Join PrimeGrid
Returning Participants
Community
Leader Boards
Results
Other
drummers-lowrise
|
Message boards :
Project Staging Area :
"Small " GenferCuda 32768 with new application
| Author |
Message |
|---|
 Crun-chi Volunteer tester
Send message
Joined: 25 Nov 09
Posts: 3233
ID: 50683
Credit: 151,443,349
RAC: 73,965
                         
|
|
I look at some results with new 1.02 (cuda32_13) application under Boinc for Genefer CUDA , and if I sow correctly it is much faster then PRPNET Genfercuda.
And one more thing: it looks like BOINC PRIMEGRID 1.02 (cuda32_13) can do small GENEFER CUDA number (32768).
When I try to do with PRPNET project , I was always failed and first successfully was GFN262144
It seems I found
http://www.primegrid.com/forum_thread.php?id=3920#46425
And since I have "problematic" 560Ti... all is ok:)
____________
92*10^1585996-1 NEAR-REPDIGIT PRIME :) :) :)
4 * 650^498101-1 CRUS PRIME
2022202116^131072+1 GENERALIZED FERMAT
Proud member of team Aggie The Pew. Go Aggie! | |
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
                 
|
I look at some results with new 1.02 (cuda32_13) application under Boinc for Genefer CUDA , and if I sow correctly it is much faster then PRPNET Genfercuda.
And one more thing: it looks like BOINC PRIMEGRID 1.02 (cuda32_13) can do small GENEFER CUDA number (32768).
When I try to do with PRPNET project , I was always failed and first successfully was GFN262144
It seems I found
http://www.primegrid.com/forum_thread.php?id=3920#46425
And since I have "problematic" 560Ti... all is ok:)
In BOINC, we did testing of small b GFN's which was within MaxErr limit of geneferCUDA. However, in PRPNet, we are well beyond that limit.
In both BOINC and PRPNet, the lowest N we can test is N=262144. The current test range of "live" data is from this N.
____________
| |
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14011
ID: 53948
Credit: 433,244,369
RAC: 880,036
                          
|
I look at some results with new 1.02 (cuda32_13) application under Boinc for Genefer CUDA , and if I sow correctly it is much faster then PRPNET Genfercuda.
And one more thing: it looks like BOINC PRIMEGRID 1.02 (cuda32_13) can do small GENEFER CUDA number (32768).
When I try to do with PRPNET project , I was always failed and first successfully was GFN262144
As far as I know, the version of GeneferCUDA being distributed on PRPNet right now (0.97) should be exactly the same speed as the boinc version 1.01 or 1.02.
If you run the benchmark tests on the two programs, 0.97 will *appear* faster, but that's just a discrepency in the way the benchmarks work in the two versions. They use essentially identical code for the calculations, so they run at the same speed.
The big difference between the two -- and why I use boinc/1.01 on prpnet instead of 0.99, is because the checkpointing is more robust.
There are some speed improvements I'm currently working on, but they only apply to the initialization of the program, so they're really only a problem at higher N values that we're not searching yet.
And one more thing: it looks like BOINC PRIMEGRID 1.02 (cuda32_13) can do small GENEFER CUDA number (32768).
When I try to do with PRPNET project , I was always failed and first successfully was GFN262144
GeneferCUDA can crunch at N=32768, but only for values of B below approximately 1.8 million. The current search range on PRPNet for N=32768 has a B higher than that, so the search range is beyond what GeneferCUDA is capable of. The same applies to N=65536, where the upper limit of B is about 1.5 million.
____________
My lucky number is 75898524288+1 | |
|
Crun-chiVolunteer tester
Send message
Joined: 25 Nov 09
Posts: 3233
ID: 50683
Credit: 151,443,349
RAC: 73,965
|
|
Thank you both, for explanation!
____________
92*10^1585996-1 NEAR-REPDIGIT PRIME :) :) :)
4 * 650^498101-1 CRUS PRIME
2022202116^131072+1 GENERALIZED FERMAT
Proud member of team Aggie The Pew. Go Aggie! | |
|
Message boards :
Project Staging Area :
"Small " GenferCuda 32768 with new application |
