Join PrimeGrid
Returning Participants
Community
Leader Boards
Results
Other
drummers-lowrise
|
1)
Message boards :
Problems and Help :
New CUDA task error ???
(Message 25869)
Posted 4671 days ago by Rüdiger
Hi,
1.28 works for me as well,
Rüdiger
|
2)
Message boards :
Problems and Help :
New CUDA task error ???
(Message 25750)
Posted 4675 days ago by Rüdiger
boinccmd --version
boinccmd, built from BOINC 6.10.17
I didn't try to upgrade my version, as 6.10.17 comes out of the box with kubuntu. Shall I try?
Rüdiger
|
3)
Message boards :
Problems and Help :
New CUDA task error ???
(Message 25748)
Posted 4675 days ago by Rüdiger
return value is 0
|
4)
Message boards :
Problems and Help :
New CUDA task error ???
(Message 25741)
Posted 4675 days ago by Rüdiger
Hi Ken,
I guess you answered to Joe still, however I ran the test as well, though I'm not allowed to access http://sites.google.com/site/kenscode/prime-programs/ppsieve-cuda.zip?attredirects=0&d=1; I used the downloaded version instead:
root@rio:/var/lib/boinc-client/projects/www.primegrid.com# ./primegrid_ppsieve_1.27_x86_64-pc-linux-gnu__cuda23 -p42070e9 -P42070003e6 -k 1201 -K 9999 -N 2000000 -z normal
ppsieve version cuda-0.1.5 (testing)
Compiled Aug 10 2010 with GCC 4.3.3
nstart=76, nstep=32, gpu_nstep=32
ppsieve initialized: 1201 <= k <= 9999, 76 <= n <= 2000000
Sieve started: 42070000000000 <= p < 42070003000000
Thread 0 starting
Detected GPU 0: GeForce 9800 GT
Detected compute capability: 1.1
Detected 14 multiprocessors.
42070000070587 | 9475*2^197534+1
42070000198537 | 3373*2^1046686+1
Thread 0 completed
Waiting for threads to exit
Sieve complete: 42070000000000 <= p < 42070003000000
Found 2 factors
count=95668,sum=0x37dacb7121ccffe4
Elapsed time: 6.46 sec. (0.01 init + 6.44 sieve) at 488406 p/sec.
Processor time: 0.13 sec. (0.02 init + 0.11 sieve) at 28597527 p/sec.
Average processor utilization: 1.36 (init), 0.02 (sieve)
When I run the same test with 1.24, I get 6 factors:
root@rio:/var/lib/boinc-client/projects/www.primegrid.com# /tmp/primegrid_ppsieve_1.24_x86_64-pc-linux-gnu__cuda23 -p42070e9 -P42070003e6 -k 1201 -K 9999 -N 2000000 -z normal
ppsieve version cuda-0.1.3c (testing)
Compiled Jul 28 2010 with GCC 4.3.3
nstart=76, nstep=32, gpu_nstep=32
ppsieve initialized: 1201 <= k <= 9999, 76 <= n <= 2000000
42070000070587 | 9475*2^197534+1
42070000198537 | 3373*2^1046686+1
42070000300049 | 9139*2^461846+1
42070000464001 | 4179*2^1577462+1
42070001011573 | 7113*2^215532+1
42070002690167 | 2553*2^1888870+1
Found 6 factors
My stderr.txt looks like:
root@rio:/var/lib/boinc-client/projects/www.primegrid.com# cat stderr.txt
Can't open init data file - running in standalone mode
Sieve started: 42070000000000 <= p < 42070003000000
Thread 0 starting
Detected GPU 0: GeForce 9800 GT
Detected compute capability: 1.1
Detected 14 multiprocessors.
Thread 0 completed
Sieve complete: 42070000000000 <= p < 42070003000000
count=95668,sum=0x37dacb7121ccffe4
Elapsed time: 6.46 sec. (0.01 init + 6.45 sieve) at 488048 p/sec.
Processor time: 0.14 sec. (0.02 init + 0.12 sieve) at 26214400 p/sec.
Average processor utilization: 1.36 (init), 0.02 (sieve)
called boinc_finish
The stderr.txt of the boinc client run is:
cat /var/lib/boinc-client/slots/0/stderr.txt
Sieve started: 457947000000000 <= p < 457948000000000
Thread 0 starting
Detected GPU 0: GeForce 9800 GT
Detected compute capability: 1.1
Detected 14 multiprocessors.
Thread 0 completed
Waiting for threads to exit
Sieve complete: 457947000000000 <= p < 457948000000000
count=29621876,sum=0x5fdb2dd654ff2536
Elapsed time: 1811.52 sec. (0.05 init + 1811.46 sieve) at 552085 p/sec.
Processor time: 9.31 sec. (0.05 init + 9.26 sieve) at 107999931 p/sec.
