One thread constantly restarting 25.7 build 3

[OneOfMany]

I put this into the 25.7 thread, but I figure it might get lost in there.

I started running Prime95 yesterday and set up the system to use 1024MB of RAM at night. It used this value all day/night and seemed to still be using it until I switched both values back to the 8MB default.

But even before I switched the value (and after) I noticed my second thread was spewing a lot of log. I noticed the message that another thread was using a lot of memory. I don't see an easy way to copy the log. CPU usage was 100% on a dual core the whole time. And at least this morning the logs for worker 2 seemed to show nothing being done, just restarting (I'm guessing) different tasks.

[S485122]

to copy the content of a window with Prime95, just use the "Edit" menu : "Edit / Copy".

Jacob

[OneOfMany]

Thanks...

When I restarted I noticed the same problem...even with mem usage set to default of 8MB (note Prime95 was using 100MB at times with this setting). But when I moved both sizes to 200MB the second thread picked up work, and the 1st got the signal that there was already a thread using lots of memory.

Worker1

[Oct 29 14:00] Optimal bounds are B1=795000, B2=795000
[Oct 29 14:00] Chance of finding a factor is an estimated 3.39%
[Oct 29 14:00] Using FFT length 2560K
[Oct 29 14:00] Ignoring suggested B1 value, using B1=565000 from the save file
[Oct 29 14:00] Exceeded limit on number of workers that can use lots of memory.
[Oct 29 14:00] Looking for work that uses less memory.
[Oct 29 14:00] Optimal P-1 factoring of M46536223 using up to 8MB of memory.
[Oct 29 14:00] Assuming no factors below 2^68 and 2 primality tests saved if a factor is found.
[Oct 29 14:00] Optimal bounds are B1=795000, B2=795000
[Oct 29 14:00] Chance of finding a factor is an estimated 3.39%
[Oct 29 14:00] Using FFT length 2560K
[Oct 29 14:00] Ignoring suggested B1 value, using B1=565000 from the save file
[Oct 29 14:00] Exceeded limit on number of workers that can use lots of memory.
[Oct 29 14:00] Resuming primality test of M46536223 using FFT length 2560K
[Oct 29 14:00] Iteration: 78092 / 46536223 [0.16%].
[Oct 29 14:00] Restarting worker with new memory settings.
[Oct 29 14:00] Stopping primality test of M46536223 at iteration 78198 [0.16%]
[Oct 29 14:00] Optimal P-1 factoring of M46536223 using up to 8MB of memory.
[Oct 29 14:00] Assuming no factors below 2^68 and 2 primality tests saved if a factor is found.
[Oct 29 14:00] Optimal bounds are B1=795000, B2=795000
[Oct 29 14:00] Chance of finding a factor is an estimated 3.39%
[Oct 29 14:00] Using FFT length 2560K
[Oct 29 14:00] Ignoring suggested B1 value, using B1=565000 from the save file
[Oct 29 14:00] Other workers are using lots of memory now.
[Oct 29 14:00] Looking for work that uses less memory.
[Oct 29 14:00] Optimal P-1 factoring of M46536223 using up to 8MB of memory.
[Oct 29 14:00] Assuming no factors below 2^68 and 2 primality tests saved if a factor is found.
[Oct 29 14:00] Optimal bounds are B1=795000, B2=795000
[Oct 29 14:00] Chance of finding a factor is an estimated 3.39%
[Oct 29 14:00] Using FFT length 2560K
[Oct 29 14:00] Ignoring suggested B1 value, using B1=565000 from the save file
[Oct 29 14:00] Other workers are using lots of memory now.
[Oct 29 14:00] Resuming primality test of M46536223 using FFT length 2560K
[Oct 29 14:00] Iteration: 78199 / 46536223 [0.16%].
[Oct 29 14:02] Iteration: 80000 / 46536223 [0.17%]. Per iteration time: 0.070 sec.
[Oct 29 14:14] Iteration: 90000 / 46536223 [0.19%]. Per iteration time: 0.069 sec.
[Oct 29 14:25] Iteration: 100000 / 46536223 [0.21%]. Per iteration time: 0.069 sec.
[Oct 29 14:37] Iteration: 110000 / 46536223 [0.23%]. Per iteration time: 0.068 sec.
[Oct 29 14:48] Iteration: 120000 / 46536223 [0.25%]. Per iteration time: 0.068 sec.
[Oct 29 14:59] Stopping primality test of M46536223 at iteration 129750 [0.27%]
