Re: Similar systems, different performance of fortran code
On Thu, 2005-04-21 at 17:16 +0100, michael wrote:
> On Fri, 2005-04-22 at 00:42 +0900, Victor Munoz wrote:
> > I recently found a C++ class to measure time spent between lines of code,
> > but this is fortran. I don't see the profiling tools I know of (gprof,
> > gconv, mpatrol) as able to measure time line by line. Any tool/suggestion to
> > do it fortran (which is not my "mother tongue")?
>
> The gprof man page says:
>
> "--line"
> The -l option enables line-by-line profiling, which causes his-
> togram hits to be charged to individual source code lines, instead
> of functions. If the program was compiled with basic-block count-
> ing enabled, this option will also identify how many times each
> line of code was executed. While line-by-line profiling can help
> isolate where in a large function a program is spending its time,
> it also significantly increases the running time of "gprof", and
> magnifies statistical inaccuracies.
>
> so if your compilers meet this criteria, away you go! (Hang on, you said
> 'Fortran' but also 'gcc' did you not...........)
whoops... that was a red herring - had forgotten gcc could handle
Fortran given gcc has the relevant libs (which again should be same on
both systems for valid comparisons)
>
>
--
Michael Bane
Atmospheric Physics Group
University of Manchester
Reply to: