mersenneforum.org  

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

Reply
 
Thread Tools
Old 2019-02-18, 03:56   #1
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

3×7×11×31 Posts
Default Prime95 version 29.6/29.7/29.8

Prime95 version 29.8 build 6 is available.

From whatsnew.txt:

Code:
1)  Support added for AVX-512 FFTs.
2)  FMA3 FFTs now have slightly higher FFT crossover points.  Soft crossovers are
    no longer used by default.  See undoc.txt.
3)  Torture test dialog box options now based on cache sizes.  Options for performing
    a weaker torture test are available.  Torture tests that use all RAM are now more
    stressful.  In-place vs. not in-place memory accesses now displayed on screen.
    On machines with more than 4GB of memory, blend defaults to 1/16th of RAM.
4)  Add & subtract operations for AVX-512 FFTs are now multithreaded.  This should
    improve performance for P-1 and ECM when using multiple threads.
5)  Benchmark results are now written to results.bench.txt.
6)  JSON results are now available for all work performed.  JSON results are
    written to results.json.txt.
7)  PRP tests with Gerbicz error checking are more immune to hardware errors.
This is a release candidate version. If no serious bugs are reported this will become the official release on the mersenne.org download page.

There are no known serious bugs. There should be no problem with replacing your existing prime95/mprime executable with the new one (the Gerbicz PRP changes will in rare situations cause a save file to be discarded and a previous save file used).

Thanks to all that helped reporting bugs during the AVX-512 development in the 29.5 thread.


Download links:
Windows 64-bit: https://mersenne.org/ftp_root/gimps/p95v298b6.win64.zip
Linux 64-bit: https://mersenne.org/ftp_root/gimps/...linux64.tar.gz
Mac OS X: https://mersenne.org/ftp_root/gimps/....MacOSX.tar.gz
Mac OS X no GUI: https://mersenne.org/ftp_root/gimps/...X.noGUI.tar.gz
Windows 32-bit: https://mersenne.org/ftp_root/gimps/p95v298b6.win32.zip
Linux 32-bit: https://mersenne.org/ftp_root/gimps/...linux32.tar.gz
FreeBSD11 64-bit: https://mersenne.org/ftp_root/gimps/...SD11-64.tar.gz
Source: https://mersenne.org/ftp_root/gimps/...8b6.source.zip
Windows 64-bit service: https://mersenne.org/ftp_root/gimps/...64.service.zip
Windows 32-bit service: https://mersenne.org/ftp_root/gimps/...32.service.zip

Last fiddled with by Prime95 on 2020-03-19 at 23:10
Prime95 is offline   Reply With Quote
Old 2019-02-18, 03:57   #2
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

157718 Posts
Default

Placeholder for bugs reported and bugs fixed.

1) Gerbicz PRP tests of (2^N+1)/factors fails in the last few iterations. Fixed in build 2.
2) Bulldozer users got a "no available FFT lengths" error trying to torture test with FMA3 or AVX FFTs. Fixed in build 2. Note that for the first time ever Bulldozer users can torture test using FMA3 and AVX FFTs. However, SSE2 FFTs may be more stressful.
3) x87 FFTs broken. Fixed in future build 2.
4) In weak torture test, Windows and Mac users are allowed to select options that cause "no FFT lengths" found errors. Fixed in build 6.
5) ScaleOutputFrequency=1 did not work testing ting numbers on AVX-512 machines. Fixed in build 3.
6) Older architectures, such as Pentium-4, may not select the best FFT implementation from gwnum.txt benchmark data. Fixed in build 3.
7) Older architectures, such as Pentium M or 4, do not display L1/L2 cache sizes. Fixed in build 3.
8) Torture test dialog box incorrectly calculated which FFTs will fit in the L3 or L4 cache. Fixed in build 7.
9) AVX-512 FFT fails on 18347731*109^1536-1 (a "rational" FFT -- no FFT weights). Fixed in 29.7.
10) FMA3 FFTs for exponents from 595.7M to 922.6M failed. Fixed in 29.7.
11) Some SkylakeX CPUs would say "no FFT sizes available" for default small torture test. Upper bound on FFT size changed to use L2+L3 cache size since the L3 cache is not inclusive. Fixed in 29.7.
12) The ability to stop individual torture test threads was not working. Running a benchmark without first stopping the torture test could cause a hang or crash. Fixed in 29.7.
13) Windows only!!! Zero-padded AVX-512 FFTs are not working. Fixed in 29.8.
14) If an error occurs writing worktodo.txt, then prime95 will hang at some later point in time. Fixed in 29.8 build 2.
15) If an error occurs deleting a worktodo.txt entry, the worker stopped computing. In 29.8 build 3, the worker will ignore the error and move onto the next entry in worktodo.txt.
16) Incorrect default value for torture test type in Linux menuing system. Fixed in 29.8 build 4.
17) URL of Mersenne Wiki was wrong. Fixed in 29.8 build 4.
18) Mojave dark mode not supported properly. Relinked with latest xcode. Fixed in 29.8 build 4, but untested for pre-Mojave OS.
19) Setting NumCPUs in local.txt to less than the number of L2 caches can cause the Torture Test dialog box to crash. Fixed in 29.8 build 4.
20) Benchmark states results are written to results.txt. Changed message to say results.bench.txt. Fixed in 29.8 build 5.
21) For PRP work, TF depth was not being written to worktodo.txt unless P-1 was required. This caused Test/Status to underestimate the chance that the PRP test would find a new Mersenne prime. Fixed in 29.8 build 5.
22) P-1 is frequently missing factors of numbers of the form 2*3^n+1. Fixed in 29.8 build 6.
23) In a throughput benchmark on machines with multiple L3 caches, some combinations of number-of-cores / number-of-workers would raise errors setting affinity or crash. Fixed in 29.8 build 6.

