I noticed that *by far* the bulk of the changes (in terms of line count) were in the simout files for the switcheroo tests, which print the cycle number on every switch, and thus are sensitive to small changes in timing. These files are very large in general... in fact, I had to temporarily disable the 400K 'large file limit' on the server to be able to push these changes. So it would be really nice if someone would tweak the switcheroo tests to not be so verbose by default, then update the reference outputs with the (much smaller) resulting simout files.
Thanks, Steve On Sat, Sep 28, 2013 at 12:27 PM, Steve Reinhardt <[email protected]> wrote: > changeset afd9ea6101d9 in /z/repo/gem5 > details: http://repo.gem5.org/gem5?cmd=changeset;node=afd9ea6101d9 > description: > tests: update reference outputs > > Apparently only stats.txt was updated the last time, so > this changeset updates other reference output files > (config.ini, simout, simerr, ruby.stats) so that > test output diffs should not be cluttered with irrelevant > changes. There are a few stats.txt updates too, but > they are in the minority. > > diffstat: > > tests/long/fs/10.linux-boot/ref/alpha/linux/tsunami-switcheroo-full/simout > | 10752 ++-- > tests/long/fs/10.linux-boot/ref/arm/linux/realview-switcheroo-full/simout > | 8198 +- > tests/long/fs/10.linux-boot/ref/arm/linux/realview-switcheroo-o3/simout > | 5212 +- > tests/long/fs/10.linux-boot/ref/arm/linux/realview-switcheroo-timing/simout > | 21275 ++++----- > tests/long/fs/10.linux-boot/ref/x86/linux/pc-switcheroo-full/simout > | 17144 ++++++++- > tests/quick/fs/10.linux-boot/ref/arm/linux/realview-switcheroo-atomic/simout > | 18629 ++++---- > _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