[Oct 29 14:59] Worker stopped.
[Oct 29 15:00] Worker starting
[Oct 29 15:00] Setting affinity to run worker on logical CPU #0
[Oct 29 15:00] Optimal P-1 factoring of M46536223 using up to 8MB of memory.
[Oct 29 15:00] Assuming no factors below 2^68 and 2 primality tests saved if a factor is found.
[Oct 29 15:00] Optimal bounds are B1=795000, B2=795000
[Oct 29 15:00] Chance of finding a factor is an estimated 3.39%
[Oct 29 15:00] Using FFT length 2560K
[Oct 29 15:00] Ignoring suggested B1 value, using B1=565000 from the save file
[Oct 29 15:00] Exceeded limit on number of workers that can use lots of memory.
[Oct 29 15:00] Looking for work that uses less memory.
[Oct 29 15:00] Optimal P-1 factoring of M46536223 using up to 8MB of memory.
[Oct 29 15:00] Assuming no factors below 2^68 and 2 primality tests saved if a factor is found.
[Oct 29 15:00] Optimal bounds are B1=795000, B2=795000
[Oct 29 15:00] Chance of finding a factor is an estimated 3.39%
[Oct 29 15:00] Using FFT length 2560K
[Oct 29 15:00] Ignoring suggested B1 value, using B1=565000 from the save file
[Oct 29 15:00] Exceeded limit on number of workers that can use lots of memory.
[Oct 29 15:00] Resuming primality test of M46536223 using FFT length 2560K
[Oct 29 15:00] Iteration: 129751 / 46536223 [0.27%].
[Oct 29 15:00] Restarting worker with new memory settings.
[Oct 29 15:00] Stopping primality test of M46536223 at iteration 129857 [0.27%]
[Oct 29 15:00] Optimal P-1 factoring of M46536223 using up to 8MB of memory.
[Oct 29 15:00] Assuming no factors below 2^68 and 2 primality tests saved if a factor is found.
[Oct 29 15:00] Optimal bounds are B1=795000, B2=795000
[Oct 29 15:00] Chance of finding a factor is an estimated 3.39%
[Oct 29 15:00] Using FFT length 2560K
[Oct 29 15:00] Ignoring suggested B1 value, using B1=565000 from the save file
[Oct 29 15:00] Other workers are using lots of memory now.
[Oct 29 15:00] Looking for work that uses less memory.
[Oct 29 15:00] Optimal P-1 factoring of M46536223 using up to 8MB of memory.
[Oct 29 15:00] Assuming no factors below 2^68 and 2 primality tests saved if a factor is found.
[Oct 29 15:00] Optimal bounds are B1=795000, B2=795000
[Oct 29 15:00] Chance of finding a factor is an estimated 3.39%
[Oct 29 15:00] Using FFT length 2560K
[Oct 29 15:00] Ignoring suggested B1 value, using B1=565000 from the save file
[Oct 29 15:00] Other workers are using lots of memory now.
[Oct 29 15:00] Resuming primality test of M46536223 using FFT length 2560K
[Oct 29 15:00] Iteration: 129858 / 46536223 [0.27%].
[Oct 29 15:01] Iteration: 130000 / 46536223 [0.27%]. Per iteration time: 0.070 sec.
[Oct 29 15:12] Iteration: 140000 / 46536223 [0.30%]. Per iteration time: 0.068 sec.
[Oct 29 15:19] Restarting worker with new memory settings.
[Oct 29 15:19] Stopping primality test of M46536223 at iteration 145919 [0.31%]
[Oct 29 15:19] Optimal P-1 factoring of M46536223 using up to 200MB of memory.
[Oct 29 15:19] Assuming no factors below 2^68 and 2 primality tests saved if a factor is found.
[Oct 29 15:19] Optimal bounds are B1=525000, B2=7875000
[Oct 29 15:19] Chance of finding a factor is an estimated 5.67%
[Oct 29 15:19] Using FFT length 2560K
[Oct 29 15:19] Exceeded limit on number of workers that can use lots of memory.
[Oct 29 15:19] Looking for work that uses less memory.
[Oct 29 15:19] Optimal P-1 factoring of M46536223 using up to 200MB of memory.
[Oct 29 15:19] Assuming no factors below 2^68 and 2 primality tests saved if a factor is found.
[Oct 29 15:19] Optimal bounds are B1=525000, B2=7875000
[Oct 29 15:19] Chance of finding a factor is an estimated 5.67%
[Oct 29 15:19] Using FFT length 2560K
[Oct 29 15:19] Exceeded limit on number of workers that can use lots of memory.
[Oct 29 15:19] Resuming primality test of M46536223 using FFT length 2560K
[Oct 29 15:19] Iteration: 145920 / 46536223 [0.31%].
[Oct 29 15:25] Iteration: 150000 / 46536223 [0.32%]. Per iteration time: 0.077 sec.