Last fiddled with by Prime95 on 2019-08-18 at 18:35
Prime95 is offline   Reply With Quote
Old 2019-02-18, 17:26   #3
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

444410 Posts
Default

[QUOTE=Prime95;508841]Prime95 version 29.6 build 1 is available.

From whatsnew.txt:

Code:
1)  Support added for AVX-512 FFTs.
2)  FMA3 FFTs now have slightly higher FFT crossover points.  Soft crossovers are
    no longer used by default.  See undoc.txt.
3)  Mprime now creates a pid file.
4)  Torture test dialog box options now based on cache sizes.  Options for performing
    a weaker torture test are available.  Torture tests that use all RAM are now more
    stressful.  In-place vs. not in-place memory accesses now displayed on screen.
5)  Add & subtract operations for AVX-512 FFTs are now multithreaded.  This should
    improve performance for P-1 and ECM when using multiple threads.
6)  Benchmark results are now written to results.bench.txt.
7)  JSON results are now available for all work performed.  JSON results are
    written to results.json.txt.
8)  Default memory available for prime95 changed from 8MB to 1/16th of RAM.
9)  PRP tests with Gerbicz error checking are more immune to hardware errors.
Note #3 does not appear in the Windows version, so numbering of the rest of the items differ by one. Maybe put OS-specific items at the end of the whatsnew lists. If they were identified separately you might get away with a single version of the file for multiple OSes.

Last fiddled with by kriesel on 2019-02-18 at 17:27
kriesel is offline   Reply With Quote
Old 2019-02-18, 18:51   #4
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

10001010111002 Posts
Default skipped interim residues; primality test type

Welcome changes and a few small things that might be cleaned up.
Following brief test was run on an HP G72 notebook (i3-M370, Win7 x64)

