* [PATCH v1] x86: Remove SSE3 instruction for broadcast in memset.S (SSE2 Only)
@ 2022-02-07 6:38 Noah Goldstein
2022-02-07 12:54 ` H.J. Lu
2022-02-07 19:39 ` [PATCH v2] x86: Remove SSSE3 " Noah Goldstein
0 siblings, 2 replies; 8+ messages in thread
From: Noah Goldstein @ 2022-02-07 6:38 UTC (permalink / raw)
To: libc-alpha
commit b62ace2740a106222e124cc86956448fa07abf4d
Author: Noah Goldstein <goldstein.w.n@gmail.com>
Date: Sun Feb 6 00:54:18 2022 -0600
x86: Improve vec generation in memset-vec-unaligned-erms.S
Revert usage of 'pshufb' in broadcast logic as it is an SSE3
instruction and memset.S is restricted to only SSE2 instructions.
---
sysdeps/x86_64/memset.S | 19 ++++++++++---------
1 file changed, 10 insertions(+), 9 deletions(-)
diff --git a/sysdeps/x86_64/memset.S b/sysdeps/x86_64/memset.S
index ccf036be53..148553cf3d 100644
--- a/sysdeps/x86_64/memset.S
+++ b/sysdeps/x86_64/memset.S
@@ -28,22 +28,23 @@
#define VMOVU movups
#define VMOVA movaps
-# define MEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
+#define MEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
movd d, %xmm0; \
- pxor %xmm1, %xmm1; \
- pshufb %xmm1, %xmm0; \
- movq r, %rax
+ movq r, %rax; \
+ punpcklbw %xmm0, %xmm0; \
+ punpcklwd %xmm0, %xmm0; \
+ pshufd $0, %xmm0, %xmm0
-# define WMEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
+#define WMEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
movd d, %xmm0; \
pshufd $0, %xmm0, %xmm0; \
movq r, %rax
-# define MEMSET_VDUP_TO_VEC0_HIGH()
-# define MEMSET_VDUP_TO_VEC0_LOW()
+#define MEMSET_VDUP_TO_VEC0_HIGH()
+#define MEMSET_VDUP_TO_VEC0_LOW()
-# define WMEMSET_VDUP_TO_VEC0_HIGH()
-# define WMEMSET_VDUP_TO_VEC0_LOW()
+#define WMEMSET_VDUP_TO_VEC0_HIGH()
+#define WMEMSET_VDUP_TO_VEC0_LOW()
#define SECTION(p) p
--
2.25.1
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v1] x86: Remove SSE3 instruction for broadcast in memset.S (SSE2 Only)
2022-02-07 6:38 [PATCH v1] x86: Remove SSE3 instruction for broadcast in memset.S (SSE2 Only) Noah Goldstein
@ 2022-02-07 12:54 ` H.J. Lu
2022-02-07 14:32 ` H.J. Lu
2022-02-07 19:39 ` [PATCH v2] x86: Remove SSSE3 " Noah Goldstein
1 sibling, 1 reply; 8+ messages in thread
From: H.J. Lu @ 2022-02-07 12:54 UTC (permalink / raw)
To: Noah Goldstein; +Cc: GNU C Library, Carlos O'Donell
On Sun, Feb 6, 2022 at 10:39 PM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
>
> commit b62ace2740a106222e124cc86956448fa07abf4d
> Author: Noah Goldstein <goldstein.w.n@gmail.com>
> Date: Sun Feb 6 00:54:18 2022 -0600
>
> x86: Improve vec generation in memset-vec-unaligned-erms.S
>
> Revert usage of 'pshufb' in broadcast logic as it is an SSE3
> instruction and memset.S is restricted to only SSE2 instructions.
