I’ve been working on porting a fairly large project to CUDA, but I’m now receiving the following error when I attempt to compile.
[codebox]
#$ nvopencc -TARG:compute_13 -m32 “/tmp/tmpxft_000050c9_00000000-7_cudaSupport.cpp3.i” -o “/tmp/tmpxft_000050c9_00000000-2_cudaSupport.ptx”
Signal: Segmentation fault in Code_Expansion phase.
(0): Error: Signal Segmentation fault in phase Code_Expansion – processing aborted
*** Internal stack backtrace:
/usr/local/cuda/open64/lib//be [0x82edd22]
/usr/local/cuda/open64/lib//be [0x82ee700]
/usr/local/cuda/open64/lib//be [0x82eea8c]
/usr/local/cuda/open64/lib//be [0x82edf73]
/usr/local/cuda/open64/lib//be [0x82eed32]
[0x4001e400]
/usr/local/cuda/open64/lib//be [0x81a1dc3]
/usr/local/cuda/open64/lib//be [0x819ea8c]
/usr/local/cuda/open64/lib//be [0x819fd4b]
/usr/local/cuda/open64/lib//be [0x819edb3]
/usr/local/cuda/open64/lib//be [0x81a35a5]
/usr/local/cuda/open64/lib//be [0x81a41cc]
/usr/local/cuda/open64/lib//be [0x81a4735]
/usr/local/cuda/open64/lib//be [0x8180fa5]
/usr/local/cuda/open64/lib//be [0x804c501]
/usr/local/cuda/open64/lib//be [0x804d297]
/usr/local/cuda/open64/lib//be [0x804e5ff]
/lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe5) [0x40161775]
/usr/local/cuda/open64/lib//be [0x804aa81]
nvopencc INTERNAL ERROR: /usr/local/cuda/open64/lib//be died due to signal 4
–error 0x100 –
make: *** [rcracki_mt] Error 255
[/codebox]
the program compiles with the -deviceemu option.
System:
Ubuntu 9.04 32bit base install(no gui)
GTX 260
2.3 driver, tools and sdk (examples compile and run)
Any help as to what could cause such an error would be greatly appreciated.