0) no benchmarking tested

1) The skipped interim residue case first reported for V29.4 remains in V29.6b1. See https://www.mersenneforum.org/showpo...&postcount=427 Note iteration 11 and 17 below have no interim residue output.
In prime.txt,
Code:
OutputIterations=1
InterimResidues=1
Worker window contains
Code:
[Feb 18 12:01] Worker starting
[Feb 18 12:01] Setting affinity to run worker on CPU core #1
[Feb 18 12:01] Setting affinity to run helper thread 1 on CPU core #2
[Feb 18 12:01] Starting primality test of M82589933 using FFT length 4480K, Pass1=896, Pass2=5K, clm=4, 2 threads
[Feb 18 12:01] Iteration: 3 / 82589933 [0.00%], ms/iter: 59.833, ETA: 57d 04:40
[Feb 18 12:01] M82589933 interim LL residue 000000000000000E at iteration 3
[Feb 18 12:01] Iteration: 4 / 82589933 [0.00%], ms/iter: 62.653, ETA: 59d 21:22
[Feb 18 12:01] M82589933 interim LL residue 00000000000000C2 at iteration 4
[Feb 18 12:01] Iteration: 5 / 82589933 [0.00%], ms/iter: 85.160, ETA: 81d 09:43
[Feb 18 12:01] M82589933 interim LL residue 0000000000009302 at iteration 5
[Feb 18 12:01] Iteration: 6 / 82589933 [0.00%], ms/iter: 63.056, ETA: 60d 06:36
[Feb 18 12:01] M82589933 interim LL residue 00000000546B4C02 at iteration 6
[Feb 18 12:01] Iteration: 7 / 82589933 [0.00%], ms/iter: 67.032, ETA: 64d 01:49
[Feb 18 12:01] M82589933 interim LL residue 1BD696D9F03D3002 at iteration 7
[Feb 18 12:01] Iteration: 8 / 82589933 [0.00%], ms/iter: 62.132, ETA: 59d 09:24
[Feb 18 12:01] M82589933 interim LL residue 8CC88407A9F4C002 at iteration 8
[Feb 18 12:01] Iteration: 9 / 82589933 [0.00%], ms/iter: 63.589, ETA: 60d 18:50
[Feb 18 12:01] M82589933 interim LL residue 55599F9D37D30002 at iteration 9
[Feb 18 12:01] Iteration: 10 / 82589933 [0.00%], ms/iter: 63.583, ETA: 60d 18:42
[Feb 18 12:01] M82589933 interim LL residue F460D65DDF4C0002 at iteration 10
[Feb 18 12:01] Iteration: 11 / 82589933 [0.00%], ms/iter: 65.185, ETA: 62d 07:26
[Feb 18 12:01] Stopping primality test of M82589933 at iteration 11 [0.00%]
[Feb 18 12:01] Worker stopped.
[Feb 18 12:01] Worker starting
[Feb 18 12:01] Setting affinity to run worker on CPU core #1
[Feb 18 12:01] Setting affinity to run helper thread 1 on CPU core #2
[Feb 18 12:01] Running Jacobi error check.  Passed.  Time:  1.159 sec.
[Feb 18 12:01] Resuming primality test of M82589933 using FFT length 4480K, Pass1=896, Pass2=5K, clm=4, 2 threads
[Feb 18 12:01] Iteration: 12 / 82589933 [0.00%].
[Feb 18 12:01] M82589933 interim LL residue 9BDB491DF4C00002 at iteration 12
[Feb 18 12:01] Iteration: 13 / 82589933 [0.00%], ms/iter: 31.265, ETA: 29d 21:15
[Feb 18 12:01] M82589933 interim LL residue 4CEBB477D3000002 at iteration 13
[Feb 18 12:01] Iteration: 14 / 82589933 [0.00%], ms/iter: 60.381, ETA: 57d 17:14
[Feb 18 12:01] M82589933 interim LL residue 0B97D1DF4C000002 at iteration 14
[Feb 18 12:01] Iteration: 15 / 82589933 [0.00%], ms/iter: 91.954, ETA: 87d 21:34
[Feb 18 12:01] M82589933 interim LL residue ACEF477D30000002 at iteration 15
[Feb 18 12:01] Iteration: 16 / 82589933 [0.00%], ms/iter: 61.444, ETA: 58d 17:37
[Feb 18 12:01] M82589933 interim LL residue 9CBD1DF4C0000002 at iteration 16
[Feb 18 12:01] Iteration: 17 / 82589933 [0.00%], ms/iter: 65.462, ETA: 62d 13:48
[Feb 18 12:01] Stopping primality test of M82589933 at iteration 17 [0.00%]
[Feb 18 12:01] Worker stopped.
[Feb 18 12:02] Worker starting
[Feb 18 12:02] Setting affinity to run worker on CPU core #1
[Feb 18 12:02] Setting affinity to run helper thread 1 on CPU core #2
[Feb 18 12:02] Running Jacobi error check.  Passed.  Time:  1.752 sec.
[Feb 18 12:02] Resuming primality test of M82589933 using FFT length 4480K, Pass1=896, Pass2=5K, clm=4, 2 threads
[Feb 18 12:02] Iteration: 18 / 82589933 [0.00%].
[Feb 18 12:02] M82589933 interim LL residue 0BD1DF4C00000002 at iteration 18
[Feb 18 12:02] Iteration: 19 / 82589933 [0.00%], ms/iter: 34.795, ETA: 33d 06:14
[Feb 18 12:02] M82589933 interim LL residue 2F477D3000000002 at iteration 19
[Feb 18 12:02] Iteration: 20 / 82589933 [0.00%], ms/iter: 68.893, ETA: 65d 20:30
[Feb 18 12:02] M82589933 interim LL residue BD1DF4C000000002 at iteration 20
[Feb 18 12:02] Iteration: 21 / 82589933 [0.00%], ms/iter: 64.226, ETA: 61d 09:27
[Feb 18 12:02] M82589933 interim LL residue F477D30000000002 at iteration 21
[Feb 18 12:02] Iteration: 22 / 82589933 [0.00%], ms/iter: 65.461, ETA: 62d 13:47
[Feb 18 12:02] M82589933 interim LL residue D1DF4C0000000002 at iteration 22
[Feb 18 12:02] Iteration: 23 / 82589933 [0.00%], ms/iter: 66.030, ETA: 63d 02:50
[Feb 18 12:02] Stopping primality test of M82589933 at iteration 23 [0.00%]
[Feb 18 12:02] Worker stopped.
2) This run indicates LL in the worker title bar, a nice addition to distinguish it from PRP3. Similarly, PRP under way indicates x.xx% of PRP (exponent) in the worker title bar. Interim residue lines contained in the worker window also indicate LL or PRP: format approximately
Mexponent interim LL residue DEADBEEF12340002 at iteration n

3) Status counts both LL and PRP3 assignment lines in worktodo.txt toward total exponents to be tested, and presumably toward odds of finding a prime.

