From: Jeff Law <law@redhat.com>
To: gcc-patches@gcc.gnu.org, richard.sandiford@arm.com
Subject: Re: [PATCH 23/23] fwprop: Rewrite to use RTL SSA
Date: Tue, 15 Dec 2020 20:52:16 -0700 [thread overview]
Message-ID: <85ed5e60-b4a1-a531-de5a-1088196a0580@redhat.com> (raw)
In-Reply-To: <mptima964v8.fsf@arm.com>
On 11/13/20 1:24 AM, Richard Sandiford via Gcc-patches wrote:
> This patch rewrites fwprop.c to use the RTL SSA framework. It tries
> as far as possible to mimic the old behaviour, even in caes where
> that doesn't fit naturally with the new framework. I've added ???
> comments to mark those places, but I think “fixing” them should
> be done separately to make bisection easier.
>
> In particular:
>
> * The old implementation iterated over uses, and after a successful
> substitution, the new insn's uses were added to the end of the list.
> The pass still processed those uses, but because it processed them at
> the end, it didn't fully optimise one instruction before propagating
> it into the next.
>
> The new version follows the same approach for comparison purposes,
> but I'd like to drop that as a follow-on patch.
>
> * The old implementation operated on single use sites (DF_REF_LOCs).
> This doesn't work well for instructions with match_dups, where it's
> necessary to update both an operand and its dups at the same time.
> For example, attempting to substitute into a divmod instruction would
> fail because only the div or the mod side would be updated.
>
> The new version again follows this to some extent for comparison
> purposes (although not exactly). Again I'd like to drop it as a
> follow-on patch.
>
> One difference is that if a register occurs in multiple MEM addresses
> in a set, the new version will try to update them all at once. This is
> what causes the SVE ACLE st4* output to improve.
>
> Also, the old version didn't naturally guarantee termination (PR79405),
> whereas the new one does.
>
> gcc/
> * fwprop.c: Rewrite to use the RTL SSA framework.
>
> gcc/testsuite/
> * gcc.dg/rtl/x86_64/test-return-const.c.before-fwprop.c: Don't
> expect insn updates to be deferred.
> * gcc.target/aarch64/sve/acle/asm/st4_s8.c: Expect the addition
> to be folded into the address.
> * gcc.target/aarch64/sve/acle/asm/st4_s8.c: Likewise.
Consider killing the ADD_NOTES bits.
s/eqaul/equal/ to fix a typo.
Naturally I'm happy at how much by-hand RTL analysis code just
disappears with this change :-)
Ideally you'll drop this in tomorrow and we can get a fresh run of all
the targets in my tester before the weekend. I won't be stressed if we
see some fallout, but I don't expect much. I'll help track them down if
they occur.
Thanks for your patience.
Jeff
next prev parent reply other threads:[~2020-12-16 3:52 UTC|newest]
Thread overview: 88+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-11-13 8:10 [00/23] Make fwprop use an on-the-side RTL SSA representation Richard Sandiford
2020-11-13 8:11 ` [01/23] vec: Silence clang warning Richard Sandiford
2020-11-25 19:58 ` Jeff Law
2020-11-13 8:12 ` [02/23] rtlanal: Remove noop_move_p REG_EQUAL condition Richard Sandiford
2020-11-25 20:00 ` Jeff Law
2020-11-13 8:12 ` [03/23] reginfo: Add a global_reg_set Richard Sandiford
2020-11-25 20:01 ` Jeff Law
2020-11-13 8:13 ` [04/23] Move iterator_range to a new iterator-utils.h file Richard Sandiford
2020-11-25 20:02 ` Jeff Law
2020-11-13 8:13 ` [05/23] Add more iterator utilities Richard Sandiford
2020-11-25 20:12 ` Jeff Law
2020-11-13 8:14 ` [06/23] Add an RAII class for managing obstacks Richard Sandiford
2020-11-25 20:15 ` Jeff Law
