mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Msieve (https://www.mersenneforum.org/forumdisplay.php?f=83)
-   -   Msieve 1.10 (https://www.mersenneforum.org/showthread.php?t=6299)

RedGolpe 2006-09-06 07:14

Msieve 1.10
 
I downloaded the binary Windows executable file for the new Msieve version (1.10) and changed the .exe file but when I stop and restart it, it won't use the previous data in msieve.dat: instead, it deletes the file and starts a new one. Unfortunately, when I saw 1.10 working properly, I deleted the 1.07 file, which was working properly. :blush:
So, could anyone help me with version 1.10 (like posting a correctly working binary for Windows) or just post the old 1.07 executable? Thank you for your time.

RedGolpe 2006-09-06 14:37

[QUOTE=RedGolpe;86436]when I saw 1.10 working properly[/QUOTE]

I obviously meant "when I THOUGHT that 1.10 was working properly".

jasonp 2006-09-06 17:10

[QUOTE=RedGolpe;86436]I downloaded the binary Windows executable file for the new Msieve version (1.10) and changed the .exe file but when I stop and restart it, it won't use the previous data in msieve.dat: instead, it deletes the file and starts a new one. Unfortunately, when I saw 1.10 working properly, I deleted the 1.07 file, which was working properly. :blush:
So, could anyone help me with version 1.10 (like posting a correctly working binary for Windows) or just post the old 1.07 executable? Thank you for your time.[/QUOTE]
If you still have the logfile from your 1.07 run, compare the "using multiplier of X" line there with the "using multiplier of Y" line in the 1.10 version. If they are different for your factorization, then you cannot restart your 1.07 job with 1.10

If you're in a position to build from source, you can force the choice of multiplier to match up between the two versions, and then restart will work as expected. If both versions already have the same multiplier, then it's a bug and I can try to find out what's going on.

Hope you didn't waste too much work,
jasonp

RedGolpe 2006-09-06 22:09

Thank you for your reply. I obviously did not make myself clear: I actually STARTED this run with 1.10 but when I stop it, it won't use the msieve.dat file. Instead, it creates a brand new one so wasting all previous computation, so I suppose there's a bug, at least in the precompiled binary for Windows. I would use the 1.07 which has always worked as expected but I don't have the executable anymore.

jasonp 2006-09-06 22:18

[QUOTE=RedGolpe;86499]Thank you for your reply. I obviously did not make myself clear: I actually STARTED this run with 1.10 but when I stop it, it won't use the msieve.dat file. Instead, it creates a brand new one so wasting all previous computation, so I suppose there's a bug, at least in the precompiled binary for Windows. I would use the 1.07 which has always worked as expected but I don't have the executable anymore.[/QUOTE]
I follow now. Can you PM me the input you used that caused this behavior? Does it happen attempting to restart from any factorization?

jasonp

RedGolpe 2006-09-07 07:14

I sent you the input file via PM. Oddly enough, for other inputs the computation resumes correctly.

jasonp 2006-09-07 12:56

[QUOTE=RedGolpe;86521]I sent you the input file via PM. Oddly enough, for other inputs the computation resumes correctly.[/QUOTE]
Still trying to get to the bottom of this (your input works fine for me). In the meantime, anybody who sees suspicious behavior from the latest binary can revert to

[url]www.boo.net/~jasonp/msieve107.tar.gz[/url]
[url]www.boo.net/~jasonp/msieve107.exe[/url]

jasonp


All times are UTC. The time now is 06:04.

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