I wonder if could be 2 queues, one for ap , sec for gpu - all mb go for gpu , and we could set cpu cores for queues ap , and units gpu with queues mb , and if ap queues would be empty they could fetch from sec queues mb - that could solve problems with gpu idle - and cpu idle , and for cpu - eliminate situation when work at 3 wu on 2 cores
Look this thread about CUDA app freezings: http://lunatics.kwsn.net/12-gpu-crunching/vbscript-fights-cuda.0.html
where can I find Mark app_info.xml ? , should I use Raistmer apps or use standalone optimized ap and mb ak v8 ?
I have all errors on cuda mb , even if i edited platform to windows_x86_64 , but what i need change it uses setiathome_6.08_windows_intelx86__cuda.exe , where can i download 64bit version?
Here V10/11 CUDA MB update (old filename used to simplify updating).Just replace your current CUDA MB executable with this one.It has some performance gain due to decreased CPU consumption during startup.ADDON: Also, it should do more smooth %done increase indication.
Just switched to the optimized apps and things seem to be running quite well. I don't have any AP workunits yet, I disabled them to make sure I got some MB for CUDA. Hopefully AP work units will go faster as well. I've only had two notable issues so far.1. Trying to track down the latest release + update + get it all together is rather confusing2. I'm on a laptop and had to update my graphics drivers via laptopvideotogo.comGiven the fairly small numbers of downloads I see right now I may be able to host your apps on my server if your interested Raistmer. Possibly could have a centralized status page as well for latest version information.Have you considered releasing optimized apps for platforms other than windows? I'm crunching MB units in about 3 hours on a Core2 6320 (1.86GHz) on FreeBSD and it would be nice if I could crunch a bit faster.
Does this have VLAR kill?