Can this practice code be compiled with Gcc using acceleration right now?
I have installed gcc-offload-nvptx lib, but failed to compile laplace2d example with gcc -fopenacc.
All of the gcc commands test failed with following output:
/usr/bin/ld: /tmp/ccgHqtwA.crtoffloadtable.o: warning: relocation against __offload_vars_end' in read-only section .rodata'
/usr/bin/ld: /tmp/cctTGv8q.o: in function main': laplace2d-parallel.c:(.text+0x287): undefined reference to GOACC_parallel_keyed'
/usr/bin/ld: laplace2d-parallel.c:(.text+0x303): undefined reference to GOACC_parallel_keyed' /usr/bin/ld: /tmp/ccT24dnM.target.o: in function init':
ccGkwQvx.c:(.text+0x27): undefined reference to GOMP_offload_register_ver' /usr/bin/ld: /tmp/ccT24dnM.target.o: in function fini':
ccGkwQvx.c:(.text+0x55): undefined reference to `GOMP_offload_unregister_ver'
/usr/bin/ld: warning: creating DT_TEXTREL in a PIE
collect2: error: ld returned 1 exit status
Originally posted by @RudiFeiman in #14 (comment)
I have installed gcc-offload-nvptx lib, but failed to compile laplace2d example with gcc -fopenacc.
All of the gcc commands test failed with following output:
Originally posted by @RudiFeiman in #14 (comment)