VLAR's to CUDA, is that something new?

Message boards : SETI@home Enhanced : VLAR's to CUDA, is that something new?
Message board moderation

To post messages, you must log in.

AuthorMessage
Grumpy Swede
Volunteer tester
Avatar

Send message
Joined: 10 Mar 12
Posts: 1700
Credit: 13,216,373
RAC: 0
Sweden
Message 54354 - Posted: 9 Jun 2015, 7:58:49 UTC
Last modified: 9 Jun 2015, 8:15:39 UTC

Is something broken, or have BETA started to send VLAR's to CUDA Nvidia cards now?

I've got a whole load of them on my new GTP980, both CUDA42 and 50.

http://setiweb.ssl.berkeley.edu/beta/show_host_detail.php?hostid=75292

Edit: I'll abort them, they run soo slooooow :-)
ID: 54354 · Report as offensive
William
Volunteer tester
Avatar

Send message
Joined: 14 Feb 13
Posts: 606
Credit: 588,843
RAC: 0
Message 54356 - Posted: 9 Jun 2015, 9:09:44 UTC

IIRC that change was made a while back, possibly after we got x41zc.
Basically testing how well fast cards cope with VLAR now.

yes, they are slow, but GPUs no longer choke on them.

I don't think we did a full review on the user experience and the general performance and I don;t think it was something recommended for main.

But with one thing and another that testing may well have been lost from the collective agenda.

One of the others might remember details.
A person who won't read has no advantage over one who can't read. (Mark Twain)
ID: 54356 · Report as offensive
Grumpy Swede
Volunteer tester
Avatar

Send message
Joined: 10 Mar 12
Posts: 1700
Credit: 13,216,373
RAC: 0
Sweden
Message 54357 - Posted: 9 Jun 2015, 9:15:21 UTC - in response to Message 54356.  

IIRC that change was made a while back, possibly after we got x41zc.
Basically testing how well fast cards cope with VLAR now.

yes, they are slow, but GPUs no longer choke on them.

I don't think we did a full review on the user experience and the general performance and I don;t think it was something recommended for main.

But with one thing and another that testing may well have been lost from the collective agenda.

One of the others might remember details.

Hmm, I thought they reversed that test, and stopped vlars again from Nvidia cards.

Ah well, I'm sure someone will remember, because I sure don't :-)
ID: 54357 · Report as offensive
Richard Haselgrove
Volunteer tester

Send message
Joined: 3 Jan 07
Posts: 1451
Credit: 3,272,268
RAC: 0
United Kingdom
Message 54365 - Posted: 10 Jun 2015, 7:08:42 UTC

IIRC, the test was specifically when the newer Kepler architecture cards were released, to find out if the CUDA application was any more successful at VLAR than it had been on older cards. It wasn't, so the change wasn't duplicated across to the main project. I don't think you'll get VLARs unless you have one of those recent cards installed.

(see message 46039)
ID: 54365 · Report as offensive
Grumpy Swede
Volunteer tester
Avatar

Send message
Joined: 10 Mar 12
Posts: 1700
Credit: 13,216,373
RAC: 0
Sweden
Message 54366 - Posted: 10 Jun 2015, 7:57:13 UTC - in response to Message 54365.  

IIRC, the test was specifically when the newer Kepler architecture cards were released, to find out if the CUDA application was any more successful at VLAR than it had been on older cards. It wasn't, so the change wasn't duplicated across to the main project. I don't think you'll get VLARs unless you have one of those recent cards installed.

(see message 46039)

Ah, Ok. Well the GTX980 is one of "those recent cards", so I guess that answers the question.

I'll let them run, if I get a new batch.
ID: 54366 · Report as offensive
William
Volunteer tester
Avatar

Send message
Joined: 14 Feb 13
Posts: 606
Credit: 588,843
RAC: 0
Message 54369 - Posted: 10 Jun 2015, 10:24:59 UTC

Plus, we need that setting to test the new apps in the first place.
Benches only tell you so much.

ah. cc 3.0+ Don't need to press Eve into service then.

@Richard Thanks for digging that out.
A person who won't read has no advantage over one who can't read. (Mark Twain)
ID: 54369 · Report as offensive

Message boards : SETI@home Enhanced : VLAR's to CUDA, is that something new?


 
©2022 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.