On 6/22/08, Ian Lance Taylor <[EMAIL PROTECTED]> wrote: > Jakub Jelinek <[EMAIL PROTECTED]> writes: > > > On Sat, Jun 21, 2008 at 08:58:05PM +0200, Laurent GUERBY wrote: > >> On Sat, 2008-06-21 at 08:10 -0400, Diego Novillo wrote: > >> > On Sat, Jun 21, 2008 at 05:21, Andrew Haley <[EMAIL PROTECTED]> wrote: > >> And for make -k check: > >> > >> -j1 2h18 > >> -j2 1h11 > >> -j4 0h55 > >> -j6 0h44 > > > > For make check, it would perhaps help to split the langest runtest.exp > > invocations (primarily check-gcc, maybe check-fortran and > > check-libstdc++-v3) > > into two parts, so that make could schedule them concurrently with higher > > make -jN -k check (e.g. for check-gcc run dg.exp in check-gcc-dg goal, > > all the ohter *.exp's in check-gcc-other and check-gcc: check-gcc-dg > > check-gcc-other). > > I think it would only be a few days of work for somebody familiar with > Tcl to add -j support to DejaGNU. I think that would be a very useful > contribution to gcc development.
I had looked into this when I first was confronted with the insane amount of time required to test on a Windows platform (mostly due to the overhead of ssh/sftp). The biggest downside was that each invocation of expect writes to a separate log file. So for instance, if you invoke "make check-gcc" from the top level makefile, this will spawn: check-gcc, check-fortran, check-objc, check-c++, and check-libstdc++-v3 for the default Win64 target. Each make target has its own sum/log file, and make parallelizes at that level. If you were to have make break that up, you would have then an additional set of sum/log files for every check-* make target. If you instead try to have expect break things up or parallelize a single check-* target, you then have to deal with a sum file that is out of order compared to any other run, and a log file at best that's just as out-of-order, and at worst that has all of the output from the various parallel tasks running into each other. Now what you could do (though I don't know how "hackey" it is) would be to run each .exp file separately and run a script to combine the results after the fact and create a new table of results data and new sum/log files. I was working on doing that, and was deciding on awk or perl as the language of choice to do such a thing. We can already specify via RUNTESTFLAGS a single exp file (and actually, a single test in a single exp file) to run. We could just do (pseudocode example) expect file1.exp > out1.txt &; expect file2.exp > out2.txt &; etc... That would then allow multiple invocations of expect that would each write to their own sum/logs. Then at the end, you have to combine them all so as to not be so untidy as to have tons of sum/log sets. Combining them isn't a hugely difficult task. It's just tedious when you aren't that advanced at scripting.