4) This worktodo.txt content briefly showed PRP in the worker title bar and then went back to computing LL iterations, despite the worktodo order.
Code:
PRP=N/A,1,2,82589933,-1
Test=N/A,82589933,76,1
Are the save files named the same for LL and PRP3? Worker window content:
Code:
[Feb 18 12:10] Worker starting
[Feb 18 12:10] Setting affinity to run worker on CPU core #1
[Feb 18 12:10] Setting affinity to run helper thread 1 on CPU core #2
[Feb 18 12:10] Error reading intermediate file: p82589933
[Feb 18 12:10] Renaming p82589933 to p82589933.bad1
[Feb 18 12:10] Trying backup intermediate file: p82589933.bu
[Feb 18 12:10] Error reading intermediate file: p82589933.bu
[Feb 18 12:10] Renaming p82589933.bu to p82589933.bad2
[Feb 18 12:10] Trying backup intermediate file: p82589933.bu2
[Feb 18 12:10] Error reading intermediate file: p82589933.bu2
[Feb 18 12:10] Renaming p82589933.bu2 to p82589933.bad3
[Feb 18 12:10] All intermediate files bad.  Temporarily abandoning work unit.
[Feb 18 12:10] Setting affinity to run helper thread 1 on CPU core #2
[Feb 18 12:10] Trying backup intermediate file: p82589933.bad3
[Feb 18 12:10] Running Jacobi error check.  Passed.  Time:  1.200 sec.
[Feb 18 12:10] Resuming primality test of M82589933 using FFT length 4480K, Pass1=896, Pass2=5K, clm=4, 2 threads
[Feb 18 12:10] Iteration: 12 / 82589933 [0.00%].
[Feb 18 12:10] M82589933 interim LL residue 9BDB491DF4C00002 at iteration 12
[Feb 18 12:10] Iteration: 13 / 82589933 [0.00%], ms/iter: 30.680, ETA: 29d 07:50
[Feb 18 12:10] M82589933 interim LL residue 4CEBB477D3000002 at iteration 13
[Feb 18 12:10] Iteration: 14 / 82589933 [0.00%], ms/iter: 63.538, ETA: 60d 17:39
[Feb 18 12:10] M82589933 interim LL residue 0B97D1DF4C000002 at iteration 14
[Feb 18 12:10] Iteration: 15 / 82589933 [0.00%], ms/iter: 61.414, ETA: 58d 16:56
[Feb 18 12:10] M82589933 interim LL residue ACEF477D30000002 at iteration 15
[Feb 18 12:10] Iteration: 16 / 82589933 [0.00%], ms/iter: 66.542, ETA: 63d 14:35
[Feb 18 12:10] M82589933 interim LL residue 9CBD1DF4C0000002 at iteration 16
[Feb 18 12:10] Iteration: 17 / 82589933 [0.00%], ms/iter: 71.187, ETA: 68d 01:08
[Feb 18 12:10] M82589933 interim LL residue 02F477D300000002 at iteration 17
[Feb 18 12:10] Iteration: 18 / 82589933 [0.00%], ms/iter: 62.478, ETA: 59d 17:20
[Feb 18 12:10] M82589933 interim LL residue 0BD1DF4C00000002 at iteration 18
[Feb 18 12:10] Iteration: 19 / 82589933 [0.00%], ms/iter: 60.849, ETA: 58d 03:59
[Feb 18 12:10] M82589933 interim LL residue 2F477D3000000002 at iteration 19
[Feb 18 12:10] Iteration: 20 / 82589933 [0.00%], ms/iter: 64.582, ETA: 61d 17:37
[Feb 18 12:10] M82589933 interim LL residue BD1DF4C000000002 at iteration 20
[Feb 18 12:10] Iteration: 21 / 82589933 [0.00%], ms/iter: 62.924, ETA: 60d 03:34
[Feb 18 12:10] M82589933 interim LL residue F477D30000000002 at iteration 21
[Feb 18 12:10] Iteration: 22 / 82589933 [0.00%], ms/iter: 63.104, ETA: 60d 07:43
[Feb 18 12:10] M82589933 interim LL residue D1DF4C0000000002 at iteration 22
[Feb 18 12:10] Iteration: 23 / 82589933 [0.00%], ms/iter: 82.314, ETA: 78d 16:24
[Feb 18 12:10] M82589933 interim LL residue 477D300000000002 at iteration 23
[Feb 18 12:10] Iteration: 24 / 82589933 [0.00%], ms/iter: 67.763, ETA: 64d 18:35
[Feb 18 12:10] M82589933 interim LL residue 1DF4C00000000002 at iteration 24
[Feb 18 12:10] Iteration: 25 / 82589933 [0.00%], ms/iter: 60.561, ETA: 57d 21:22
[Feb 18 12:10] M82589933 interim LL residue 77D3000000000002 at iteration 25
[Feb 18 12:10] Iteration: 26 / 82589933 [0.00%], ms/iter: 65.237, ETA: 62d 08:38
[Feb 18 12:10] M82589933 interim LL residue DF4C000000000002 at iteration 26
[Feb 18 12:10] Iteration: 27 / 82589933 [0.00%], ms/iter: 63.078, ETA: 60d 07:06
[Feb 18 12:10] Stopping primality test of M82589933 at iteration 27 [0.00%]
[Feb 18 12:10] Worker stopped.
There were a few intermediate save files labeled bad, in the working folder. Presumably because they were from LL and the first worktodo item was PRP, for the same exponent. This case should be very rare. It may occur by accident or in some testing scenarios.

