298 + -1.800050 ->37020.709248 2024-05-21 22:20:09 \| GPUGRID \| [prio] -0.307616 rsf 0.500000 rt 37022.509298 rs 240706.388640 2024-05-21 22:20:09 \| GPUGRID \| [prio] -0.295702 rsf 0.500000 rt 35588.641740 rs 240706.388640 2024-05-21 22:20:19 \| \| [work_fetch] Request work fetch: Backoff ended for World Community Grid 2024-05-21 22:20:19 \| \| choose_project(): 1716322819.176867 2024-05-21 22:20:19 \| GPUGRID \| [prio] -0.295702 rsf 0.500000 rt 35588.641740 rs 240706.388640 2024-05-21 22:20:19 \| GPUGRID \| [prio] -0.658988 rsf 0.200000 rt 37020.709248 rs 280890.522194 2024-05-21 22:20:19 \| \| [work_fetch] ------- start work fetch state ------- 2024-05-21 22:20:19 \| \| [work_fetch] target work buffer: 17280.00 + 25920.00 sec 2024-05-21 22:20:19 \| \| [work_fetch] --- project states --- 2024-05-21 22:20:19 \| GPUGRID \| [work_fetch] REC 37020.709 prio -0.659 can request work 2024-05-21 22:20:19 \| \| [work_fetch] --- state for CPU --- 2024-05-21 22:20:19 \| \| [work_fetch] shortfall 186815.93 nidle 0.00 saturated 24370.99 busy 0.00 2024-05-21 22:20:19 \| GPUGRID \| [work_fetch] share 0.000 no applications 2024-05-21 22:20:19 \| \| [work_fetch] --- state for NVIDIA GPU --- 2024-05-21 22:20:19 \| \| [work_fetch] shortfall 0.00 nidle 0.00 saturated 45037.92 busy 0.00 2024-05-21 22:20:19 \| GPUGRID \| [work_fetch] share 0.000 project is backed off (resource backoff: 28552.76, inc 38400.00) 2024-05-21 22:20:19 \| \| [work_fetch] --- state for Intel GPU --- 2024-05-21 22:20:19 \| \| [work_fetch] shortfall 43200.00 nidle 1.00 saturated 0.00 busy 0.00 2024-05-21 22:20:19 \| GPUGRID \| [work_fetch] share 0.000 project is backed off (resource backoff: 72345.83, inc 86400.00) 2024-05-21 22:20:19 \| \| [work_fetch] --- state for NVIDIA GeForce RTX 4080 Laptop GPU --- 2024-05-21 22:20:19 \| \| [work_fetch] shortfall 43200.00 nidle 1.00 saturated 0.00 busy 0.00 2024-05-21 22:20:19 \| GPUGRID \| [work_fetch] share 0.000 project is backed off (resource backoff: 75821.13, inc 86400.00) 2024-05-21 22:20:19 \| \| [work_fetch] --- state for Intel(R) UHD Graphics --- 2024-05-21 22:20:19 \| \| [work_fetch] shortfall 43200.00 nidle 1.00 saturated 0.00 busy 0.00 2024-05-21 22:20:19 \| GPUGRID \| [work_fetch] share 0.000 project is backed off (resource backoff: 5017.80, inc 86400.00) 2024-05-21 22:20:19 \| \| [work_fetch] ------- end work fetch state ------- 2024-05-21 22:20:19 \| GPUGRID \| choose_project: scanning 2024-05-21 22:20:19 \| GPUGRID \| can't fetch CPU: no applications 2024-05-21 22:20:19 \| GPUGRID \| can't fetch NVIDIA GPU: project is backed off 2024-05-21 22:20:19 \| GPUGRID \| can't fetch Intel GPU: project is backed off 2024-05-21 22:20:19 \| GPUGRID \| can't fetch NVIDIA GeForce RTX 4080 Laptop GPU: project is backed off 2024-05-21 22:20:19 \| GPUGRID \| can't fetch Intel(R) UHD Graphics: project is backed off 2024-05-21 22:20:19 \| \| [work_fetch] No project chosen for work fetch 2024-05-21 22:21:09 \| GPUGRID \| [prio] recent est credit: 0.00G in 60.57 sec, 37020.709248 + -1.798850 ->37018.910398 2024-05-21 22:21:09 \| GPUGRID \| [prio] -0.307547 rsf 0.500000 rt 37020.709248 rs 240748.213919 2024-05-21 22:21:09 \| GPUGRID \| [prio] -0.295697 rsf 0.500000 rt 35594.238733 rs 240748.213919 2024-05-21 22:21:19 \| \| choose_project(): 1716322879.711470 2024-05-21 22:21:19 \| GPUGRID \| [prio] -0.295697 rsf 0.500000 rt 35594.238733 rs 240748.213919 2024-05-21 22:21:19 \| GPUGRID \| [prio] -0.658863 rsf 0.200000 rt 37018.910398 rs 280930.394911 2024-05-21 22:21:19 \| \| [work_fetch] ------- start work fetch state ------- 2024-05-21 22:21:19 \| \| [work_fetch] target work buffer: 17280.00 + 25920.00 sec 2024-05-21 22:21:19 \| \| [work_fetch] --- project states --- 2024-05-21 22:21:19 \| GPUGRID \| [work_fetch] REC 37018.910 prio -0.659 can request work 2024-05-21 22:21:19 \| \| [work_fetch] --- state for CPU --- 2024-05-21 22:21:19 \| \| [work_fetch] shortfall 190218.61 nidle 0.00 saturated 24119.01 busy 0.00 2024-05-21 22:21:19 \| GPUGRID \| [work_fetch] share 0.000 no applications 2024-05-21 22:21:19 \| \| [work_fetch] --- state for NVIDIA GPU --- 2024-05-21 22:21:19 \| \| [work_fetch] shortfall 0.00 nidle 0.00 saturated 45029.53 busy 0.00 2024-05-21 22:21:19 \| GPUGRID \| [work_fetch] share 0.000 project is backed off (resource backoff: 28492.22, inc 38400.00) 2024-05-21 22:21:19 \| \| [work_fetch] --- state for Intel GPU --- 2024-05-21 22:21:19 \| \| [work_fetch] shortfall 43200.00 nidle 1.00 saturated 0.00 busy 0.00 2024-05-21 22:21:19 \| GPUGRID \| [work_fetch] share 0.000 project is backed off (resource backoff: 72285.30, inc 86400.00) 2024-05-21 22:21:19 \| \| [work_fetch] --- state for NVIDIA GeForce RTX 4080 Laptop GPU --- 2024-05-21 22:21:19 \| \| [work_fetch] shortfall 43200.00 nidle 1.00 saturated 0.00 busy 0.00 2024-05-21 22:21:19 \| GPUGRID \| [work_fetch] share 0.000 project is backed off (resource backoff: 75760.60, inc 86400.00) 2024-05-21 22:21:19 \| \| [work_fetch] --- state for Intel(R) UHD Graphics --- 2024-05-21 22:21:19 \| \| [work_fetch] shortfall 43200.00 nidle 1.00 saturated 0.00 busy 0.00 2024-05-21 22:21:19 \| GPUGRID \| [work_fetch] share 0.000 project is backed off (resource backoff: 4957.26, inc 86400.00) 2024-05-21 22:21:19 \| \| [work_fetch] ------- end work fetch state ------- 2024-05-21 22:21:19 \| GPUGRID \| choose_project: scanning 2024-05-21 22:21:19 \| GPUGRID \| can't fetch CPU: no applications 2024-05-21 22:21:19 \| GPUGRID \| can't fetch NVIDIA GPU: project is backed off 2024-05-21 22:21:19 \| GPUGRID \| can't fetch Intel GPU: project is backed off 2024-05-21 22:21:19 \| GPUGRID \| can't fetch NVIDIA GeForce RTX 4080 Laptop GPU: project is backed off 2024-05-21 22:21:19 \| GPUGRID \| can't fetch Intel(R) UHD Graphics: project is backed off 2024-05-21 22:21:19 \| \| [work_fetch] No project chosen for work fetch
	ID: 61520 \| Rating: 0 \| rate: / Reply Quote

Keith Myers Send message Joined: 13 Dec 17 Posts: 1360 Credit: 7,911,840,522 RAC: 5,650,417 Level Scientific publications	Message 61522 - Posted: 22 May 2024 \| 2:01:30 UTC
	You have a couple things wrong I believe. First you are in contention with WCG as it is also requesting work. Second, you should remove the additional work request limit or set it to only 0.1 days. Then you need to bump up your primary daily work request from the 17260 seconds as the GPUGrid work request of ~35,000 seconds exceeds that. You shouldn't need that much time to finish any of the QChem tasks, but if you are also requesting Python, ATM or acemd3 tasks, then you need the additional time to finish computation on those tasks. But the 3080 laptop gpu should be able to do those in those times as long as you don't have a lot of cpu contention. That's an issue with laptops that share cpu resources with Intel igpus which your output shows as also being seen. I'd reduce your cpu contention to 50% to adequately support your Nvidia gpu. Also make sure you have done a recent cpu benchmark to profile your host and have its processing speed coincident hopefully with the task rsc_fpops_estimate values the tasks are set up with.
	ID: 61522 \| Rating: 0 \| rate: / Reply Quote

Endasil Send message Joined: 22 May 15 Posts: 4 Credit: 6,210,000 RAC: 0 Level Scientific publications	Message 61531 - Posted: 25 May 2024 \| 20:58:04 UTC - in response to Message 61522. Last modified: 25 May 2024 \| 20:58:38 UTC
	Sorry for the multipost, I did not intend to post my message here , I must had multiple tabs open and mixed stuff up confused. But since the reply is here: How do I increase my primary daily work request to 35,000 seconds? How do I reduce my CPU contention to 50%?
	ID: 61531 \| Rating: 0 \| rate: / Reply Quote

Keith Myers Send message Joined: 13 Dec 17 Posts: 1360 Credit: 7,911,840,522 RAC: 5,650,417 Level Scientific publications	Message 61532 - Posted: 26 May 2024 \| 6:05:49 UTC - in response to Message 61531.
	Sorry for the multipost, I did not intend to post my message here , I must had multiple tabs open and mixed stuff up confused. But since the reply is here: How do I increase my primary daily work request to 35,000 seconds? How do I reduce my CPU contention to 50%? You change your work request in the Computing Preferences on the website or in the Computing Preferences locally in the Boinc Manager. There are 86,400 seconds in a day. Your current preferences are set for 0.2 days. If you want 35,000 seconds of work, then set your daily request to 0.4 or 0.5 days. Same places to set your cpu usage. Instead of using 100% of your cpu cores, set to use 50% of your cpu cores. https://www.gpugrid.net/prefs.php?subset=global
	ID: 61532 \| Rating: 0 \| rate: / Reply Quote

Post to thread

Message boards : Frequently Asked Questions (FAQ) : CPU Only

	About	Science	Volunteers	Performance	Forum	Join us	Donate