From: Andrew Stubbs <ams@codesourcery.com>
To: Jeff Law <jlaw@ventanamicro.com>,
"gcc-patches@gcc.gnu.org" <gcc-patches@gcc.gnu.org>
Cc: Jivan Hakobyan <jivanhakobyan9@gmail.com>
Subject: Re: [RFA] New pass for sign/zero extension elimination
Date: Mon, 27 Nov 2023 11:30:58 +0000 [thread overview]
Message-ID: <2e63f6e7-6a8e-4112-9cdc-aa14fa7c2824@codesourcery.com> (raw)
In-Reply-To: <6d5f8ba7-0c60-4789-87ae-68617ce6ac2c@ventanamicro.com>
I tried this patch for AMD GCN. We have a similar problem with excess
extends, but also for vector modes. Each lane has a minimum 32 bits and
GCC's normal assumption is that vector registers have precisely the
number of bits they need, so the amdgcn backend patterns have explicit
sign/zero extends for QImode and HImode for the instructions that might
need it. It would be cool if this pass could eliminate some of those,
but at this point I just wanted to check it didn't break anything.
Unfortunately I get a crash building libgcc:
> during RTL pass: ext_dce
> conftest.c: In function 'main':
> conftest.c:16:1: internal compiler error: RTL check: expected code 'subreg', have 'reg' in ext_dce_process_uses, at ext-dce.cc:421
> 16 | }
> | ^
> 0x8c7aa3 rtl_check_failed_code1(rtx_def const*, rtx_code, char const*, int, char const*)
> /scratch/astubbs/omp/upA/gcnbuild/src/gcc-mainline/gcc/rtl.cc:770
> 0xa76a27 ext_dce_process_uses
> /scratch/astubbs/omp/upA/gcnbuild/src/gcc-mainline/gcc/ext-dce.cc:421
> 0x1aeca5c ext_dce_process_bb
> /scratch/astubbs/omp/upA/gcnbuild/src/gcc-mainline/gcc/ext-dce.cc:651
> 0x1aeca5c ext_dce
> /scratch/astubbs/omp/upA/gcnbuild/src/gcc-mainline/gcc/ext-dce.cc:802
> 0x1aeca5c execute
> /scratch/astubbs/omp/upA/gcnbuild/src/gcc-mainline/gcc/ext-dce.cc:868
> Please submit a full bug report, with preprocessed source (by using -freport-bug).
> Please include the complete backtrace with any bug report.
> See <https://gcc.gnu.org/bugs/> for instructions.
> configure:3812: $? = 1
> configure: failed program was:
> | /* confdefs.h */
> | #define PACKAGE_NAME "GNU C Runtime Library"
> | #define PACKAGE_TARNAME "libgcc"
> | #define PACKAGE_VERSION "1.0"
> | #define PACKAGE_STRING "GNU C Runtime Library 1.0"
> | #define PACKAGE_BUGREPORT ""
> | #define PACKAGE_URL "http://www.gnu.org/software/libgcc/"
> | /* end confdefs.h. */
> |
> | int
> | main ()
> | {
> |
> | ;
> | return 0;
> | }
I have no idea if this is an unhandled case or a case that shouldn't
exist, but it's trying to do "SUBREG_BYTE (dst).is_constant ()" for a
very simple instruction:
(set (reg/i:SI 168 v8)
(const_int 0 [0]))
This seems pretty basic to me, but there is some hidden complexity. It's
possible that the pass has correctly identified that "v8" can hold more
that just a single integer: in this case we're using a single lane of a
vector register. No extend is needed here though. The register has 2048
bits, but only 32 are active in SImode.
Andrew
next prev parent reply other threads:[~2023-11-27 11:31 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-11-20 0:47 Jeff Law
2023-11-20 1:22 ` Oleg Endo
2023-11-20 2:51 ` Jeff Law
2023-11-20 2:57 ` Oleg Endo
2023-11-20 2:23 ` Xi Ruoyao
2023-11-20 2:46 ` Jeff Law
2023-11-20 2:52 ` Jeff Law
2023-11-20 3:32 ` Xi Ruoyao
2023-11-20 3:48 ` Jeff Law
2023-11-20 18:26 ` Richard Sandiford
2023-11-22 17:59 ` Jeff Law
2023-11-27 20:15 ` Richard Sandiford
2023-11-20 18:56 ` Dimitar Dimitrov
2023-11-22 22:23 ` Jeff Law
2023-11-26 16:42 ` rep.dot.nop
2023-11-27 16:14 ` Jeff Law
2023-11-27 11:30 ` Andrew Stubbs [this message]
2023-11-27 16:16 ` Jeff Law
2023-12-01 1:08 ` Hans-Peter Nilsson
2023-12-01 15:09 ` Jeff Law
2023-12-01 16:17 ` Hans-Peter Nilsson
2023-11-27 17:36 Joern Rennecke
2023-11-27 17:57 ` Joern Rennecke
2023-11-27 20:03 ` Richard Sandiford
2023-11-27 20:18 ` Jeff Law
2023-11-28 13:36 ` Joern Rennecke
2023-11-28 14:09 ` Joern Rennecke
2023-11-30 17:33 ` Jeff Law
2023-11-28 13:13 ` Joern Rennecke
2023-11-28 5:50 ` Jeff Law
2023-11-27 18:19 Joern Rennecke
2023-11-28 5:51 ` Jeff Law
2023-11-29 17:37 Joern Rennecke
2023-11-29 19:13 ` Jivan Hakobyan
2023-11-30 15:37 ` 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=2e63f6e7-6a8e-4112-9cdc-aa14fa7c2824@codesourcery.com \
--to=ams@codesourcery.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=jivanhakobyan9@gmail.com \
--cc=jlaw@ventanamicro.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).