The worker was stopped and all intermediate files for the exponent were moved to a separate folder. On resume, PRP3 started from iteration 1.
Code:
[Feb 18 12:19] Worker starting
[Feb 18 12:19] Setting affinity to run worker on CPU core #1
[Feb 18 12:19] Setting affinity to run helper thread 1 on CPU core #2
[Feb 18 12:19] Starting Gerbicz error-checking PRP test of M82589933 using FFT length 4480K, Pass1=896, Pass2=5K, clm=4, 2 threads
[Feb 18 12:19] Iteration: 1 / 82589933 [0.00%], ms/iter: 1332.191, ETA: 1273d 10:40
[Feb 18 12:19] Iteration: 2 / 82589933 [0.00%], ms/iter: 212.367, ETA: 203d 00:02
[Feb 18 12:19] M82589933 interim PRP residue 000000000000001B at iteration 1
[Feb 18 12:19] Iteration: 3 / 82589933 [0.00%], ms/iter: 213.918, ETA: 204d 11:38
[Feb 18 12:19] M82589933 interim PRP residue 000000000000088B at iteration 2
[Feb 18 12:19] Iteration: 4 / 82589933 [0.00%], ms/iter: 246.031, ETA: 235d 04:21
[Feb 18 12:19] M82589933 interim PRP residue 0000000000DAF26B at iteration 3
[Feb 18 12:19] Iteration: 5 / 82589933 [0.00%], ms/iter: 210.429, ETA: 201d 03:34
[Feb 18 12:19] M82589933 interim PRP residue 000231C54B5F6A2B at iteration 4
[Feb 18 12:19] Iteration: 6 / 82589933 [0.00%], ms/iter: 226.473, ETA: 216d 11:39
[Feb 18 12:19] M82589933 interim PRP residue D310B7D97DD4E9AB at iteration 5
[Feb 18 12:19] Iteration: 7 / 82589933 [0.00%], ms/iter: 205.084, ETA: 196d 00:58
[Feb 18 12:20] M82589933 interim PRP residue 2AC0B180838228AB at iteration 6
[Feb 18 12:20] Iteration: 8 / 82589933 [0.00%], ms/iter: 244.553, ETA: 233d 18:26
[Feb 18 12:20] M82589933 interim PRP residue 9B5ACA650265A6AB at iteration 7
[Feb 18 12:20] Iteration: 9 / 82589933 [0.00%], ms/iter: 211.244, ETA: 201d 22:16
[Feb 18 12:20] M82589933 interim PRP residue B47759B0D250A2AB at iteration 8
[Feb 18 12:20] Iteration: 10 / 82589933 [0.00%], ms/iter: 250.999, ETA: 239d 22:19
[Feb 18 12:20] M82589933 interim PRP residue DF36E033DAB69AAB at iteration 9
[Feb 18 12:20] Iteration: 11 / 82589933 [0.00%], ms/iter: 215.197, ETA: 205d 16:58
[Feb 18 12:20] M82589933 interim PRP residue C94525688DC28AAB at iteration 10
[Feb 18 12:20] Iteration: 12 / 82589933 [0.00%], ms/iter: 218.688, ETA: 209d 01:04
[Feb 18 12:20] M82589933 interim PRP residue FC7BEC947CDA6AAB at iteration 11
[Feb 18 12:20] Iteration: 13 / 82589933 [0.00%], ms/iter: 213.036, ETA: 203d 15:24
[Feb 18 12:20] Stopping PRP test of M82589933 at iteration 13 [0.00%]
[Feb 18 12:20] Worker stopped.
5) Note the seemingly out of order output above for PRP3; interim residue for iteration 11 appearing after iteration timing for iteration 12, for example. This is also behavior that was observed in v29.4. See also https://www.mersenneforum.org/showpo...&postcount=428

6) The skipped-interim-residue observed for LL also occurs for PRP3. Note there's no interim residue given for iteration 12 below. It's not known whether this skipping may also occur for cases of higher spacing than 1, but it seems likely it does, with lowered probability as the interim interval is increased.
Code:
[Feb 18 12:20] M82589933 interim PRP residue C94525688DC28AAB at iteration 10
[Feb 18 12:20] Iteration: 12 / 82589933 [0.00%], ms/iter: 218.688, ETA: 209d 01:04
[Feb 18 12:20] M82589933 interim PRP residue FC7BEC947CDA6AAB at iteration 11
[Feb 18 12:20] Iteration: 13 / 82589933 [0.00%], ms/iter: 213.036, ETA: 203d 15:24
[Feb 18 12:20] Stopping PRP test of M82589933 at iteration 13 [0.00%]
[Feb 18 12:20] Worker stopped.
[Feb 18 12:29] Worker starting
[Feb 18 12:29] Setting affinity to run worker on CPU core #1
[Feb 18 12:29] Setting affinity to run helper thread 1 on CPU core #2
[Feb 18 12:29] Resuming Gerbicz error-checking PRP test of M82589933 using FFT length 4480K, Pass1=896, Pass2=5K, clm=4, 2 threads
[Feb 18 12:29] Iteration: 14 / 82589933 [0.00%].
[Feb 18 12:29] M82589933 interim PRP residue 628CE0A31369AAAB at iteration 13
[Feb 18 12:29] Iteration: 15 / 82589933 [0.00%], ms/iter: 121.541, ETA: 116d 04:21
[Feb 18 12:29] M82589933 interim PRP residue B332521E7C28AAAB at iteration 14
kriesel is offline   Reply With Quote
Old 2019-02-18, 20:50   #5
GP2
 
GP2's Avatar
 
Sep 2003

22·3·5·43 Posts
Default

I am doing PRP tests of Wagstaff exponents.

