mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Software

Reply
 
Thread Tools
Old 2011-03-09, 20:04   #155
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

1100111111112 Posts
Default

Is it possible to configure Prime95 to:
  • automatically report results to PrimeNet
  • request assignment IDs for manually generated work
  • not to request new assignments if it runs out of work
  • not automatically unreserve work it thinks it doesn't want to do right now?
James Heinrich is offline   Reply With Quote
Old 2011-03-09, 22:26   #156
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

11100111100112 Posts
Default

Quote:
Originally Posted by James Heinrich View Post
Is it possible to configure Prime95 to:
  • automatically report results to PrimeNet
  • request assignment IDs for manually generated work
  • not to request new assignments if it runs out of work
  • not automatically unreserve work it thinks it doesn't want to do right now?
1) UsePrimenet=1
2) Don't put an assignment ID in the worktodo.txt line
3) NoMoreWork=1
4) Put N/A as the assignment ID on the worktodo.txt line

2 and 4 combined means you must know what work the server will accept and reject. Alternatively, try getting assignment IDs for all work, diff the worktodo file to see what was unreserved, re-insert the unreserved work with an N/A.
Prime95 is online now   Reply With Quote
Old 2011-03-09, 22:32   #157
ixfd64
Bemusing Prompter
 
ixfd64's Avatar
 
"Danny"
Dec 2002
California

94916 Posts
Default

Quote:
automatically report results to PrimeNet
Um, doesn't Prime95 do that already?

Quote:
not to request new assignments if it runs out of work
Advanced > Quit GIMPS...
ixfd64 is offline   Reply With Quote
Old 2011-03-09, 22:35   #158
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

CFF16 Posts
Default

Quote:
Originally Posted by Prime95 View Post
3) NoMoreWork=1
That was the key I was looking for. Thanks. It doesn't seem to be mentioned in either the readme or undoc.
James Heinrich is offline   Reply With Quote
Old 2011-03-10, 04:43   #159
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

7,411 Posts
Default

Quote:
Originally Posted by James Heinrich View Post
That was the key I was looking for. Thanks. It doesn't seem to be mentioned in either the readme or undoc.
That's what Advanced / Quit Gimps writes to the prime.txt file.
Prime95 is online now   Reply With Quote
Old 2011-03-10, 07:53   #160
ATH
Einyen
 
ATH's Avatar
 
Dec 2003
Denmark

17·181 Posts
Default

Quote:
Originally Posted by James Heinrich View Post
Is it possible to configure Prime95 to:
  • automatically report results to PrimeNet
  • request assignment IDs for manually generated work
  • not to request new assignments if it runs out of work
  • not automatically unreserve work it thinks it doesn't want to do right now?
I use MaxExponents=0 in prime.txt for this.
ATH is offline   Reply With Quote
Old 2011-03-10, 23:01   #161
Danath
 
Danath's Avatar
 
May 2003
Republic of Moldova

2816 Posts
Default

I tried this new version on a Celeron 420 (1.6 GHz, 512 KB cache, based on Conroe core), and it looks like it's not choosing the optimal path for all FFT sizes:

