I am resending this message as the previous one had one wrong response
email address "gcc-pat...@gcc.gnu.org"
-------- Forwarded Message --------
Subject: Re: [PATCH V2 0/4] Add DF_LIVE_SUBREG data and apply to IRA
and LRA
Date: Wed, 1 May 2024 08:35:27 -0400
From: Vladimir Makarov <vmaka...@redhat.com>
To: Lehua Ding <lehua.d...@rivai.ai>, gcc-pat...@gcc.gnu.org,
richard.sandif...@arm.com
CC: juzhe.zh...@rivai.ai, shuo.c...@rivai.ai, jin....@rivai.ai
On 4/24/24 06:01, Lehua Ding wrote:
Hi Vladimir and Richard,
These patches are used to add a new data flow DF_LIVE_SUBREG,
which will track subreg liveness and then apply it to IRA and LRA
passes (enabled via -O3 or -ftrack-subreg-liveness). These patches
are for GCC 15. And these codes are pushed to the devel/subreg-coalesce
branch. In addition, my colleague Shuo Chen will also be involved in some
of the remain work, thank you for your support.
Thank you for creation of the branch. It helped me a lot.
These patches are separated from the subreg-coalesce patches submitted
a few months ago. I refactored the code according to comments. The next
patches will support subreg coalesce base on they. Here are some data
abot build time of SPEC INT 2017 (x86-64 target):
Thank you for refactoring patches.
baseline baseline(+track-subreg-liveness)
specint2017 build time : 1892s 1883s
Interesting and surprisingly unexpected improvement.
Regarding build times, I've run it a few times, but they all seem to take
much less time. Since the difference is small, it's possible that it's
just
a change in environment. But it's theoretically possible, since supporting
subreg-liveness could have reduced the number of living regs.
For memory usage, I trided PR 69609 by valgrind, peak memory size grow
from
2003910656 to 2003947520, very small increase.
I'll soon finish code review of IRA and LRA changes and send it today or
tomorrow.
But In brief I have no objections to the patches, just some minor
requests to improve them.
