seems like it waits to run out of work before requesting more...
Are you running multiple BOINC projects on that host? In more than one instance when my Einstein work requests dried-up, I found that suspending Einstein work in order to force some SETI work to run would rather shortly lead to resumed work requests (after I undid the suspensions of course) on the Einstein side. This is particularly vexing to me as I allocate SETI either 0% or 1% on my boxes. As my own personal resolution, I have suspended download of SETI work units, so I'm essentially using SETI as a manual backup, not automatic.
I was wondering... one box is fetching WU's fine on 6.12.34 but the 560 Ti box isn't... seems like it waits to run out of work before requesting more...
I wonder if that 6.13.0 needs to be changed... hmm, have to read up on this stuff sometime..
:D
The line will have no effect on whether/when you request work.
BTW, did you ever say why you felt the need to run with an app_info.xml file? Sure, it's a good idea to understand it if you're going to use it, but it's usually easier to do neither... :-)
I use it to run 2 WU's per GPU as it gets a higher output that way. All my boxes now running BOINC 7.0.25 and the next thing to do is change my preferences to use 0.5 GPU's on the main boxes and use an app_info to spec 1.0 GPU's for the cheapo 9800 GT box.. I think that would work out better for me..
I was wondering... one box is fetching WU's fine on 6.12.34 but the 560 Ti box isn't... seems like it waits to run out of work before requesting more...
That's probably because you are running BOINC 7.0.x and haven't adjusted your cache settings to allow for the significant changes in function of those settings. If you have the first setting (connect to internet ... (CI)) set close to or equal to zero, your cache will run down to almost empty before new work will be fetched.
For BOINC 7.0.x and above, you should now think of the first setting as a 'low water mark'. The second setting (extra days ... (ED)) is added to the low water mark to give the 'high water mark'. If you want an exact x day cache, set the CI setting to x and set the ED setting to something very low like 0.01 - (not sure if zero is allowed - probably it is). If you would like a cache that oscillates between x days and x+y days, set CI to x and ED to y.
RE: seems like it waits to
)
Are you running multiple BOINC projects on that host? In more than one instance when my Einstein work requests dried-up, I found that suspending Einstein work in order to force some SETI work to run would rather shortly lead to resumed work requests (after I undid the suspensions of course) on the Einstein side. This is particularly vexing to me as I allocate SETI either 0% or 1% on my boxes. As my own personal resolution, I have suspended download of SETI work units, so I'm essentially using SETI as a manual backup, not automatic.
RE: RE: I was
)
I use it to run 2 WU's per GPU as it gets a higher output that way. All my boxes now running BOINC 7.0.25 and the next thing to do is change my preferences to use 0.5 GPU's on the main boxes and use an app_info to spec 1.0 GPU's for the cheapo 9800 GT box.. I think that would work out better for me..
8-)
RE: ...and use an app_info
)
You could achieve the same result with another host venue for that box.
Gruß,
Gundolf
Computer sind nicht alles im Leben. (Kleiner Scherz)
RE: RE: ...and use an
)
Indeed I could and I plum fergit about that!
Thanks for tweaking my brain!
Dadgum brain of mine seems to suffer since passing 55 yo...
8-)
RE: I was wondering... one
)
That's probably because you are running BOINC 7.0.x and haven't adjusted your cache settings to allow for the significant changes in function of those settings. If you have the first setting (connect to internet ... (CI)) set close to or equal to zero, your cache will run down to almost empty before new work will be fetched.
For BOINC 7.0.x and above, you should now think of the first setting as a 'low water mark'. The second setting (extra days ... (ED)) is added to the low water mark to give the 'high water mark'. If you want an exact x day cache, set the CI setting to x and set the ED setting to something very low like 0.01 - (not sure if zero is allowed - probably it is). If you would like a cache that oscillates between x days and x+y days, set CI to x and ED to y.
Cheers,
Gary.
A preview of the new dual
)
A preview of the new dual GK104 Kepler was brought online this evening:
http://www.hardocp.com/article/2012/04/28/geforce_gtx_690_perfection_inside_out
The card has a boost clock up to 1015 MHz for the GPUs, 10-phase power, vapor chamber cooling, and built for a max 300w power draw.
It only has 2GB VRAM at 256
)
It only has 2GB VRAM at 256 bus. Remember, it's in SLI, so it's shared.
Next waiting game is the anticipated 685 or 780 coming out holiday season.