Intel(R) Celeron(R) CPU 420 @ 1.60GHz
CPU speed: 1599.97 MHz
CPU features: Prefetch, MMX, SSE, SSE2
L1 cache size: 32 KB
L2 cache size: 512 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
TLBS: 256
Prime95 32-bit version 26.5, RdtscTiming=1
Best time for 768K FFT length: 33.776 ms., avg: 35.123 ms.
Best time for 896K FFT length: 33.539 ms., avg: 33.858 ms.
Best time for 1024K FFT length: 40.043 ms., avg: 40.246 ms.
Best time for 1280K FFT length: 54.130 ms., avg: 54.312 ms.
Best time for 1536K FFT length: 47.279 ms., avg: 47.536 ms.
Best time for 1792K FFT length: 94.949 ms., avg: 95.461 ms.
Best time for 2048K FFT length: 63.452 ms., avg: 63.808 ms.
Best time for 2560K FFT length: 87.990 ms., avg: 88.536 ms.
Best time for 3072K FFT length: 227.910 ms., avg: 228.951 ms.
Best time for 3584K FFT length: 267.396 ms., avg: 268.235 ms.
Best time for 4096K FFT length: 179.772 ms., avg: 180.507 ms.
Best time for 5120K FFT length: 245.666 ms., avg: 246.679 ms.
Best time for 6144K FFT length: 321.693 ms., avg: 322.619 ms.
Best time for 7168K FFT length: 400.121 ms., avg: 401.850 ms.
Best time for 8192K FFT length: 553.498 ms., avg: 555.347 ms.
Best time for 58 bit trial factors: 6.908 ms.
Best time for 59 bit trial factors: 6.896 ms.
Best time for 60 bit trial factors: 6.911 ms.
Best time for 61 bit trial factors: 6.887 ms.
Best time for 62 bit trial factors: 6.893 ms.
Best time for 63 bit trial factors: 11.123 ms.
Best time for 64 bit trial factors: 11.154 ms.
Best time for 65 bit trial factors: 10.792 ms.
Best time for 66 bit trial factors: 10.738 ms.
Best time for 67 bit trial factors: 10.727 ms.

Several benchmarks showed the same results. As you can see, the timings for some FFTs are much worse compared to larger ones (at least, I see the problem at 768K, 1280K, 1792K, 3072K, 3584K). The problem might be also at some other FFT sizes, not reflected in the benchmark.

I used the "Time" option from the menu for a few virtual exponents, and got results consistent with the benchmark:

[Mar 10 20:11] Using Core2 type-3 FFT length 3M, Pass1=3K, Pass2=1K
[Mar 10 20:11] p: 55000000. Time: 233.395 ms.
[Mar 10 20:11] p: 55000000. Time: 232.954 ms.
[Mar 10 20:11] p: 55000000. Time: 233.966 ms.
[Mar 10 20:11] p: 55000000. Time: 231.208 ms.
[Mar 10 20:11] p: 55000000. Time: 229.640 ms.
[Mar 10 20:11] p: 55000000. Time: 230.923 ms.
[Mar 10 20:11] p: 55000000. Time: 230.646 ms.
[Mar 10 20:11] p: 55000000. Time: 236.376 ms.
[Mar 10 20:11] p: 55000000. Time: 232.002 ms.
[Mar 10 20:11] p: 55000000. Time: 237.006 ms.
[Mar 10 20:11] Iterations: 10. Total time: 2.328 sec.
[Mar 10 20:11] Estimated time to complete this exponent: 148 days, 4 hours, 50 minutes.
[Mar 10 20:11] Worker stopped.
[Mar 10 20:12] Worker starting
[Mar 10 20:12] Using Core2 type-3 FFT length 2800K, Pass1=448, Pass2=6400
[Mar 10 20:12] p: 52000000. Time: 103.169 ms.
[Mar 10 20:12] p: 52000000. Time: 104.073 ms.
[Mar 10 20:12] p: 52000000. Time: 103.392 ms.
[Mar 10 20:12] p: 52000000. Time: 103.584 ms.
[Mar 10 20:12] p: 52000000. Time: 102.842 ms.
[Mar 10 20:12] p: 52000000. Time: 103.134 ms.
[Mar 10 20:12] p: 52000000. Time: 104.650 ms.
[Mar 10 20:12] p: 52000000. Time: 102.559 ms.
[Mar 10 20:12] p: 52000000. Time: 102.483 ms.
[Mar 10 20:12] p: 52000000. Time: 103.139 ms.
[Mar 10 20:12] Iterations: 10. Total time: 1.033 sec.
[Mar 10 20:12] Estimated time to complete this exponent: 62 days, 4 hours, 8 minutes.