> ---
> sysdeps/x86_64/memset.S | 19 ++++++++++---------
> 1 file changed, 10 insertions(+), 9 deletions(-)
>
> diff --git a/sysdeps/x86_64/memset.S b/sysdeps/x86_64/memset.S
> index ccf036be53..148553cf3d 100644
> --- a/sysdeps/x86_64/memset.S
> +++ b/sysdeps/x86_64/memset.S
> @@ -28,22 +28,23 @@
> #define VMOVU movups
> #define VMOVA movaps
>
> -# define MEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> +#define MEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> movd d, %xmm0; \
> - pxor %xmm1, %xmm1; \
> - pshufb %xmm1, %xmm0; \
> - movq r, %rax
> + movq r, %rax; \
> + punpcklbw %xmm0, %xmm0; \
> + punpcklwd %xmm0, %xmm0; \
> + pshufd $0, %xmm0, %xmm0
>
> -# define WMEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> +#define WMEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> movd d, %xmm0; \
> pshufd $0, %xmm0, %xmm0; \
> movq r, %rax
>
> -# define MEMSET_VDUP_TO_VEC0_HIGH()
> -# define MEMSET_VDUP_TO_VEC0_LOW()
> +#define MEMSET_VDUP_TO_VEC0_HIGH()
> +#define MEMSET_VDUP_TO_VEC0_LOW()
What are these changes for?
> -# define WMEMSET_VDUP_TO_VEC0_HIGH()
> -# define WMEMSET_VDUP_TO_VEC0_LOW()
> +#define WMEMSET_VDUP_TO_VEC0_HIGH()
> +#define WMEMSET_VDUP_TO_VEC0_LOW()
What are these changes for?
> #define SECTION(p) p
>
> --
> 2.25.1
>
--
H.J.
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v1] x86: Remove SSE3 instruction for broadcast in memset.S (SSE2 Only)
2022-02-07 12:54 ` H.J. Lu
@ 2022-02-07 14:32 ` H.J. Lu
2022-02-07 19:40 ` Noah Goldstein
0 siblings, 1 reply; 8+ messages in thread
From: H.J. Lu @ 2022-02-07 14:32 UTC (permalink / raw)
To: Noah Goldstein; +Cc: GNU C Library, Carlos O'Donell
On Mon, Feb 7, 2022 at 4:54 AM H.J. Lu <hjl.tools@gmail.com> wrote:
>
> On Sun, Feb 6, 2022 at 10:39 PM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
> >
> > commit b62ace2740a106222e124cc86956448fa07abf4d
> > Author: Noah Goldstein <goldstein.w.n@gmail.com>
> > Date: Sun Feb 6 00:54:18 2022 -0600
> >
> > x86: Improve vec generation in memset-vec-unaligned-erms.S
> >
> > Revert usage of 'pshufb' in broadcast logic as it is an SSE3
pshufb is an SSSE3, not SSE3, instruction.
> > instruction and memset.S is restricted to only SSE2 instructions.
> > ---
> > sysdeps/x86_64/memset.S | 19 ++++++++++---------
> > 1 file changed, 10 insertions(+), 9 deletions(-)
> >
> > diff --git a/sysdeps/x86_64/memset.S b/sysdeps/x86_64/memset.S
> > index ccf036be53..148553cf3d 100644
> > --- a/sysdeps/x86_64/memset.S
> > +++ b/sysdeps/x86_64/memset.S
> > @@ -28,22 +28,23 @@
> > #define VMOVU movups
> > #define VMOVA movaps
> >
> > -# define MEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> > +#define MEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> > movd d, %xmm0; \
> > - pxor %xmm1, %xmm1; \
> > - pshufb %xmm1, %xmm0; \
> > - movq r, %rax
> > + movq r, %rax; \
> > + punpcklbw %xmm0, %xmm0; \
> > + punpcklwd %xmm0, %xmm0; \
> > + pshufd $0, %xmm0, %xmm0
> >
> > -# define WMEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> > +#define WMEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> > movd d, %xmm0; \
> > pshufd $0, %xmm0, %xmm0; \
> > movq r, %rax
> >
> > -# define MEMSET_VDUP_TO_VEC0_HIGH()
> > -# define MEMSET_VDUP_TO_VEC0_LOW()
> > +#define MEMSET_VDUP_TO_VEC0_HIGH()
> > +#define MEMSET_VDUP_TO_VEC0_LOW()
>
> What are these changes for?
>
> > -# define WMEMSET_VDUP_TO_VEC0_HIGH()
> > -# define WMEMSET_VDUP_TO_VEC0_LOW()
> > +#define WMEMSET_VDUP_TO_VEC0_HIGH()
> > +#define WMEMSET_VDUP_TO_VEC0_LOW()
>
> What are these changes for?
>
> > #define SECTION(p) p
> >
> > --
> > 2.25.1
> >
>
>
> --
> H.J.
--
H.J.
