cuda-gdb Error message -
i tried debug cuda application cuda-gdb got weird error.
i set option -g -g -o0 build application. run program without cuda-gdb, didn't correct result. hence decided use cuda-gdb, however, got following error message while running program cuda-gdb
 error: failed read valid warps mask (dev=1, sm=0, error=16). 
what means? why sm=0 , what's meaning of error=16?
update 1: tried use cuda-gdb cuda samples, fails same problem. installed cuda 6.0 toolkit followed instruction of nvidia. problem of system?
update 2:
- os - centos 6.5
 - gpu 
- 1 quadro 400
 - 2 tesla c2070
 - i'm using 1 gpu program, i've got same bug message gpu selected
 
 - cuda version - 6.0
 - gpu driver 
- nvrm version: nvidia unix x86_64 kernel module 331.62 wed mar 19 18:20:03 pdt 2014
 - gcc version: gcc version 4.4.7 20120313 (red hat 4.4.7-4) (gcc)
 
 
update 3: tried more information in cuda-gdb, got following results
 (cuda-gdb) info cuda devices error: failed read valid warps mask (dev=1, sm=0, error=16). (cuda-gdb) info cuda sms focus not set on active cuda kernel. (cuda-gdb) info cuda lanes focus not set on active cuda kernel. (cuda-gdb) info cuda kernels no cuda kernels. (cuda-gdb) info cuda contexts no cuda contexts. 
actually, issue specific old nvidia gpus(like "quadro 400", "geforce gt220", or "geforce gt 330m", etc).
on liam kim's setup, cuda-gdb should work fine set environment variable "cuda_visible_devices", , let cuda-gdb running on tesla c2070 gpus specifically. i.e $export cuda_visible_devices=0 (or 2) - exact cuda devices index found running cuda sample - "devicequery".
and now, issue has been fixed, fix availble cuda developers in next cuda release(it posted out around july, 2014).
Comments
Post a Comment