From: "H.J. Lu" <hjl.tools@gmail.com>
To: "H.J. Lu via Gcc-patches" <gcc-patches@gcc.gnu.org>,
Richard Biener <richard.guenther@gmail.com>,
"H.J. Lu" <hjl.tools@gmail.com>,
Richard Sandiford <richard.sandiford@arm.com>
Subject: Re: [PATCH v2] DSE: Use the constant store source if possible
Date: Wed, 25 May 2022 11:56:14 -0700 [thread overview]
Message-ID: <CAMe9rOo3c4BgJx8yChVqcyhZz8ix6E2iwYFvGHaOyT5FBHG8pw@mail.gmail.com> (raw)
In-Reply-To: <mpt7d6aqdc2.fsf@arm.com>
On Wed, May 25, 2022 at 12:30 AM Richard Sandiford
<richard.sandiford@arm.com> wrote:
>
> "H.J. Lu via Gcc-patches" <gcc-patches@gcc.gnu.org> writes:
> > On Mon, May 23, 2022 at 12:38:06PM +0200, Richard Biener wrote:
> >> On Sat, May 21, 2022 at 5:02 AM H.J. Lu via Gcc-patches
> >> <gcc-patches@gcc.gnu.org> wrote:
> >> >
> >> > When recording store for RTL dead store elimination, check if the source
> >> > register is set only once to a constant. If yes, record the constant
> >> > as the store source. It eliminates unrolled zero stores after memset 0
> >> > in a loop where a vector register is used as the zero store source.
> >> >
> >> > gcc/
> >> >
> >> > PR rtl-optimization/105638
> >> > * dse.cc (record_store): Use the constant source if the source
> >> > register is set only once.
> >> >
> >> > gcc/testsuite/
> >> >
> >> > PR rtl-optimization/105638
> >> > * g++.target/i386/pr105638.C: New test.
> >> > ---
> >> > gcc/dse.cc | 19 ++++++++++
> >> > gcc/testsuite/g++.target/i386/pr105638.C | 44 ++++++++++++++++++++++++
> >> > 2 files changed, 63 insertions(+)
> >> > create mode 100644 gcc/testsuite/g++.target/i386/pr105638.C
> >> >
> >> > diff --git a/gcc/dse.cc b/gcc/dse.cc
> >> > index 30c11cee034..0433dd3d846 100644
> >> > --- a/gcc/dse.cc
> >> > +++ b/gcc/dse.cc
> >> > @@ -1508,6 +1508,25 @@ record_store (rtx body, bb_info_t bb_info)
> >> >
> >> > if (tem && CONSTANT_P (tem))
> >> > const_rhs = tem;
> >> > + else
> >> > + {
> >> > + /* If RHS is set only once to a constant, set CONST_RHS
> >> > + to the constant. */
> >> > + df_ref def = DF_REG_DEF_CHAIN (REGNO (rhs));
> >> > + if (def != nullptr
> >> > + && !DF_REF_IS_ARTIFICIAL (def)
> >> > + && !DF_REF_NEXT_REG (def))
> >> > + {
> >> > + rtx_insn *def_insn = DF_REF_INSN (def);
> >> > + rtx def_body = PATTERN (def_insn);
> >> > + if (GET_CODE (def_body) == SET)
> >> > + {
> >> > + rtx def_src = SET_SRC (def_body);
> >> > + if (CONSTANT_P (def_src))
> >> > + const_rhs = def_src;
> >>
> >> doesn't DSE have its own tracking of stored values? Shouldn't we
> >
> > It tracks stored values only within the basic block. When RTL loop
> > invariant motion hoists a constant initialization out of the loop into
> > a separate basic block, the constant store value becomes unknown
> > within the original basic block.
> >
> >> improve _that_ if it is not enough? I also wonder if you need to
> >
> > My patch extends DSE stored value tracking to include the constant which
> > is set only once in another basic block.
> >
> >> verify the SET isn't partial?
> >>
> >
> > Here is the v2 patch to check that the constant is set by a non-partial
> > unconditional load.
> >
> > OK for master?
> >
> > Thanks.
> >
> > H.J.
> > ---
> > RTL DSE tracks redundant constant stores within a basic block. When RTL
> > loop invariant motion hoists a constant initialization out of the loop
> > into a separate basic block, the constant store value becomes unknown
> > within the original basic block. When recording store for RTL DSE, check
> > if the source register is set only once to a constant by a non-partial
> > unconditional load. If yes, record the constant as the constant store
> > source. It eliminates unrolled zero stores after memset 0 in a loop
> > where a vector register is used as the zero store source.
> >
> > gcc/
> >
> > PR rtl-optimization/105638
> > * dse.cc (record_store): Use the constant source if the source
> > register is set only once.
> >
> > gcc/testsuite/
> >
> > PR rtl-optimization/105638
> > * g++.target/i386/pr105638.C: New test.
