Hi,
This week I looked into DENBench:
* sad8_c (hot function from mp4encode) needs SLP reduction, but it
also contains cond_expr which cannot be vectorized as reduction, so I
don't think there is anything I can do here
* fdct_int32 (another hot function from mp4encode) now gets vectorized
with vzi
Hello,
* Continue looking into DENbench benchmarks.
* While testing SMS I realized that my current implementation of doloop
pattern for ARM does not follow SMS's requirement to have the doloop
instructions be decoupled from the other loop's instructions. This happens
because doloop uses CC regist
Hi there. I've created a new mailing list for the automated build
results of Linaro GCC, Linaro GDB, and (once-weekly) FSF 4.5 and 4.6.
You can subscribe by going to:
https://launchpad.net/~linaro-toolchain-builds
I've subscribed everyone in the toolchain working group. If you'd
like to get th