Several dozen machines resumed 29.5 savefiles without a problem, but one has a problem that seems similar to the one reported by Simon.

All of the exponents on all of the machines seem to have a problem similar to the one reported by Simon.

I tried moving away the 29.5 savefiles and starting the same exponent from scratch with 29.6, but the same problem occurred.

The Wagstaff exponent in question is 9081307 (edit: it's not just this exponent). I am testing another exponent now to see if the problem is the exponent or the machine.


The exponent passes the Gerbicz error check at iteration p−1 (9081306), but then fails somehow in the final processing. It continues in an infinite error loop and does not attempt to resume from earlier savefiles. Only when I interrupt the program with SIGINT does it start to process earlier savefiles, but then it terminates right after, obviously.

When I manually delete the more recent savefiles and try to resume from older savefiles (at iteration 9 million and at 8 million), the same problem happens: success until iteration p−1 and then the same infinite error loop.


Code:
PRP=1,2,9081307,1,"3"
Code:
WorkerThreads=1
CoresPerTest=2
HyperthreadLL=1
Code:
PRPBase=3
PRPResidueType=5
results.txt:

Code:
[Mon Feb 18 05:44:04 2019]
ERROR: Comparing PRP double-check values failed.  Rolling back to iteration 9081306.
Continuing from last save file.
ERROR: Comparing PRP double-check values failed.  Rolling back to iteration 9081306.
Continuing from last save file.
ERROR: Comparing PRP double-check values failed.  Rolling back to iteration 9081306.
Continuing from last save file.
ERROR: Comparing PRP double-check values failed.  Rolling back to iteration 9081306.
Continuing from last save file.
...
(about 80,000 lines like this, and growing rapidly)
I sent a SIGINT to stop the program. Then the following lines appeared at the bottom of the results.txt file, after all the tens of thousands of ERROR lines:

Code:
Error reading intermediate file: p9081307
Renaming p9081307 to p9081307.bad1
Trying backup intermediate file: p9081307.bu
The rename succeeded, but obviously the program did terminate right after from the SIGINT.


I tried restarting from each savefile (at iterations 8 million, 9 million, and higher), and all of them eventually gave the same error.

For instance, resuming from iteration 9 million (the .bu3 file):