2020-11-13 8:14 ` [07/23] Add a class that multiplexes two pointer types Richard Sandiford
2020-11-25 20:23 ` Jeff Law
2020-11-26 16:15 ` Richard Sandiford
2020-11-30 1:28 ` Jeff Law
2020-11-25 23:33 ` Martin Sebor
2020-11-26 17:06 ` Richard Sandiford
2020-11-27 18:12 ` Richard Sandiford
2020-11-28 0:17 ` Martin Sebor
2020-12-17 0:17 ` Richard Sandiford
2020-12-17 14:21 ` Tom Tromey
2020-12-17 15:38 ` Richard Sandiford
2020-12-17 15:44 ` Nathan Sidwell
2021-01-04 15:32 ` Jeff Law
2020-11-13 8:15 ` [08/23] Add an alternative splay tree implementation Richard Sandiford
2020-12-02 20:36 ` Jeff Law
2020-12-17 0:29 ` Richard Sandiford
2021-01-04 15:27 ` Jeff Law
2021-01-01 8:25 ` Andreas Schwab
2021-01-04 14:53 ` Richard Sandiford
2021-01-04 15:02 ` Andreas Schwab
2021-01-04 15:42 ` Richard Sandiford
2021-01-05 12:13 ` Richard Biener
2020-11-13 8:15 ` [09/23] Add a cut-down version of std::span (array_slice) Richard Sandiford
2020-11-30 19:56 ` Jeff Law
2022-08-03 15:13 ` Martin Jambor
2022-08-03 15:31 ` Richard Sandiford
2022-08-10 16:03 ` Martin Jambor
2022-08-11 6:58 ` Richard Biener
2022-08-16 7:59 ` Richard Sandiford
2020-11-13 8:16 ` [10/23] Tweak the way that is_a is implemented Richard Sandiford
2020-12-02 5:15 ` Jeff Law
2020-11-13 8:16 ` [11/23] Split update_cfg_for_uncondjump out of combine Richard Sandiford
2020-11-30 6:14 ` Jeff Law
2020-11-13 8:17 ` [12/23] Export print-rtl.c:print_insn_with_notes Richard Sandiford
2020-11-25 20:24 ` Jeff Law
2020-11-13 8:18 ` [13/23] recog: Split out a register_asm_p function Richard Sandiford
2020-11-25 20:24 ` Jeff Law
2020-11-13 8:18 ` [14/23] simplify-rtx: Put simplify routines into a class Richard Sandiford
2020-11-30 19:54 ` Jeff Law
2020-11-13 8:19 ` [15/23] recog: Add a validate_change_xveclen function Richard Sandiford
2020-11-30 20:03 ` Jeff Law
2020-11-13 8:19 ` [16/23] recog: Add a way of temporarily undoing changes Richard Sandiford
2020-11-25 20:27 ` Jeff Law
2020-12-17 0:22 ` Richard Sandiford
2020-11-13 8:20 ` [17/23] recog: Add a class for propagating into insns Richard Sandiford
2020-12-03 22:32 ` Jeff Law
2020-11-13 8:20 ` [18/23] recog: Add an RAII class for undoing insn changes Richard Sandiford
2020-11-25 20:27 ` Jeff Law
2020-11-13 8:20 ` [19/23] rtlanal: Add some new helper classes Richard Sandiford
2020-12-13 17:30 ` Jeff Law
2020-12-14 16:37 ` Richard Sandiford
2020-12-14 20:02 ` Jeff Law
2020-11-13 8:21 ` [20/23] rtlanal: Add simple_regno_set Richard Sandiford
2020-11-25 20:31 ` Jeff Law
2020-12-17 0:47 ` Richard Sandiford
2021-01-04 15:28 ` Jeff Law
2020-11-13 8:22 ` [21/23] doc: Add documentation for rtl-ssa Richard Sandiford
2020-11-30 6:26 ` Jeff Law
2020-11-13 8:23 ` [PATCH 22/23] Add rtl-ssa Richard Sandiford
2020-12-16 3:31 ` Jeff Law
2020-12-17 0:33 ` Richard Sandiford
2020-12-19 20:01 ` Jeff Law
2020-11-13 8:24 ` [PATCH 23/23] fwprop: Rewrite to use RTL SSA Richard Sandiford
2020-12-16 3:52 ` Jeff Law [this message]
2020-12-17 0:34 ` Richard Sandiford
2020-11-25 19:58 ` [00/23] Make fwprop use an on-the-side RTL SSA representation Jeff Law
2020-11-26 16:03 ` Richard Sandiford
2020-11-27 15:56 ` Michael Matz
2020-11-27 16:31 ` Richard Sandiford
2020-11-30 21:13 ` Jeff Law
2020-12-01 0:03 ` Michael Matz
2020-12-01 10:15 ` Richard Sandiford
2020-12-02 0:25 ` Jeff Law
2020-11-30 6:45 ` Jeff Law
2020-11-30 14:12 ` Richard Sandiford
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=85ed5e60-b4a1-a531-de5a-1088196a0580@redhat.com \
--to=law@redhat.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=richard.sandiford@arm.com \
/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).