Brian,
Upstream closed this bug report as invalid:
"That is because the early complete unrolling comes and unrolls the
loop so the autovectorizer does not have a loop to work on anymore.
If I increase it to be 16 instead of 4, the loop is vectorizer.
So the original testcase is invalid as two th
forwarded 512050 http://gcc.gnu.org/PR38899
thanks
* brian m. carlson [2009-01-17 17:13]:
> I believe that it's the maintainers job to file the bug upstream, and
OK, I've verified that 4.3 and 4.4 from SVN produce the same code.
--
Martin Michlmayr
http://www.cyrius.com/
--
To UNSUBSCRIBE,
Processing commands for cont...@bugs.debian.org:
> forwarded 512050 http://gcc.gnu.org/PR38899
Bug#512050: gcc-4.3: pessimizes function without SSE intrinsics
Noted your statement that Bug has been forwarded to http://gcc.gnu.org/PR38899.
> thanks
Stopping processing here.
Please contact
On Fri, Jan 16, 2009 at 08:10:15PM +0100, Martin Michlmayr wrote:
* brian m. carlson [2009-01-16 18:38]:
Obviously, since the two functions do the exact same thing, they should
be optimized to be identical. Instead, mul is pessimized.
Can you check if this happens with gcc-4.3 and trunk from
* brian m. carlson [2009-01-16 18:38]:
> Obviously, since the two functions do the exact same thing, they should
> be optimized to be identical. Instead, mul is pessimized.
Can you check if this happens with gcc-4.3 and trunk from SVN, and if
so, file the bug upstream?
--
Martin Michlmayr
http
5 matches
Mail list logo