On the other hand, version 25.11 gives the expected results:

Intel(R) Celeron(R) CPU 420 @ 1.60GHz
CPU speed: 1599.95 MHz
CPU features: RDTSC, CMOV, Prefetch, MMX, SSE, SSE2
L1 cache size: 32 KB
L2 cache size: 512 KB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
TLBS: 256
Prime95 32-bit version 25.11, RdtscTiming=1
Best time for 768K FFT length: 26.684 ms.
Best time for 896K FFT length: 33.084 ms.
Best time for 1024K FFT length: 36.473 ms.
Best time for 1280K FFT length: 45.905 ms.
Best time for 1536K FFT length: 55.025 ms.
Best time for 1792K FFT length: 67.185 ms.
Best time for 2048K FFT length: 75.364 ms.
Best time for 2560K FFT length: 102.764 ms.
Best time for 3072K FFT length: 129.004 ms.
Best time for 3584K FFT length: 157.560 ms.
Best time for 4096K FFT length: 186.269 ms.
Best time for 5120K FFT length: 248.426 ms.
Best time for 6144K FFT length: 310.377 ms.
Best time for 7168K FFT length: 417.948 ms.
Best time for 8192K FFT length: 492.939 ms.
Danath is offline   Reply With Quote
Old 2011-03-19, 19:42   #162
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

9,391 Posts
Default

worktodo.txt needs to be updated when temp p* files are written (e.g. if the interval is 15 minutes).
It seems to be out of date.

I find that when electricity is lost and then restored, the computer boots and starts with the numbers that are already processed (because they are still in the worktodo.txt); what's worse is that then they go for the whole 100% workunit (naturally, their savefiles are already deleted).

This is observed on the type of work (each test takes a few hours; not your usual LL tests):
Code:
[Worker #1]
PRP=4,257,323502,1
PRP=4,257,323838,1
[Worker #2]
PRP=4,257,322350,1
PRP=4,257,322742,1
[Worker #3]
PRP=4,257,...etc
[Worker #4]
PRP=4,257,...
Batalov is offline   Reply With Quote
Old 2011-03-21, 00:32   #163
Falkentyne
 
Mar 2011

168 Posts
Default

Hi,
Prime95 26.5 32 bit won't start on my system. Nothing happens when I start it, no window, nothing. Process isn't even showing in memory. I saw a few other people with the same problem. The previous 26.4 works perfectly, however.

Any ideas what the issue is? I haven't tried the 64 bit yet. Using windows xp with an i7-2600k, hyperthreading off.
Falkentyne is offline   Reply With Quote
Old 2011-03-23, 05:35   #164
clarke
 
Feb 2009

22×7 Posts
Default

Prime95 26.5 on Celeron D340 2.9GHz (256k L2) throws error after passing 960k blend self-test:
[Work thread Mar 22 20:42] Self-test 960K passed!
[Work thread Mar 22 20:42] Cannot initialize FFT code, errcode=1002
[Work thread Mar 22 20:42] Number sent to gwsetup is too large for the FFTs to handle.
[Work thread Mar 22 20:42] Torture Test completed 38 tests in 2 hours, 31 minutes - 0 errors, 0 warnings.
Computer isn't oveclocked and from here http://mersenneforum.org/showthread....200#post237200 and http://mersenneforum.org/showthread....998#post241998 it seems you've addressed this issue already. Or it is not fixed?
Attached Files
File Type: txt test.txt (31.2 KB, 224 views)
clarke is offline   Reply With Quote
Old 2011-03-23, 16:22   #165
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

7,411 Posts
Default

Quote:
Originally Posted by Falkentyne View Post
Prime95 26.5 32 bit won't start on my system.

Any ideas what the issue is? I haven't tried the 64 bit yet. Using windows xp with an i7-2600k, hyperthreading off.
Try putting "NumCPUs=4" and "CpuNumHyperthreads=1" in local.txt.

Last fiddled with by Prime95 on 2011-03-23 at 19:13
Prime95 is online now   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Prime95 version 27.3 Prime95 Software 148 2012-03-18 19:24
Prime95 version 26.3 Prime95 Software 76 2010-12-11 00:11
Prime95 version 25.5 Prime95 PrimeNet 369 2008-02-26 05:21
Prime95 version 25.4 Prime95 PrimeNet 143 2007-09-24 21:01
When the next prime95 version ? pacionet Software 74 2006-12-07 20:30

All times are UTC. The time now is 16:31.

Mon Apr 19 16:31:54 UTC 2021 up 11 days, 11:12, 1 user, load averages: 2.87, 2.87, 2.44

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.