> > ---
> > gcc/dse.cc | 22 ++++++++++++
> > gcc/testsuite/g++.target/i386/pr105638.C | 44 ++++++++++++++++++++++++
> > 2 files changed, 66 insertions(+)
> > create mode 100644 gcc/testsuite/g++.target/i386/pr105638.C
> >
> > diff --git a/gcc/dse.cc b/gcc/dse.cc
> > index 30c11cee034..af8e88dac32 100644
> > --- a/gcc/dse.cc
> > +++ b/gcc/dse.cc
> > @@ -1508,6 +1508,28 @@ record_store (rtx body, bb_info_t bb_info)
> >
> > if (tem && CONSTANT_P (tem))
> > const_rhs = tem;
> > + else
> > + {
> > + /* If RHS is set only once to a constant, set CONST_RHS
> > + to the constant. */
> > + df_ref def = DF_REG_DEF_CHAIN (REGNO (rhs));
> > + if (def != nullptr
> > + && !DF_REF_IS_ARTIFICIAL (def)
> > + && !(DF_REF_FLAGS (def)
> > + & (DF_REF_PARTIAL | DF_REF_CONDITIONAL))
> > + && !DF_REF_NEXT_REG (def))
>
> Can we introduce a helper for this? There are already similar tests
> in ira and loop-iv, and it seems a bit too complex to have to open-code
> each time.
I can use find_single_def_src in loop-iv.cc:
/* If REGNO has a single definition, return its known value, otherwise return
null. */
rtx
find_single_def_src (unsigned int regno)
and do
/* If RHS is set only once to a constant, set CONST_RHS
to the constant. */
rtx def_src = find_single_def_src (REGNO (rhs));
if (def_src != nullptr && CONSTANT_P (def_src))
{
df_ref def = DF_REG_DEF_CHAIN (REGNO (rhs));
if (!(DF_REF_FLAGS (def)
& (DF_REF_PARTIAL | DF_REF_CONDITIONAL)))
{
rtx_insn *def_insn = DF_REF_INSN (def);
rtx def_body = single_set (def_insn);
if (rhs == SET_DEST (def_body))
const_rhs = def_src;
}
}
}
Will it work?
Thanks.
> Thanks,
> Richard
>
> > + {
> > + rtx_insn *def_insn = DF_REF_INSN (def);
> > + rtx def_body = PATTERN (def_insn);
> > + if (GET_CODE (def_body) == SET)
> > + {
> > + rtx def_src = SET_SRC (def_body);
> > + if (CONSTANT_P (def_src)
> > + && GET_MODE (def_src) == GET_MODE (rhs))
> > + const_rhs = def_src;
> > + }
> > + }
> > + }
> > }
> > }
> >
> > diff --git a/gcc/testsuite/g++.target/i386/pr105638.C b/gcc/testsuite/g++.target/i386/pr105638.C
> > new file mode 100644
> > index 00000000000..ff40a459de1
> > --- /dev/null
> > +++ b/gcc/testsuite/g++.target/i386/pr105638.C
> > @@ -0,0 +1,44 @@
> > +/* { dg-do compile { target { ! ia32 } } } */
> > +/* { dg-options "-std=gnu++20 -O2 -march=skylake" } */
> > +/* { dg-final { scan-assembler-not "vpxor" } } */
> > +
> > +#include <stdint.h>
> > +#include <vector>
> > +#include <tr1/array>
> > +
> > +class FastBoard {
> > +public:
> > + typedef std::pair<int, int> movescore_t;
> > + typedef std::tr1::array<movescore_t, 24> scoredlist_t;
> > +
> > +protected:
> > + std::vector<int> m_critical;
> > +
> > + int m_boardsize;
> > +};
> > +
> > +class FastState {
> > +public:
> > + FastBoard board;
> > +
> > + int movenum;
> > +protected:
> > + FastBoard::scoredlist_t scoredmoves;
> > +};
> > +
> > +class KoState : public FastState {
> > +private:
> > + std::vector<uint64_t> ko_hash_history;
> > + std::vector<uint64_t> hash_history;
> > +};
> > +
> > +class GameState : public KoState {
> > +public:
> > + void foo ();
> > +private:
> > + std::vector<KoState> game_history;
> > +};
> > +
> > +void GameState::foo() {
> > + game_history.resize(movenum);
> > +}
--
H.J.
next prev parent reply other threads:[~2022-05-25 18:56 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-05-21 3:01 [PATCH] DSE: Use the constant " H.J. Lu
2022-05-23 10:38 ` Richard Biener
2022-05-23 18:34 ` [PATCH v2] DSE: Use the constant store " H.J. Lu
2022-05-24 6:42 ` Richard Biener
2022-05-24 20:10 ` H.J. Lu
2022-05-25 9:22 ` Richard Biener
2022-05-25 9:30 ` Richard Sandiford
2022-05-25 19:01 ` H.J. Lu
2022-05-25 7:30 ` Richard Sandiford
2022-05-25 18:56 ` H.J. Lu [this message]
2022-05-26 15:14 ` Richard Sandiford
2022-05-26 20:43 ` [PATCH v3] " H.J. Lu
2022-05-28 18:37 ` Jeff Law
2022-05-29 21:43 ` H.J. Lu
2022-05-29 22:55 ` Jeff Law
2022-05-30 8:28 ` Richard Sandiford
2022-05-30 22:58 ` Jeff Law
2022-05-30 8:35 ` Richard Sandiford
2022-05-31 17:12 ` [PATCH v4] " H.J. Lu
2022-06-01 7:20 ` Richard Sandiford
2022-06-01 21:07 ` H.J. Lu
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=CAMe9rOo3c4BgJx8yChVqcyhZz8ix6E2iwYFvGHaOyT5FBHG8pw@mail.gmail.com \
--to=hjl.tools@gmail.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=richard.guenther@gmail.com \
--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).