FASTA program discussion list

[Haruna Cofer <>]

Hello again, Bill!

I seem to have encountered an error in the MPI version of FASTY fasta-36.3.8
(also occurs in fasta-36.3.8a).  The code fails with:

cannot allocate buf2_data[-44544]

It looks like this happens because int_msg_b[1] in work_comp() is not
initialized.  So I added this line to init_thr():

    int_msg_b[1] = m_bufi_p->max_chain_seqs * (m_msp->nitt1+1 -
m_msp->revcomp);

and the code seems to get past this now.  However, it then fails with a 
SIGSEGV
at a different place:

MPT: (gdb) #0  0x00002aaaab16b345 in waitpid () from /lib64/libpthread.so.0
MPT: #1  0x00002aaaab8730ac in mpi_sgi_system (command=<optimized out>) at
sig.c:99
MPT: #2  MPI_SGI_stacktraceback (header=<optimized out>) at sig.c:319
MPT: #3  0x00002aaaab873b8f in first_arriver_handler (signo=11,
MPT:     stack_trace_sem=0x2aaab4220500) at sig.c:468
MPT: #4  0x00002aaaab873e1f in slave_sig_handler (signo=11,
MPT:     siginfo=<optimized out>, extra=<optimized out>) at sig.c:542
MPT: #5  <signal handler called>
MPT: #6  0x0000000000431901 in do_work ()
MPT: #7  0x000000000041582f in buf_do_work ()
MPT: #8  0x0000000000409549 in work_comp (my_worker=1) at work_thr2.c:413
MPT: #9  0x000000000040279b in main ()
MPT: (gdb)

Enabling -DDEBUG output reveals this:

[0] incorrect cur_aa1b_size: -1944560025 != 526207 [-1408627136]

I'm still debugging this one, so please let me know if you have any insights 
on
it.  Thank you again, Bill!

-- Haruna  :)