On 05/12/2011 04:25 PM, Jette, Moe wrote:
The function _slurm_destroy_association_shares_object() is declared, although several header files are used in the path:/g/g0/jette/slurm.chaos/src/common/slurm_protocol_defs.h included from /g/g0/jette/slurm.chaos/src/common/checkpoint.h:47, from /g/g0/jette/slurm.chaos/src/slurmctld/slurmctld.h:70 from /g/g0/jette/slurm.chaos/src/common/slurm_accounting_ from /g/g0/jette/slurm.chaos/src/common/assoc_mgr.h:48, from /g/g0/jette/slurm.chaos/src/common/assoc_mgr.c:39: Just out of curiosity, can anyone tell me why there has been a lot of recent interest in SLURM on OSX? Are people building Mac clusters? ________________________________________ btw, is the function _slurm_destroy_association_shares_object() declared in any header that gets included while building assoc_mgr.c? (might help, however, the behaviour you're seeing here is still plain *wrong* for gcc, prototype declared or not) regards, jaKa
As to why the use of OS X is increasing in general, I can't say, but in my specific case I'm working for a group of biomedical researchers (working with DNA sequencing) at a university who have an 8-unit xserve cluster. From what I understand, several of the programs they use only work on macs, so they tend to use it for much of their number crunching.
I've tried including the slurm_protocol_defs.h header file straight into assoc_mgr.c with the same result, so I've pretty much concluded it's a compiler bug as well. Now if I could just figure out how to work around it...
In the mean time I'm trying to see if cross-compiling SLURM on my linux box might be a viable way of getting this working, but that's proving to be no small challenge in and of itself - just getting the compiler to compile is a hassle.
Thanks, Tyler
