mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Msieve

Reply
 
Thread Tools
Old 2018-09-30, 00:15   #1
firejuggler
 
firejuggler's Avatar
 
"Vincent"
Apr 2010
Over the rainbow

22·11·61 Posts
Default Msieve failed to polyselect

Since I have updated my drivers, my Msieve ( wich worked before) fail to work with this message

Code:
error (sort_engine.cu:163) a PTX JIT compilation failed
I use a Geforce 750 Ti (compute capability 5.0).
I tried rollnack the drivers, update my cuda toolkit, , re-update the drivers, to no avail.
Any idea?

Last fiddled with by firejuggler on 2018-09-30 at 00:20
firejuggler is offline   Reply With Quote
Old 2018-09-30, 07:21   #2
VictordeHolland
 
VictordeHolland's Avatar
 
"Victor de Hollander"
Aug 2011
the Netherlands

117610 Posts
Default

Is this on Windows or Linux?
Driver/ CUDA toolkit version?
Msieve version?

Last fiddled with by VictordeHolland on 2018-09-30 at 07:31
VictordeHolland is offline   Reply With Quote
Old 2018-09-30, 09:17   #3
Gimarel
 
Apr 2010

22·47 Posts
Default

I think that cuda got more picky on compute capabilities.
You can try the following patch.
Code:
Index: Makefile
===================================================================
--- Makefile    (Revision 1025)
+++ Makefile    (Arbeitskopie)
@@ -347,4 +347,4 @@
     $(NVCC) -arch sm_50 -ptx -o $@ $<
 
 cub/built:
-    cd cub && make WIN=$(WIN) WIN64=$(WIN64) sm=200,300,350,520 && cd ..
+    cd cub && make WIN=$(WIN) WIN64=$(WIN64) sm=200,300,350,500 && cd ..
Index: cub/Makefile
===================================================================
--- cub/Makefile    (Revision 1025)
+++ cub/Makefile    (Arbeitskopie)
@@ -27,8 +27,8 @@
     SM_ARCH = 200
 endif
 
-ifeq (520, $(findstring 520, $(SM_ARCH)))
-    SM_TARGETS     += -gencode=arch=compute_52,code=\"sm_52,compute_52\" 
+ifeq (500, $(findstring 500, $(SM_ARCH)))
+    SM_TARGETS     += -gencode=arch=compute_50,code=\"sm_50,compute_50\" 
     SM_DEF         += -DSM520
 endif
 ifeq (370, $(findstring 370, $(SM_ARCH)))
Gimarel is offline   Reply With Quote
Old 2018-10-01, 13:47   #4
jasonp
Tribal Bullet
 
jasonp's Avatar
 
Oct 2004

3×1,181 Posts
Default

Ouch, the compile times are going to balloon if we have to separately compile for every single compute version out there.
jasonp is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Base method of montgomery polyselect implementation csg Homework Help 2 2017-12-13 07:04
Playing with CADO's individual polyselect binaries Dubslow CADO-NFS 22 2016-07-04 19:54
To CPU or to GPU: stage1 polyselect research Karl M Johnson Msieve 8 2011-11-12 19:46
Assertion failed Andi47 GMP-ECM 2 2009-08-04 14:11
integrator failed MatWur-S530113 Msieve 5 2009-01-28 03:54

All times are UTC. The time now is 18:07.


Mon Dec 6 18:07:46 UTC 2021 up 136 days, 12:36, 0 users, load averages: 2.13, 1.92, 1.84

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.