Deprecated: Function get_magic_quotes_gpc() is deprecated in /disks/centurion/b/carolyn/b/home/boincadm/projects/beta/html/inc/util.inc on line 663
No consensus

No consensus

Message boards : SETI@home Enhanced : No consensus
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Ananas
Volunteer tester

Send message
Joined: 21 Jun 05
Posts: 43
Credit: 155,681
RAC: 0
Germany
Message 41556 - Posted: 25 Nov 2011, 3:11:11 UTC
Last modified: 25 Nov 2011, 3:28:07 UTC

ATI 365 vs. nVidia 365 vs. CPU 6.97 - 3 program versions (the two GPU seem to be the same version but different routines used), 3 different results

http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=3596456

p.s.:

The spike count differs between both GPU (14) and CPU (15)

The sort order of the spike listing is different between the GPUs, the score differs also (CPU doesn't have that)
ID: 41556 · Report as offensive
Profile Mike
Volunteer tester
Avatar

Send message
Joined: 16 Jun 05
Posts: 2530
Credit: 1,074,556
RAC: 0
Germany
Message 41559 - Posted: 25 Nov 2011, 9:46:15 UTC


Power of some signal varies too much.

Should be this spike.

power: 24.001, index(bin)=83950, i guess.


With each crime and every kindness we birth our future.
ID: 41559 · Report as offensive
Josef W. Segur
Volunteer tester

Send message
Joined: 14 Oct 05
Posts: 1137
Credit: 1,848,733
RAC: 0
United States
Message 41568 - Posted: 26 Nov 2011, 3:39:24 UTC - in response to Message 41559.  


Power of some signal varies too much.

Should be this spike.

power: 24.001, index(bin)=83950, i guess.

Yes, that one wasn't found on the NVIDIA GPU. Similarly, the NVIDIA found a later one at 24.003 which the ATI didn't. The spike threshold is 24.000 so very small difference in calculations can cause such signals to be reported or not.

Almost certainly the CPU found both of those, hence one more than either of the GPUs.

Although the OpenCL source code is common for the two GPUs, it goes through layers of translation before producing the actual code suitable for those two different GPUs. Result differences are guaranteed, keeping them as small as practical is the goal. Had those spikes been at 25.0 an inconclusive would have been very unlikely since the validator would have considered them equal as long as they didn't differ by more than 0.25.
                                                                Joe
ID: 41568 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 41570 - Posted: 26 Nov 2011, 22:01:50 UTC

Actually, spike finding code differs for HD5/NV and usual ATi builds.
There are few kernels that make use of local memory. There is no local memory for HD4xxx GPUs so usual ATi build doesn't use local memory at all.

News about SETI opt app releases: https://twitter.com/Raistmer
ID: 41570 · Report as offensive

Message boards : SETI@home Enhanced : No consensus


 
©2023 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.