Code:
$ ./mprime -d
[Main thread Feb 18 18:23] Mersenne number primality test program version 29.6
[Main thread Feb 18 18:23] Optimizing for CPU architecture: Core i3/i5/i7, L2 cache size: 2x1 MB, L3 cache size: 25344 KB
[Main thread Feb 18 18:23] Starting worker.
[Work thread Feb 18 18:23] Worker starting
[Work thread Feb 18 18:23] Setting affinity to run worker on CPU core #1
[Work thread Feb 18 18:23] Setting affinity to run helper thread 1 on CPU core #1
[Work thread Feb 18 18:23] Setting affinity to run helper thread 3 on CPU core #2
[Work thread Feb 18 18:23] Setting affinity to run helper thread 2 on CPU core #2
[Work thread Feb 18 18:23] Trying backup intermediate file: p9081307.bu3
[Work thread Feb 18 18:23] Resuming Gerbicz error-checking PRP test of (2^9081307+1)/3 using all-complex AVX-512 FFT length 480K, Pass1=128, Pass2=3840, clm=2, 4 threads
[Work thread Feb 18 18:23] Iteration: 9000001 / 9081307 [99.10%].
[Work thread Feb 18 18:23] Iteration: 9010000 / 9081307 [99.21%], ms/iter:  0.712, ETA: 00:00:50
[Work thread Feb 18 18:23] Iteration: 9020000 / 9081307 [99.32%], ms/iter:  0.710, ETA: 00:00:43
[Work thread Feb 18 18:23] Iteration: 9030000 / 9081307 [99.43%], ms/iter:  0.710, ETA: 00:00:36
[Work thread Feb 18 18:23] Iteration: 9040000 / 9081307 [99.54%], ms/iter:  0.712, ETA: 00:00:29
[Work thread Feb 18 18:23] Iteration: 9050000 / 9081307 [99.65%], ms/iter:  0.710, ETA: 00:00:22
[Work thread Feb 18 18:23] Iteration: 9060000 / 9081307 [99.76%], ms/iter:  0.710, ETA: 00:00:15
[Work thread Feb 18 18:23] Iteration: 9070000 / 9081307 [99.87%], ms/iter:  0.711, ETA: 00:00:08
[Work thread Feb 18 18:23] Iteration: 9080000 / 9081307 [99.98%], ms/iter:  0.711, ETA: 00:00:00
[Work thread Feb 18 18:24] Gerbicz error check passed at iteration 9081225.
[Work thread Feb 18 18:24] Gerbicz error check passed at iteration 9081306.
[Work thread Feb 18 18:24] ERROR: Comparing PRP double-check values failed.  Rolling back to iteration 9081306.
[Work thread Feb 18 18:24] Continuing from last save file.
[Work thread Feb 18 18:24] Setting affinity to run helper thread 1 on CPU core #1
[Work thread Feb 18 18:24] Setting affinity to run helper thread 3 on CPU core #2
[Work thread Feb 18 18:24] Setting affinity to run helper thread 2 on CPU core #2
[Work thread Feb 18 18:24] Trying backup intermediate file: p9081307.bu3
[Work thread Feb 18 18:24] Resuming Gerbicz error-checking PRP test of (2^9081307+1)/3 using all-complex AVX-512 FFT length 480K, Pass1=128, Pass2=3840, clm=2, 4 threads
[Work thread Feb 18 18:24] Iteration: 9000001 / 9081307 [99.10%].
[Work thread Feb 18 18:24] Hardware errors have occurred during the test!
[Work thread Feb 18 18:24] 1 Gerbicz/double-check error.
[Work thread Feb 18 18:24] Confidence in final result is excellent.
[Work thread Feb 18 18:24] Iteration: 9010000 / 9081307 [99.21%], ms/iter:  0.714, ETA: 00:00:50
[Work thread Feb 18 18:24] Hardware errors have occurred during the test!
[Work thread Feb 18 18:24] 1 Gerbicz/double-check error.
[Work thread Feb 18 18:24] Confidence in final result is excellent.
[Work thread Feb 18 18:24] Iteration: 9020000 / 9081307 [99.32%], ms/iter:  0.713, ETA: 00:00:43
[Work thread Feb 18 18:24] Hardware errors have occurred during the test!
[Work thread Feb 18 18:24] 1 Gerbicz/double-check error.
[Work thread Feb 18 18:24] Confidence in final result is excellent.
[Work thread Feb 18 18:24] Iteration: 9030000 / 9081307 [99.43%], ms/iter:  0.712, ETA: 00:00:36
[Work thread Feb 18 18:24] Hardware errors have occurred during the test!
[Work thread Feb 18 18:24] 1 Gerbicz/double-check error.
[Work thread Feb 18 18:24] Confidence in final result is excellent.
[Work thread Feb 18 18:24] Iteration: 9040000 / 9081307 [99.54%], ms/iter:  0.713, ETA: 00:00:29
[Work thread Feb 18 18:24] Hardware errors have occurred during the test!
[Work thread Feb 18 18:24] 1 Gerbicz/double-check error.
[Work thread Feb 18 18:24] Confidence in final result is excellent.
[Work thread Feb 18 18:24] Iteration: 9050000 / 9081307 [99.65%], ms/iter:  0.713, ETA: 00:00:22
[Work thread Feb 18 18:24] Hardware errors have occurred during the test!
[Work thread Feb 18 18:24] 1 Gerbicz/double-check error.
[Work thread Feb 18 18:24] Confidence in final result is excellent.
[Work thread Feb 18 18:24] Iteration: 9060000 / 9081307 [99.76%], ms/iter:  0.714, ETA: 00:00:15
[Work thread Feb 18 18:24] Hardware errors have occurred during the test!
[Work thread Feb 18 18:24] 1 Gerbicz/double-check error.
[Work thread Feb 18 18:24] Confidence in final result is excellent.
[Work thread Feb 18 18:24] Iteration: 9070000 / 9081307 [99.87%], ms/iter:  0.715, ETA: 00:00:08
[Work thread Feb 18 18:24] Hardware errors have occurred during the test!
[Work thread Feb 18 18:24] 1 Gerbicz/double-check error.
[Work thread Feb 18 18:24] Confidence in final result is excellent.
[Work thread Feb 18 18:24] Iteration: 9080000 / 9081307 [99.98%], ms/iter:  0.713, ETA: 00:00:00
[Work thread Feb 18 18:24] Hardware errors have occurred during the test!
[Work thread Feb 18 18:24] 1 Gerbicz/double-check error.
[Work thread Feb 18 18:24] Confidence in final result is excellent.
[Work thread Feb 18 18:24] Gerbicz error check passed at iteration 9081225.
[Work thread Feb 18 18:24] Gerbicz error check passed at iteration 9081306.
[Work thread Feb 18 18:24] ERROR: Comparing PRP double-check values failed.  Rolling back to iteration 9081306.
[Work thread Feb 18 18:24] Continuing from last save file.
[Work thread Feb 18 18:24] Setting affinity to run helper thread 1 on CPU core #1
[Work thread Feb 18 18:24] Setting affinity to run helper thread 3 on CPU core #2
[Work thread Feb 18 18:24] Setting affinity to run helper thread 2 on CPU core #2
[Work thread Feb 18 18:24] Trying backup intermediate file: p9081307.bu3
[Work thread Feb 18 18:24] Resuming Gerbicz error-checking PRP test of (2^9081307+1)/3 using all-complex AVX-512 FFT length 480K, Pass1=128, Pass2=3840, clm=2, 4 threads
[Work thread Feb 18 18:24] Iteration: 9000001 / 9081307 [99.10%].
[Work thread Feb 18 18:24] Hardware errors have occurred during the test!
[Work thread Feb 18 18:24] 2 Gerbicz/double-check errors.
[Work thread Feb 18 18:24] Confidence in final result is excellent.
[Work thread Feb 18 18:25] Iteration: 9010000 / 9081307 [99.21%], ms/iter:  0.711, ETA: 00:00:50
[Work thread Feb 18 18:25] Hardware errors have occurred during the test!
[Work thread Feb 18 18:25] 2 Gerbicz/double-check errors.
[Work thread Feb 18 18:25] Confidence in final result is excellent.
[Work thread Feb 18 18:25] Iteration: 9020000 / 9081307 [99.32%], ms/iter:  0.711, ETA: 00:00:43
[Work thread Feb 18 18:25] Hardware errors have occurred during the test!
[Work thread Feb 18 18:25] 2 Gerbicz/double-check errors.
[Work thread Feb 18 18:25] Confidence in final result is excellent.
[Work thread Feb 18 18:25] Iteration: 9030000 / 9081307 [99.43%], ms/iter:  0.711, ETA: 00:00:36
[Work thread Feb 18 18:25] Hardware errors have occurred during the test!
[Work thread Feb 18 18:25] 2 Gerbicz/double-check errors.
[Work thread Feb 18 18:25] Confidence in final result is excellent.
[Main thread Feb 18 18:25] Stopping all worker threads.
[Work thread Feb 18 18:25] Stopping PRP test of (2^9081307+1)/3 at iteration 9036608 [99.50%]
[Work thread Feb 18 18:25] Worker stopped.
[Main thread Feb 18 18:25] Execution halted.
If you let it continue, eventually it reaches "15 or more Gerbicz/double-check errors."

