PrimeGrid
Please visit donation page to help the project cover running costs for this month

Toggle Menu

Join PrimeGrid

Returning Participants

Community

Leader Boards

Results

Other

drummers-lowrise

Advanced search

Message boards : Number crunching : Collab and Suspending GPU

Author Message
Profile jshriverProject donor
Send message
Joined: 25 Jul 19
Posts: 61
ID: 1162187
Credit: 58,115,421
RAC: 179,700
321 LLR Ruby: Earned 2,000,000 credits (2,076,316)Cullen LLR Amethyst: Earned 1,000,000 credits (1,727,522)ESP LLR Amethyst: Earned 1,000,000 credits (1,041,879)Generalized Cullen/Woodall LLR Amethyst: Earned 1,000,000 credits (1,090,339)PPS LLR Turquoise: Earned 5,000,000 credits (5,058,884)PSP LLR Amethyst: Earned 1,000,000 credits (1,237,964)SoB LLR Amethyst: Earned 1,000,000 credits (1,013,853)SR5 LLR Amethyst: Earned 1,000,000 credits (1,012,112)SGS LLR Ruby: Earned 2,000,000 credits (2,291,198)TRP LLR Ruby: Earned 2,000,000 credits (4,300,773)Woodall LLR Amethyst: Earned 1,000,000 credits (1,094,441)321 Sieve Turquoise: Earned 5,000,000 credits (5,361,499)PPS Sieve Turquoise: Earned 5,000,000 credits (9,664,657)AP 26/27 Turquoise: Earned 5,000,000 credits (5,781,490)GFN Jade: Earned 10,000,000 credits (13,349,191)PSA Ruby: Earned 2,000,000 credits (2,036,020)
Message 143172 - Posted: 9 Sep 2020 | 10:44:32 UTC

Anyone else have issues with Google Collab constantly saying "computer in use, suspending GPU"?

I ended up disabling that feature from my preferences entirely, yet this has been happening the past week.

I've also ran htop to see what was running and nothing comes up aside from the boinc client.
So its not another non-boinc process consuming resources.

Is there a way to forge gpu alway on from boinccmd? Using boinccmd --set_gpu_mode always currently.

Toughts?

Dr Who Fan
Avatar
Send message
Joined: 26 Sep 06
Posts: 108
ID: 3556
Credit: 15,055,347
RAC: 1,955
321 LLR Silver: Earned 100,000 credits (101,760)Cullen LLR Silver: Earned 100,000 credits (120,652)ESP LLR Silver: Earned 100,000 credits (102,526)Generalized Cullen/Woodall LLR Silver: Earned 100,000 credits (115,730)PPS LLR Gold: Earned 500,000 credits (863,322)PSP LLR Silver: Earned 100,000 credits (132,546)SoB LLR Silver: Earned 100,000 credits (433,700)SR5 LLR Silver: Earned 100,000 credits (110,499)SGS LLR Silver: Earned 100,000 credits (162,863)TRP LLR Silver: Earned 100,000 credits (100,020)Woodall LLR Silver: Earned 100,000 credits (103,446)321 Sieve Ruby: Earned 2,000,000 credits (2,006,012)Generalized Cullen/Woodall Sieve (suspended) Ruby: Earned 2,000,000 credits (2,740,001)PPS Sieve Turquoise: Earned 5,000,000 credits (5,066,749)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Silver: Earned 100,000 credits (119,708)TRP Sieve (suspended) Gold: Earned 500,000 credits (527,925)AP 26/27 Gold: Earned 500,000 credits (788,467)GFN Amethyst: Earned 1,000,000 credits (1,445,029)
Message 143184 - Posted: 9 Sep 2020 | 16:27:52 UTC

I think Google is messing around with it!

My Google Colab (free version) has been been acting very weird lately. For me, on and off about the last almost two weeks. Here is what I have been experiencing:
* Suddenly last week it won't let me run my Wu-Prop NCI project concurrent with any other projects though nothing has changed in my code / images that have been successfully running it past many months.
* No GPU instance available or if I get a GPU instance the session only runs for 1 to maximum of 3 hours before stopping
* CPU only instance normally are running about 6 hours.
* CPU + GPU instance hours now vary all over the place but no more than 4 hours before stopping.
*Recently I had a CPU ONLY instance running for almost 13 hours on the SR5/SR Base project before stopping.

____________

Post to thread

Message boards : Number crunching : Collab and Suspending GPU

[Return to PrimeGrid main page]
DNS Powered by DNSEXIT.COM
Copyright © 2005 - 2020 Rytis Slatkevičius (contact) and PrimeGrid community. Server load 2.30, 2.12, 2.23
Generated 1 Nov 2020 | 2:50:15 UTC