Hello everyone!
I’ve installed the trial version of PGI on the cluster (on the shared disk on the manager node. the shared disk is available on every computational node).
CUDA Toolkit 4.1 was installed in the manager node:
[nguyen@nks-g6 msu_spring_2012]$ nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2011 NVIDIA Corporation
Built on Thu_Nov_17_17:38:12_PST_2011
Cuda compilation tools, release 4.1, V0.2.1221
PGI was installed without built-in CUDA support. A symbolic link had been created to directory /usr/local/cuda:
[nguyen@nks-g6 msu_spring_2012]$ cd /ifs/opt/pgi/linux86-64/2012/cuda
[nguyen@nks-g6 cuda]$ ls -la
total 8
drwxr-xr-x 2 root root 4096 Apr 5 20:28 .
drwxr-xr-x 7 root root 4096 Apr 5 00:13 ..
lrwxrwxrwx 1 root root 15 Apr 5 00:13 4.1 -> /usr/local/cuda
When I tried to compile programs written in CUDA Fortran or OpenACC, I got the following error:
[nguyen@nks-g6 msu_spring_2012]$ pgfortran -Minfo=all -ta=nvidia vadd.cuf
NOTE: your trial license will expire in 12 days, 3.56 hours.
NOTE: your trial license will expire in 12 days, 3.56 hours.
Command-line error: invalid option: --no_export
1 catastrophic error detected in this compilation.
Compilation terminated.
PGF90-F-0000-Internal compiler error. Device compiler exited with error status code 0 (vadd.cuf: 49)
PGF90/x86-64 Linux 12.3-0: compilation aborted
[nguyen@nks-g6 msu_spring_2012]$ vim makefile
[nguyen@nks-g6 msu_spring_2012]$ make J1.exe
pgfortran -acc -Mcuda -Minfo=all -ta=nvidia -c J1.f90 -Minfo=accel
NOTE: your trial license will expire in 12 days, 3.55 hours.
NOTE: your trial license will expire in 12 days, 3.55 hours.
jacobi:
18, Generating copyout(newa(2:m-1,2:n-1))
Generating copyin(a(:m,:n))
Generating copyout(a(2:m-1,2:n-1))
19, Loop is parallelizable
20, Loop is parallelizable
Accelerator kernel generated
19, !$acc do parallel, vector(8) ! blockidx%y threadidx%y
20, !$acc do parallel, vector(8) ! blockidx%x threadidx%x
24, Max reduction generated for change
27, Loop is parallelizable
Accelerator kernel generated
27, !$acc do parallel, vector(8) ! blockidx%x threadidx%x
!$acc do parallel, vector(8) ! blockidx%y threadidx%y
Command-line error: invalid option: --no_export
1 catastrophic error detected in this compilation.
Compilation terminated.
pgfortran-Fatal-/ifs/opt/pgi/linux86-64/12.3/bin/pgf902 TERMINATED by signal 11
Arguments to /ifs/opt/pgi/linux86-64/12.3/bin/pgf902
/ifs/opt/pgi/linux86-64/12.3/bin/pgf902 /tmp/pgfortranRI1gb8cjjpmH.ilm -fn J1.f90 -opt 2 -terse 1 -inform warn -x 51 0x20 -x 119 0xa10000 -x 122 0x40 -x 123 0x1000 -x 127 4 -x 127 17 -x 19 0x400000 -x 28 0x40000 -x 120 0x10000000 -x 70 0x8000 -x 122 1 -x 125 0x20000 -quad -x 59 4 -x 59 4 -tp nehalem -x 120 0x1000 -x 124 0x1400 -y 15 2 -x 57 0x3b0000 -x 58 0x48000000 -x 49 0x100 -x 120 0x200 -astype 0 -x 124 1 -x 186 0x80000 -x 180 0x400 -x 180 0x4000000 -x 163 0x100 -x 137 1 -x 176 0x140000 -x 177 0x0202007f -x 186 1 -x 0 0x1000000 -x 2 0x100000 -x 0 0x2000000 -x 161 53239 -x 162 53239 -x 163 0x1 -x 186 1 -accel nvidia -x 176 0x140000 -x 177 0x0202007f -x 0 0x1000000 -x 2 0x100000 -x 0 0x2000000 -x 161 16384 -x 162 16384 -cmdline '+pgfortran J1.f90 -acc -Mcuda -Minfo=all -ta=nvidia -c -Minfo=accel' -asm /tmp/pgfortranZI1gzN9UBfC5.s
make: *** [J1.o] Error 127
I haven’t figure out what was wrong yet. Please somebody help me!!!
Thanks you so much.