Yeah, my first post was from June 2011, so 9 months ago. I am just surprised that it took so long to get to market, that's all.
Yeah, my first post was from June 2011, so 9 months ago. I am just surprised that it took so long to get to market, that's all.
I think that items like this should be placed in emails or somehow made more accessible for new members and less well informed members to find and act on.
Being a greenhorn myself, I often get confused or have ideas and wonder where to find the specific information and when I look it gets me even more confused at times sifting through the chafe to find the good stuff.
I don't read old posts and searching for info about stuff like this in the forum is painful at best. Things get lost in there.
Any good information that allows a member to easily increase the output is great news and welcome at that.
I am glad John P. Myers threw this out there again. Thanks! To all of you gurus who share with us non gurus.
@kaptainkarl1: You can call me JPM like everyone else. It's much easier to type
A special note/reminder for those adding this (or any other) GPU to their systems:
If you are getting one and it will be the only GPU in your system, you can skip this.
If you are getting one and adding it to a system that already has 1 AMD/ATI GPU, but no other Nvidia GPUs, you can skip this
If you are getting more than 1 of these, or are adding 1 or more of these to a system that already has 1 or more Nvidia GPUs, please take note
By default, BOINC will not use every GPU in your system. It leaves one untouched, thinking maybe you need it for something else i guess. I dunno. I don't get it personally and it seems a bit silly This is something BOINC does and is not the fault of your GPU manufacturer, not a driver issue and nothing to do with a bug in your OS.
Fortunately, preventing it from happening is really really simple.
** Go to C:\ProgramData\BOINC\
look for a file named cc_config.xml
If it is there, simply add <use_all_gpus>1</use_all_gpus> in between the <options> tags, save and you're done.
If it is not there (often it isn't unless you've made this file before), then open Notepad or any crappy text editor and copy/paste the following:
Then simply save it as C:\ProgramData\BOINC\cc_config.xmlCode:<cc_config> <options> <use_all_gpus>1</use_all_gpus> </options> </cc_config>
You're done!
**Normally, the C:\ProgramData directory (i think they call them "folders" these days? bleh) is hidden. You may have to type it in manually to go there. To make all files unhidden: (in Win 7)
open Folder Options
click the View tab
put a bullet in the circle next to Show Hidden Files, Folders and Drives
click Apply/OK and you're done
EDIT: After creating cc_config.xml, you must either restart BOINC for it to take effect, or in BOINC, click Advanced >> Read config file (call me superstitious, but restarting feels safer)
Last edited by John P. Myers; 03-24-12 at 05:38 PM.
This situation happens *only* when there are two different GPUs with *significantly* different crunching power. BOINC uses the fastest GPU for calculating work requests x the number of GPUs. So for example, assume you have a 9500 GT and a GTX 580. BOINC would automatically calculate work fetch based on the 580, times two cards. That could end up downloading so much work that it cannot return it all in time, depending on the size of your cache and the deadlines of the project. Because the 9500 is not really a 580. So by default, if BOINC detects a large enough difference between the two (or however many) GPUs, it will ignore one of them. I forget which, the faster or slower. Anyway, if you want, you can over ride this behavior with the method you describe. But understand that you are taking a risk of over-downloading tasks, and being late. So be sure to keep an eye on your queue, and keep it small until you have a good handle on the right cache size to keep without issues.
That is absolutely true. Thank you for pointing that out. There are 2 ways to prevent it:
1) click Tools >> Computing Preferences >> Network Usage and change Additional Work Buffer to some value less than 0.50 (i personally keep mine at 0.25). The higher the GPU count, the lower this number may need to be. Also during some competitions where buffering is permitted, you may need to keep an eye on things.
2) F$ made a thread on how to use another cc_config option which prevents certain GPUs from running certain projects so BOINC will not download the tasks for it as it will the others. This excluded GPU can then be set to run a different project, though you will have to exclude this different project from the other GPUs. See here: http://www.setiusa.us/showthread.php...ll=1#post24981
EDIT: The reason i keep my Additional Work Buffer that low has nothing to do with worrying about getting too many tasks to turn in on time. It's just a personal preference to keep my task list as uncluttered as possible. Back when i first started crunching, i used to like to see hundreds of work units in the queue. Now, not so much. It doesn't really matter either way (except someone, somewhere probably won't be happy waiting 10 days for validation), just don't be late
Last edited by John P. Myers; 03-24-12 at 07:23 PM.
Assume info. Thanks you all for posting this thread, for having it posted to the front and for how to get it to work well. I am a little overspent at the moment, put I I have ear marked the thread and will fill empty slots later on.
Aye agree thanks for all the info but the project i have atm is trying to see if i can use the x1 PCIE slot like someone else was doing on the hydra project, sorry can't recall off hand which team member but it really sounded cool to think i could fit a few extra decent cards on those x1 slots that are just collecting dust. So I purchased a few x1 extention cables and will let ya know how it goes not sure if the x1 will be enough bandwidth but since its for older cards im hoping it will work.