Worker 2

[Oct 29 15:19] Using FFT length 2560K
[Oct 29 15:19] Ignoring suggested B1 value, using B1=565000 from the save file
[Oct 29 15:19] Other workers are using lots of memory now.
[Oct 29 15:19] Looking for work that uses less memory.
[Oct 29 15:19] Optimal P-1 factoring of M46536289 using up to 8MB of memory.
[Oct 29 15:19] Assuming no factors below 2^68 and 2 primality tests saved if a factor is found.
[Oct 29 15:19] Optimal bounds are B1=795000, B2=795000
[Oct 29 15:19] Chance of finding a factor is an estimated 3.39%
[Oct 29 15:19] Using FFT length 2560K
[Oct 29 15:19] Restarting worker with new memory settings.
[Oct 29 15:19] Ignoring suggested B1 value, using B1=565000 from the save file
[Oct 29 15:19] Available memory is 175MB.
[Oct 29 15:19] Using 168MB of memory. Processing 4 relative primes (0 of 480 already processed).
[Oct 29 15:19] Optimal P-1 factoring of M46536289 using up to 200MB of memory.
[Oct 29 15:19] Assuming no factors below 2^68 and 2 primality tests saved if a factor is found.
[Oct 29 15:19] Optimal bounds are B1=525000, B2=7875000
[Oct 29 15:19] Chance of finding a factor is an estimated 5.67%
[Oct 29 15:19] Using FFT length 2560K
[Oct 29 15:20] Available memory is 175MB.
[Oct 29 15:20] Using 168MB of memory. Processing 4 relative primes (0 of 480 already processed).
[Oct 29 15:20] M46536289 stage 2 is 0.54% complete.

[uigrad]

The only reason I have a Windows computer is for playing games. Since games seem to take all the memory available, I've decided to not let prime95 take any factoring assignments. I stick with LL assignments only, and they take very little memory.

The readme.txt says:

Stage 2 of P-1 factoring step prior to running a Lucas-Lehmer test is
slightly more effective if it is given more memory to work with. However,
if you let the program use too much memory then the performance of ALL programs
will suffer. The good news is that 98% of the time the program uses a minimal
amount of memory. In fact, the program will work just fine if never
let it use more than the minimum.

According to the chart in the readme, you will need a minimum of 75MB memory per thread that is doing P-1 factoring, and ideally should have 200MB each for those threads. I don't know how much memory your computer has, but it sounds like you don't have enough to run two factoring threads simultaneously.

If you have any assignments that are LL tests only, you could move them to the top of your worktodo.txt file, and then restart prime95. If not, you might want to unreserve the work for one of your cores, set prime95 to only pick up LL work for that core, and then restart it.

I know I've seen other threads here that explain the memory in more depth, but I can't find them right now.