mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Msieve (https://www.mersenneforum.org/forumdisplay.php?f=83)
-   -   CUDA 64-bit compilation (https://www.mersenneforum.org/showthread.php?t=19094)

wombatman 2014-01-28 05:16

CUDA 64-bit compilation
 
[STRIKE]I'm trying to compile MSieve on Win 7 Ultimate--it actually compiles successfully, but when I actually run it, I immediately get a c000001d exception. Debugging in VS2012 gives the error at this line:

[CODE]0000000000402B53 vmovaps xmmword ptr [rsp+0B40h],xmm6[/CODE]

MSieve was compiled with core-avx-i as the -march flag. Any suggestions on what may be causing the error?


Edit: Alright, it's not a CUDA problem. Even just building with "make all WIN=1" and setting the march flag to native, I get the same error.[/STRIKE]

Edit 2: Compiling with march as -corei7 allows the program to work as long as CUDA is not involved. With CUDA=1, I now get a [B]c0000005[/B] exception.

wombatman 2014-02-08 18:40

Still can't get around the C0000005 exception error.

I get this far when I run Msieve with CUDA enabled:
[CODE]Msieve v. 1.52 (SVN 959M)
Wed Feb 05 21:57:01 2014
random seeds: 291dab40 2432dee9
factoring 4085684988171507780203327299922394806290879077936133966700748091011780
92705586290904977944808947918659909212360990889226885021373244890483039858860243
2892817233 (160 digits)
searching for 15-digit factors
commencing number field sieve (160-digit input)
commencing number field sieve polynomial selection
polynomial degree: 5
max stage 1 norm: 2.30e+024
max stage 2 norm: 2.53e+022
min E-value: 9.32e-013
poly select deadline: 1080000
time limit set to 300.00 CPU-hours
expecting poly E from 1.33e-012 to > 1.52e-012
searching leading coefficients from 1 to 31874397[/CODE]

Any ideas on where to start looking for issues would be greatly appreciated.


All times are UTC. The time now is 20:34.

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