This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] [aarch64] Fix falkor pipeline description for dup<q>


On 08/02/2018 04:19 PM, Kyrill Tkachov wrote:
Hi Siddhesh,

On 02/08/18 11:23, Siddhesh Poyarekar wrote:
There was a typo in the pipeline description where DUP was assigned to
the vector pipes for quad mode ops when it really only uses the VTOG
pipes.  Fixing this does not show any noticeable difference in
performance (there's a very small bump of 1.7% in x264 but that's
about it) in my tests but is the more precise description of operations
for falkor.


You know the microarchitecture better, so I'm assuming that's correct for Falkor.
I assume this has been tested with a bootstrap and test.

Sorry I should have mentioned this: bootstrapped and tested --with-cpu=falkor and no new regressions resulted from this patch.


        * gcc/config/aarch64/falkor.md (falkor_am_1_vxvy_vxvy): Move
        neon_dup_q to...
        (falkor_am_1_gtov_gtov): ... a new insn reservation.


No gcc/ prefix in the ChangeLog path.

Oops, will fix.

You should be able to use the "*" construct here like this:
falkor_gtov*2

Fixed and tested.

Otherwise looks ok to me (but you'll need a maintainer to approve)

Thanks, I'll send the updated patch for James' review.

Siddhesh


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]