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 : Proth Prime Search : Suspending a PPS (Sieve) cuda32 tasks seems to cause a computation error

Author Message
VapourLock
Avatar
Send message
Joined: 12 Feb 11
Posts: 8
ID: 85904
Credit: 10,881,291
RAC: 0
PPS LLR Silver: Earned 100,000 credits (232,320)SoB LLR Bronze: Earned 10,000 credits (74,453)SGS LLR Gold: Earned 500,000 credits (804,571)321 Sieve Bronze: Earned 10,000 credits (13,319)PPS Sieve Turquoise: Earned 5,000,000 credits (9,754,711)
Message 33623 - Posted: 1 Mar 2011 | 18:57:20 UTC
Last modified: 1 Mar 2011 | 19:07:40 UTC

Each night when I go to bed, I turn on Proth Prime Search (Sieve)'s cuda32 work units (on a GTX285 system) so they'll have a bunch of hours interrupted.

Takes about an hour and a half to finish each work unit.

When I wake up in the morning, I then suspend the cuda32 tasks due to UI lag issues, and then unsuspend them again that night. (and so on each night)

Something I've noticed is that I'm occasionally getting "Computation error" reported for work units, and they _generally_ seem to be for the work units that were partially completed when they were suspended.

It's almost like the suspend/unsuspend action on an in progress cuda32 unit is corrupting it somehow, or causing it's internal validation to go wrong.

Has anyone else seen or noticed this?

Profile Michael GoetzProject donor
Volunteer moderator
Project administrator
Avatar
Send message
Joined: 21 Jan 10
Posts: 13261
ID: 53948
Credit: 223,153,378
RAC: 29,479
The "Shut up already!" badge:  This loud mouth has mansplained on the forums over 10 thousand times!  Sheesh!!!Discovered the World's First GFN-19 prime!!!Discovered 1 mega primeFound 1 prime in the 2018 Tour de PrimesFound 1 prime in the 2019 Tour de PrimesFound 1 prime in the 2020 Tour de Primes321 LLR Ruby: Earned 2,000,000 credits (2,822,730)Cullen LLR Ruby: Earned 2,000,000 credits (2,005,249)ESP LLR Turquoise: Earned 5,000,000 credits (5,009,577)Generalized Cullen/Woodall LLR Ruby: Earned 2,000,000 credits (2,145,754)PPS LLR Jade: Earned 10,000,000 credits (12,903,521)PSP LLR Turquoise: Earned 5,000,000 credits (5,197,957)SoB LLR Sapphire: Earned 20,000,000 credits (34,291,181)SR5 LLR Jade: Earned 10,000,000 credits (10,007,110)SGS LLR Ruby: Earned 2,000,000 credits (2,276,011)TRP LLR Ruby: Earned 2,000,000 credits (4,089,728)Woodall LLR Ruby: Earned 2,000,000 credits (2,195,123)321 Sieve Jade: Earned 10,000,000 credits (10,061,196)Cullen/Woodall Sieve (suspended) Ruby: Earned 2,000,000 credits (4,170,256)Generalized Cullen/Woodall Sieve (suspended) Turquoise: Earned 5,000,000 credits (5,059,304)PPS Sieve Sapphire: Earned 20,000,000 credits (22,885,121)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Amethyst: Earned 1,000,000 credits (1,035,522)TRP Sieve (suspended) Ruby: Earned 2,000,000 credits (2,051,121)AP 26/27 Jade: Earned 10,000,000 credits (10,118,303)GFN Emerald: Earned 50,000,000 credits (72,383,585)PSA Jade: Earned 10,000,000 credits (12,445,029)
Message 33630 - Posted: 1 Mar 2011 | 19:11:27 UTC - in response to Message 33623.
Last modified: 1 Mar 2011 | 19:12:41 UTC

Each night when I go to bed, I turn on Proth Prime Search (Sieve)'s cuda32 work units (on a GTX285 system) so they'll have a bunch of hours interrupted.

Takes about an hour and a half to finish each work unit.

When I wake up in the morning, I then suspend the cuda32 tasks due to UI lag issues, and then unsuspend them again that night. (and so on each night)

Something I've noticed is that I'm occasionally getting "Computation error" reported for work units, and they _generally_ seem to be for the work units that we're partially completed when they were suspended.

It's almost like the suspend/unsuspend action on an in progress cuda32 unit is corrupting it somehow, or causing it's internal validation to go wrong.

Has anyone else seen or noticed this?



Nope, I'm able to suspend CUDA work without the WUs crashing.

Furthermore, 90 minutes sounds way too slow for a 285. Until it died, I was running a 280, which is slower, and the WUs were, if I remember correctly, in the 15 minute range. WU size has been doubled since then, I think, which means about a 30 minute run time. 90 minutes doesn't seem right.

So, something definitely isn't right. I'd start with the drivers -- are you running the most recent drivers? I don't know anything about BOINC under Darwin, so I can't be a lot of help with specifics, unfortunately. It's possible that this is normal behavior with Darwin.
____________
My lucky number is 75898524288+1

VapourLock
Avatar
Send message
Joined: 12 Feb 11
Posts: 8
ID: 85904
Credit: 10,881,291
RAC: 0
PPS LLR Silver: Earned 100,000 credits (232,320)SoB LLR Bronze: Earned 10,000 credits (74,453)SGS LLR Gold: Earned 500,000 credits (804,571)321 Sieve Bronze: Earned 10,000 credits (13,319)PPS Sieve Turquoise: Earned 5,000,000 credits (9,754,711)
Message 33631 - Posted: 1 Mar 2011 | 19:17:33 UTC - in response to Message 33630.

Yeah, this is running the most recent drivers from NVIDIA, and everything else works with it fine.

So, I guess it might just be a Darwin thing. :/

Message boards : Proth Prime Search : Suspending a PPS (Sieve) cuda32 tasks seems to cause a computation error

[Return to PrimeGrid main page]
DNS Powered by DNSEXIT.COM
Copyright © 2005 - 2020 Rytis Slatkevičius (contact) and PrimeGrid community. Server load 1.18, 1.47, 1.80
Generated 15 Aug 2020 | 19:14:22 UTC