mpirun

[cfz1998]

Could FunGAP support mpirun for maker? The maker process is so slower.

mpirun -n num_cores maker

[mbnmbn00]

Thanks for the suggestion. If you look at maker_opts.ctl file, there's an option called cpus=, and FunGAP changes this number to user-provided one.

Do you think mpirun makes it run faster than putting the number in cpus=? If you do, I definitely check that out and update the code.

[cfz1998]

In other cases (like plant genome), mpirun may run faster than putting the number in cpus=. But this may introduces some configuration file issues.

And i think the codes in the run_maker.py line 466 -472 may cause an exception. For me, the path of blastn not in "os.path.dirname(maker_bin)". I think it would be better to change it to this

    # Program paths
    for program in [
            'makeblastdb', 'blastn', 'blastx', 'tblastx', 'RepeatMasker',
            'exonerate', 'snap', 'augustus', 'tRNAscan-SE', 'snoscan']:
        replace('maker_exe.ctl', '{}='.format(program), '{}={}'.format(
            program, D_CONF[program.upper()+"_PATH"]
        ))

[mbnmbn00]

Thanks. I have updated the code! Let me know if you have further questions!!