Average processor utilization: 0.92 (init), 0.01 (sieve)
Sieve started: 457947000000000 <= p < 457948000000000
Thread 0 starting
Detected GPU 0: GeForce 9800 GT
Detected compute capability: 1.1
Detected 14 multiprocessors.
Thread 0 completed
Waiting for threads to exit
Sieve complete: 457947000000000 <= p < 457948000000000
count=29621876,sum=0x5fdb2dd654ff2536
Elapsed time: 1806.76 sec. (0.06 init + 1806.70 sieve) at 553539 p/sec.
Processor time: 9.24 sec. (0.05 init + 9.19 sieve) at 108822564 p/sec.
Average processor utilization: 0.91 (init), 0.01 (sieve)
Sieve started: 457947000000000 <= p < 457948000000000
Thread 0 starting
Detected GPU 0: GeForce 9800 GT
Detected compute capability: 1.1
Detected 14 multiprocessors.
|
5)
Message boards :
Problems and Help :
New CUDA task error ???
(Message 25727)
Posted 4676 days ago by Rüdiger
Hi Ken,
I seem to have a similar problem, originally reported in the thread mentioned by you: http://www.primegrid.com/forum_thread.php?id=2638 in message http://www.primegrid.com/forum_thread.php?id=2638&nowrap=true#25700
With 1.24, my 9800GT crunched a block in about 30 minutes, 1.27 now states that it should finish in 00:01:28 minutes, but the progress just stays at 0%, until it gets killed due to timeout.
Do you need more input to figure out what is wrong? Can I somehow enforce a switchback to 1.24?
Thanks,
Rüdiger
|
6)
Message boards :
Proth Prime Search :
PPS sieve (cuda23) memory error
(Message 25725)
Posted 4676 days ago by Rüdiger
Hi,
as Scott seems to have solved his problem, shall I create a new thread for 'my' problem? His solution is not suitable for me, as I am using the latest driver (258.96) already.
Rüdiger
|
7)
Message boards :
Proth Prime Search :
PPS sieve (cuda23) memory error
(Message 25700)
Posted 4676 days ago by Rüdiger
Hi,
pps 1.27 for linux seems to be broken: I encounter a similar error as Scott since my computer upgraded from "Proth Prime Search (Sieve) v1.24 (cuda23)" to "Proth Prime Search (Sieve) v1.27 (cuda23)"
I'm running kubuntu x64, using a Nvidia 9800 GT
20-Aug-2010 01:43:22 [PrimeGrid] Starting pps_sr2sieve_1603578_1
20-Aug-2010 01:43:22 [PrimeGrid] Starting task pps_sr2sieve_1603578_1 using pps_sr2sieve version 127
20-Aug-2010 02:13:41 [PrimeGrid] Task pps_sr2sieve_1603578_1 exited with zero status but no 'finished' file
20-Aug-2010 02:13:41 [PrimeGrid] If this happens repeatedly you may need to reset the project.
20-Aug-2010 02:13:41 [PrimeGrid] Restarting task pps_sr2sieve_1603578_1 using pps_sr2sieve version 127
20-Aug-2010 02:44:01 [PrimeGrid] Task pps_sr2sieve_1603578_1 exited with zero status but no 'finished' file
20-Aug-2010 02:44:01 [PrimeGrid] If this happens repeatedly you may need to reset the project.
20-Aug-2010 02:44:01 [PrimeGrid] Restarting task pps_sr2sieve_1603578_1 using pps_sr2sieve version 127
20-Aug-2010 03:14:21 [PrimeGrid] Task pps_sr2sieve_1603578_1 exited with zero status but no 'finished' file
20-Aug-2010 03:14:21 [PrimeGrid] If this happens repeatedly you may need to reset the project.
20-Aug-2010 03:14:21 [PrimeGrid] Restarting task pps_sr2sieve_1603578_1 using pps_sr2sieve version 127
20-Aug-2010 03:35:31 [PrimeGrid] Aborting task pps_sr2sieve_1603578_1: exceeded elapsed time limit 6724.601457
20-Aug-2010 03:36:32 [PrimeGrid] Computation for task pps_sr2sieve_1603578_1 finished
20-Aug-2010 03:36:32 [PrimeGrid] Output file pps_sr2sieve_1603578_1_0 for task pps_sr2sieve_1603578_1 absent
Sometimes I also see memory problems:
20-Aug-2010 10:44:31 [PrimeGrid] Task pps_sr2sieve_1612901_0: no shared memory segment
20-Aug-2010 10:44:31 [PrimeGrid] Task pps_sr2sieve_1612901_0 exited with zero status but no 'finished' file
All computations fail since this switchover; they get killed after an hour. Also others seem to see the same problem, as the same work unit gets handled by several people unsuccessfully with v1.27, see e.g. http://www.primegrid.com/workunit.php?wuid=128271014
Can you revert to v1.24?
Thanks,
Rüdiger
|
|
