Hi Alexander, thanks for the tip! This does seem to get me closer to solving the problem. At least I now see activity on the GPU. Where could I have come across this explanation and the hint that --disable-multilib will result in an incomplete build? We did configure with that switch (as we always do) and I was struggling for a while to get it working without (building with 32-bit support doesn't work, so eventually I've tried with --with-multilib-list=m64 instead - not sure this breaks things again). For a somewhat longer example, even when compiled with -foffload=-lgfortran I now get unresolved symbol _gfortran_os_error collect2: error: ld returned 1 exit status I suspect this is a similar problem? Thanks! On 29/01/2020 13:44, Alexander Monakov wrote: > On Wed, 29 Jan 2020, dbroemmel wrote: > >> Adding --save-temps and checking test.s I can see the lines >> // BEGIN GLOBAL VAR DECL: __nvptx_uni >> .extern .shared .u32 __nvptx_uni[32]; >> >> Running strace along with the compiler, I can see mkoffload failing and >> various accesses to crt*.o but not crt0.o which is where I suspect the >> missing symbol comes from? This is, however, clutching at straws since >> the OpenACC case also doesn't seem to touch crt0.o. >> >> What are we missing here? Any hint appreciated, thanks, > > Definition for that symbol is provided by libgcc.a from the offloading > compiler, namely mgomp.o member in the libgcc.a archive. > > Note that OpenMP offloading in GCC uses additional code generation > conventions on top of "standard" PTX conventions (see documentation > for the option '-mgomp' in the manual), and for that reason support > libraries are multilibbed: installed tree has an mgomp/ subdirectory > with versions of libgomp.a, libgcc.a and others for OpenMP offloading. > > mkoffload also knows to select the multilibs by passing -mgomp when > linking for OpenMP (i.e. when -fopenmp flag is present). > > If you configured with --disable-multilib, or accidentally have old > mkoffload or incomplete install tree, that would explain the problem. > > HTH > Alexander >
