Message boards :
Number crunching :
I surrender...
Message board moderation
Author | Message |
---|---|
Send message Joined: 7 Nov 07 Posts: 13 Credit: 20,143,762 RAC: 0 |
I recently received a new computer with a GPU that could crunch MW, so I merrily downloaded Boinc and signed up for MW. Sure enuf downloaded a bunch of MW GPU units and a half dozen or so of the N-body 0.66 Multi-threaded ones as well. I figured all was well as the new guy is a 6 core (12 w/HT enabled), but... The GPU units all crunched but not one of the MT WU's. I downloaded some Rosetta and some Enigma WU's just to see what would happen and they all crunched merrily to completion, but the N-body guys haven't budged a bit. Any ideas??? Computer is here: 298955 |
Send message Joined: 12 Nov 07 Posts: 2425 Credit: 524,164 RAC: 0 |
All except the 2 MW that say they are (cuda) are Cpu wus. It looks like you are only getting 1 GPU wu at a time. You may need to tweek your settings if you want to run multiple GPu wus. Doesn't expecting the unexpected make the unexpected the expected? If it makes sense, DON'T do it. |
Send message Joined: 7 Nov 07 Posts: 13 Credit: 20,143,762 RAC: 0 |
un-hunh... All of the GPU units processed (and most merrilly went off to la-la-land)... It's the multi-treaded CPU units that refuse to run at all... |
Send message Joined: 1 Sep 08 Posts: 204 Credit: 219,354,537 RAC: 0 |
Any error message? What's in the BOINC messages if you start it, with only MW allowed for the CPU? MrS Scanning for our furry friends since Jan 2002 |
Send message Joined: 7 Nov 07 Posts: 13 Credit: 20,143,762 RAC: 0 |
No error messages that I can see in the Log, every time a GPU unit finished there was a schedule request for more GPU work, usually only one more unit downloaded with a message saying the unit had reached the max-tasks-in-process limit. I have one more Rosetta WU running on the CPU at 99.84 pct. Once that one is finished and uploaded I will exit Boinc and bounce the machine to see if maybe the GPU WU's locked something up. |
Send message Joined: 7 Nov 07 Posts: 13 Credit: 20,143,762 RAC: 0 |
Oh well... took a couple of hours to do the shut-down/re-start as Mr Softie decided he just had to co-opt the machine for updates. Situation is still the same, seven CPU n-body 0.66 MT WU's sitting there all by themselves in 'ready to start' status, but none of them running. I am going to Abort these WU's and see what happens with another batch later. At least that way they will get re-sent to someone else. |
Send message Joined: 23 Jan 08 Posts: 7 Credit: 156,801,302 RAC: 7,344 |
I am having similar non-results. Have disconnected and reconnected without success. Any suggestions??? Thanks dp Stderr output <core_client_version>6.12.26</core_client_version> <![CDATA[ <message> Incorrect function. (0x1) - exit code 1 (0x1) </message> <stderr_txt> Error loading Lua script 'astronomy_parameters.txt': [string "number_parameters: 4..."]:1: '<name>' expected near '4' Error reading astronomy parameters from file 'astronomy_parameters.txt' Trying old parameters file Using SSE3 path Failed to get CAL device attributes: Parameter passed in is invalid (CAL_RESULT_INVALID_PARAMETER) Error getting device information: Parameter passed in is invalid (CAL_RESULT_INVALID_PARAMETER) Failed to get CAL info: Parameter passed in is invalid (CAL_RESULT_INVALID_PARAMETER) Failed to setup CAL 03:33:26 (5972): called boinc_finish </stderr_txt> |
Send message Joined: 1 Sep 08 Posts: 204 Credit: 219,354,537 RAC: 0 |
dp, your problem is very different. It looks like your ATI driver is way out of date. pieface, it could be that the n_body WUs require something that is not yet available to BOINC (be it system memory or whatever). Aborting these and downloading more WUs will very probably change nothing. Could you post the messages upon BOINC startup, preferably with only the MW WUs on the machine? MrS Scanning for our furry friends since Jan 2002 |
Send message Joined: 7 Nov 07 Posts: 13 Credit: 20,143,762 RAC: 0 |
Hmmm... Sorry, I had already aborted them when I checked the message boards again. I ran some different projects overnite, still getting the new machine identified to all the other projects I run and didn't encounter any problems. When I came back around to MW this afteroon it merrilly downloaded the whole spectrum of types, GPU running in around 10 mins, multi-thread running very quickly also (using all 10 'cores' that I let Boinc have) and now the single threaded ones are starting to run their chunk of time. I have no idea what it could have been, probably something related to running with a useable GPU for the first time got it confused. Thanks for all the time!! Hopefully this was a one-shot wonder and won't happen again... |
Send message Joined: 23 Jan 08 Posts: 7 Credit: 156,801,302 RAC: 7,344 |
Thanks for getting back. Updated the driver, detached and re-attached to the MW project, but still no luck. Any suggestions? dp |
Send message Joined: 28 Feb 10 Posts: 120 Credit: 109,840,492 RAC: 0 |
Please look at http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=2468 Here is a solution which has worked by others. For futher questions don't be shy to PM me. regards franz |
Send message Joined: 1 Sep 08 Posts: 204 Credit: 219,354,537 RAC: 0 |
You're getting a different error now - moving in the right direction :) When I switched from 0.62 to 0.82 I had to manually reset the "result duration correction factor" to not get the "maximum elapsed time exceeded" error. Don't know what the solution in the other thread is - just ask if you need further assistance. MrS Scanning for our furry friends since Jan 2002 |
©2024 Astroinformatics Group