From: Paul Koning <paulkoning@comcast.net>
To: Hans-Peter Nilsson <hp@axis.com>
Cc: Roger Sayle <roger@nextmovesoftware.com>,
jeffreyalaw@gmail.com, gcc-patches@gcc.gnu.org,
segher@kernel.crashing.org
Subject: Re: [committed] Convert xstormy16 to LRA
Date: Thu, 11 May 2023 12:55:26 -0400 [thread overview]
Message-ID: <367738E7-2E0E-4406-8BBC-6DC70EA491DD@comcast.net> (raw)
In-Reply-To: <20230511150540.9606F20420@pchp3.se.axis.com>
> On May 11, 2023, at 11:05 AM, Hans-Peter Nilsson via Gcc-patches <gcc-patches@gcc.gnu.org> wrote:
>
> ...
> Yes, very interesting. Thank you for sharing this. I've
> seen regressions with LRA for CRIS too, for
> "double-register-sized" types, which for CRIS, a 32-bit
> target, translates to 64-bit types (DFmode and DImode), and
> where LRA does a much worse job than reload; spills a lot
> more often to stack, even after trying every
> register-allocation-related hook I found (and also an LRA
> patch which helped only by a fraction, but regressed results
> on x86_64-linux, so let's quickly forget it again).
That observation makes me a bit worried. While CRIS may not be a priority platform, that description makes it sound like a case that would be significant in any 32 bit platform, which would include priority ones like i386 and ARM.
If that's true, I wonder about dropping Reload. While I understand it's been years since LRA was first introduced, wouldn't we even so want to go by the rule that a newer replacement mechanism doesn't replace an older one until the replacement demonstrates comparable or better output compared with the older one?
paul
next prev parent reply other threads:[~2023-05-11 16:55 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-05-01 23:37 Roger Sayle
2023-05-02 12:40 ` Paul Koning
2023-05-02 13:18 ` Roger Sayle
2023-05-02 13:49 ` Segher Boessenkool
2023-05-02 16:20 ` Roger Sayle
2023-05-02 21:55 ` Segher Boessenkool
2023-05-02 14:11 ` Paul Koning
2023-05-02 15:47 ` Segher Boessenkool
2023-05-02 15:57 ` Jeff Law
2023-05-11 15:05 ` Hans-Peter Nilsson
2023-05-11 16:55 ` Paul Koning [this message]
2023-05-11 18:15 ` Jeff Law
2023-05-11 18:32 ` Hans-Peter Nilsson
2023-05-12 13:53 ` Hans-Peter Nilsson
2023-05-12 14:01 ` Hans-Peter Nilsson
2023-05-12 14:04 ` Roger Sayle
2023-05-13 0:56 ` Hans-Peter Nilsson
2023-05-13 1:11 ` Hans-Peter Nilsson
-- strict thread matches above, loose matches on Subject: below --
2023-05-01 13:42 Jeff Law
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=367738E7-2E0E-4406-8BBC-6DC70EA491DD@comcast.net \
--to=paulkoning@comcast.net \
--cc=gcc-patches@gcc.gnu.org \
--cc=hp@axis.com \
--cc=jeffreyalaw@gmail.com \
--cc=roger@nextmovesoftware.com \
--cc=segher@kernel.crashing.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).