This is the mail archive of the
mailing list for the GCC project.
Re: [lno] fix mem-tag problem in vectorizer, and test-cases updates
- From: Mark Mitchell <mark at codesourcery dot com>
- To: Dorit Naishlos <DORIT at il dot ibm dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Tue, 22 Jun 2004 19:52:50 -0700
- Subject: Re: [lno] fix mem-tag problem in vectorizer, and test-cases updates
- Organization: CodeSourcery, LLC
- References: <OF12E4D71E.46879A09-ONC2256EBB.00608F8F-C2256EBB.00619E94@il.ibm.com>
Dorit Naishlos wrote:
Fixed the problem raised in http://gcc.gnu.org/ml/gcc/2004-06/msg01023.html
and http://gcc.gnu.org/ml/gcc/2004-06/msg00961.html. Test-cases 40,42 and 46
test the fix (the call to bar was removed from the vectorized function).
Other updates to the test-cases include some cleanups,
and new test-cases to test different combinations of
known/unknown misalignment of stores/loads,
with known/unknown loop-bound.
At some point, I'd suggest you move the tests from
tree-ssa-vect/tree-ssa-vect-n.c to vect/vect-n.c. The tree-ssa prefix
isn't necessary because these are tests for functionality (particular
optimizations), not tests of tree-ssa per se.
And I'm glad to see that there are all these tests!