This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [cxx-mem-model] testing infrastructure
- From: Mike Stump <mikestump at comcast dot net>
- To: Aldy Hernandez <aldyh at redhat dot com>
- Cc: gcc-patches <gcc-patches at gcc dot gnu dot org>, Jeff Law <law at redhat dot com>, Tom Tromey <tromey at redhat dot com>, Benjamin De Kosnik <bkoz at redhat dot com>, Andrew MacLeod <amacleod at redhat dot com>
- Date: Thu, 10 Mar 2011 13:09:33 -0800
- Subject: Re: [cxx-mem-model] testing infrastructure
- References: <4D792D6E.4050803@redhat.com>
On Mar 10, 2011, at 11:58 AM, Aldy Hernandez wrote:
> The following is a test harness for testing atomicity problems and data races in multi-threaded environments.
The idea of developing a methodology for testing corner cases I think is interesting and neat. volatile would be another area that one could use this sort of testing for, if they want. As time goes on, I suspect we'll have an ever increasing need to handle multicore issues, so, I think expanding in this area is worthwhile.
There is a speed hit and the requirement to have a target (or simulator) for testing. If one just checks generated assembly, one avoids needing a target, and avoids the speed hit, but at the cost of having to come up with assembles to match for the target, which is annoying. I can imagine validating sequences with this sort of test suite and then for future testing, checking the cache for a sequence that is known to be ok, to speed testing. I'll just note that this sort of speed hack is orthogonal and could be used for a large number of execute test cases, not just this class of testing.