Re: Parallel debugging
On Wed, Jul 29, 2009 at 09:04:47PM -0400, Daniel Underwood wrote: Thanks for the reply. I should have said that I'm also interested in profilers. I have limited experience (in Linux) using gprof and valgrind. gprof is part of the base system. Valgrind is available in ports, but only for the i386 architecture. Personally, I haven't felt the need to use a profiler in at least a decade. Current machines are so fast that even interpreted languages are fast enough for a lot of (smaller) programs. Roland -- R.F.Smith http://www.xs4all.nl/~rsmith/ [plain text _non-HTML_ PGP/GnuPG encrypted/signed email much appreciated] pgp: 1A2B 477F 9970 BA3C 2914 B7CE 1277 EFB0 C321 A725 (KeyID: C321A725) pgpTnAHS6Sf9K.pgp Description: PGP signature
Re: Parallel debugging
Valgrind is rather hopeless on fbsd for multithreaded programs Interesting. Thanks for the note. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Parallel debugging
Roland Smith wrote: gprof is part of the base system. Valgrind is available in ports, but only for the i386 architecture. Roland Unfortunately, Valgrind is rather hopeless on fbsd for multithreaded programs. I had to resort to using Ubuntu when Valgrinding. N :o) ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Parallel debugging
On Wed, Jul 29, 2009 at 12:58:19PM -0400, Daniel Underwood wrote: I'm programming a multithreaded scientific program using pthreads and c++. Currently The model is SMP, but may investigate MMP using MPI in the future. Can anyone suggest a good debugger and possibly an associated GUI? (E.g., gdb and ddd.) Both gdb and ddd can in principle be used with multithreaded programs. IMHO there is no good way to analyze a multithreaded program with a debugger, AFAIK. E.g. do not expect the program running in the debugger to behave the same as when running outside of the debugger, especially if you are dealing with timing issues or heisenbugs. The UNIX philiosophy would be to make multiple programs that can exchange (preferably plain text) data via pipes or shared memory. That way you can test the components separately. Look e.g. at postfix. Roland -- R.F.Smith http://www.xs4all.nl/~rsmith/ [plain text _non-HTML_ PGP/GnuPG encrypted/signed email much appreciated] pgp: 1A2B 477F 9970 BA3C 2914 B7CE 1277 EFB0 C321 A725 (KeyID: C321A725) pgpR5QeeU7NFy.pgp Description: PGP signature
Re: Parallel debugging
Thanks for the reply. I should have said that I'm also interested in profilers. I have limited experience (in Linux) using gprof and valgrind. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org