Am I doing something wrong regarding resource share?

Author	Message
Neal Chantrill Send message Joined: 17 Jan 09 Posts: 98 Credit: 72,182,367 RAC: 0	Message 33869 - Posted: 27 Nov 2009, 23:19:18 UTC I have Collatz at 1 and Milkyway at 500 yet Collatz seems to get the lion share of work. This is GPU by the way. I'm a little confused. Click here for The Clangers forum! ID: 33869 · Rating: 0 · rate: / Reply Quote

The Gas Giant Send message Joined: 24 Dec 07 Posts: 1947 Credit: 240,884,648 RAC: 0	Message 33882 - Posted: 28 Nov 2009, 6:51:35 UTC - in response to Message 33869. I have Collatz at 1 and Milkyway at 500 yet Collatz seems to get the lion share of work. This is GPU by the way. I'm a little confused. What you're experiencing is a bug that the devs have not acknowledged is a problem. Paul D Buck and myself have previously reported it, but alas. BOINC is working on a FIFO basis for GPU work. IF your cache is small enough then BOINC will come close to honouring resource share. ID: 33882 · Rating: 0 · rate: / Reply Quote

Neal Chantrill Send message Joined: 17 Jan 09 Posts: 98 Credit: 72,182,367 RAC: 0	Message 33883 - Posted: 28 Nov 2009, 8:05:51 UTC Thanks. :o) Click here for The Clangers forum! ID: 33883 · Rating: 0 · rate: / Reply Quote

verstapp Send message Joined: 26 Jan 09 Posts: 589 Credit: 497,834,261 RAC: 0	Message 33888 - Posted: 28 Nov 2009, 11:14:04 UTC >IF your cache is small enough But isn't there another thread in this topic complaining about too-small caches? :p Cheers, PeterV . ID: 33888 · Rating: 0 · rate: / Reply Quote

Paul D. Buck Send message Joined: 12 Apr 08 Posts: 621 Credit: 161,934,067 RAC: 0	Message 33936 - Posted: 29 Nov 2009, 11:53:04 UTC - in response to Message 33888. >IF your cache is small enough But isn't there another thread in this topic complaining about too-small caches? :p We mean the local cache settings to about 0.1 days size ... Collatz will fill your full needs, GPU Grid and MW will not (usually) ... in any case, BOINC will ask to fill the cache in many cases and then for GPU work will work on the tasks in strict FIFO order ... As a consequence, if you have 1.5 days as your cache settings you will get several hundred tasks from collatz all at once ... then work contentedly for that 1.5 days ... then try to get MW work ... if none is available, you will get another 1.5 days of Collatz .... so, minor glitches on work feeding will cause Collatz to seem to dominate. Same can be true of GPU Grid if that is an alternative. as TGG noted this issue has been raised ... and ignored... ID: 33936 · Rating: 0 · rate: / Reply Quote

Nick Send message Joined: 1 Nov 09 Posts: 21 Credit: 18,543,759 RAC: 0	Message 33943 - Posted: 29 Nov 2009, 16:24:07 UTC I wonder if my problem is related. I have a Quadcore system and I am running SETI@Home, Cosmology@Home and Milkyway@Home. SETI@Home uses only the GPU with both Cosmology@Home and Milkyway@Home set to only use the CPUs. Seti@Home is set to use 50% of the resources and the other two set to use 25% of the resources, yet Milkyway@Home will not run unless I suspend Cosmology@Home. Any one have any idea what's wrong? ID: 33943 · Rating: 0 · rate: / Reply Quote