Last fiddled with by GP2 on 2019-02-18 at 21:07
GP2 is offline   Reply With Quote
Old 2019-02-18, 20:52   #6
GP2
 
GP2's Avatar
 
Sep 2003

22×3×5×43 Posts
Default

PS, if you have any test debug programs, I can try running them on this particular virtual machine.

However it's a spot instance, so it could go away at any time with only two minutes' warning.

The problem is occurring for all Wagstaff exponents on all machines.

Last fiddled with by GP2 on 2019-02-18 at 21:08
GP2 is offline   Reply With Quote
Old 2019-02-18, 20:53   #7
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

157718 Posts
Default

Quote:
Originally Posted by GP2 View Post
For instance, resuming from iteration 9 million (the .bu3 file):
Please email this save file (as well as worktodo.txt). Thanks.
Prime95 is offline   Reply With Quote
Old 2019-02-18, 20:53   #8
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

3×7×11×31 Posts
Default

BUG: x87 FFTs are not working in the 32-bit builds.
Prime95 is offline   Reply With Quote
Old 2019-02-18, 21:25   #9
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

22×11×101 Posts
Default Minimum processor, OS version?

Tried the win32 Prime95 29.6b1 on Pentium 133 NT 4 sp6a, and got Dr Watson illegal instruction fatal error very early, before the app GUI appeared or any progress files were created. I guess that's expected. Last version I had run on this box successfully was 25.11 it appears.

V29.6b1 seems to run on Pentium M on Vista or XP.
Attached Thumbnails
Click image for larger version

Name:	296b1-p133-nt4-boom.png
Views:	235
Size:	25.9 KB
ID:	19908  

Last fiddled with by kriesel on 2019-02-18 at 21:43
kriesel is offline   Reply With Quote
Old 2019-02-18, 21:46   #10
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

22·11·101 Posts
Default readme update

Although the default behavior has been changed to 1/16 of RAM, the readme.txt still says
Quote:
If at all in doubt, leave the settings at 8MB.
kriesel is offline   Reply With Quote
Old 2019-02-18, 22:05   #11
chalsall
If I May
 
chalsall's Avatar
 
"Chris Halsall"
Sep 2002
Barbados

2·4,657 Posts
Default

Quote:
Originally Posted by kriesel View Post
Although the default behavior has been changed to 1/16 of RAM, the readme.txt still says
Please note that except for those who work outside of the nominal range(s), P-1 will have already been done, and so the RAM needed will be small.

Not to say the "readme" shouldn't be updated to be accurate....
chalsall is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Prime95 version 29.2 Prime95 Software 71 2017-09-16 16:55
Prime95 version 29.1 Prime95 Software 95 2017-08-22 22:46
Prime95 version 26.5 Prime95 Software 175 2011-04-04 22:35
Prime95 version 25.9 Prime95 Software 143 2010-01-05 22:53
Prime95 version 25.8 Prime95 Software 159 2009-09-21 16:30

All times are UTC. The time now is 10:53.

Mon Sep 28 10:53:34 UTC 2020 up 18 days, 8:04, 0 users, load averages: 1.57, 1.53, 1.48

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, 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.