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v1] x86: Remove SSE3 instruction for broadcast in memset.S (SSE2 Only)
2022-02-07 14:32 ` H.J. Lu
@ 2022-02-07 19:40 ` Noah Goldstein
0 siblings, 0 replies; 8+ messages in thread
From: Noah Goldstein @ 2022-02-07 19:40 UTC (permalink / raw)
To: H.J. Lu; +Cc: GNU C Library, Carlos O'Donell
On Mon, Feb 7, 2022 at 8:33 AM H.J. Lu <hjl.tools@gmail.com> wrote:
>
> On Mon, Feb 7, 2022 at 4:54 AM H.J. Lu <hjl.tools@gmail.com> wrote:
> >
> > On Sun, Feb 6, 2022 at 10:39 PM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
> > >
> > > commit b62ace2740a106222e124cc86956448fa07abf4d
> > > Author: Noah Goldstein <goldstein.w.n@gmail.com>
> > > Date: Sun Feb 6 00:54:18 2022 -0600
> > >
> > > x86: Improve vec generation in memset-vec-unaligned-erms.S
> > >
> > > Revert usage of 'pshufb' in broadcast logic as it is an SSE3
>
> pshufb is an SSSE3, not SSE3, instruction.
Fixed. The commit message is different but V2 is up.
>
> > > instruction and memset.S is restricted to only SSE2 instructions.
> > > ---
> > > sysdeps/x86_64/memset.S | 19 ++++++++++---------
> > > 1 file changed, 10 insertions(+), 9 deletions(-)
> > >
> > > diff --git a/sysdeps/x86_64/memset.S b/sysdeps/x86_64/memset.S
> > > index ccf036be53..148553cf3d 100644
> > > --- a/sysdeps/x86_64/memset.S
> > > +++ b/sysdeps/x86_64/memset.S
> > > @@ -28,22 +28,23 @@
> > > #define VMOVU movups
> > > #define VMOVA movaps
> > >
> > > -# define MEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> > > +#define MEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> > > movd d, %xmm0; \
> > > - pxor %xmm1, %xmm1; \
> > > - pshufb %xmm1, %xmm0; \
> > > - movq r, %rax
> > > + movq r, %rax; \
> > > + punpcklbw %xmm0, %xmm0; \
> > > + punpcklwd %xmm0, %xmm0; \
> > > + pshufd $0, %xmm0, %xmm0
> > >
> > > -# define WMEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> > > +#define WMEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> > > movd d, %xmm0; \
> > > pshufd $0, %xmm0, %xmm0; \
> > > movq r, %rax
> > >
> > > -# define MEMSET_VDUP_TO_VEC0_HIGH()
> > > -# define MEMSET_VDUP_TO_VEC0_LOW()
> > > +#define MEMSET_VDUP_TO_VEC0_HIGH()
> > > +#define MEMSET_VDUP_TO_VEC0_LOW()
> >
> > What are these changes for?
> >
> > > -# define WMEMSET_VDUP_TO_VEC0_HIGH()
> > > -# define WMEMSET_VDUP_TO_VEC0_LOW()
> > > +#define WMEMSET_VDUP_TO_VEC0_HIGH()
> > > +#define WMEMSET_VDUP_TO_VEC0_LOW()
> >
> > What are these changes for?
Undid them in V2. Realized I had misindented them in my
last commit.
> >
> > > #define SECTION(p) p
> > >
> > > --
> > > 2.25.1
> > >
> >
> >
> > --
> > H.J.
>
>
>
> --
> H.J.
^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH v2] x86: Remove SSSE3 instruction for broadcast in memset.S (SSE2 Only)
2022-02-07 6:38 [PATCH v1] x86: Remove SSE3 instruction for broadcast in memset.S (SSE2 Only) Noah Goldstein
2022-02-07 12:54 ` H.J. Lu
@ 2022-02-07 19:39 ` Noah Goldstein
2022-02-07 19:49 ` H.J. Lu
1 sibling, 1 reply; 8+ messages in thread
From: Noah Goldstein @ 2022-02-07 19:39 UTC (permalink / raw)
To: libc-alpha
commit b62ace2740a106222e124cc86956448fa07abf4d
Author: Noah Goldstein <goldstein.w.n@gmail.com>
Date: Sun Feb 6 00:54:18 2022 -0600
x86: Improve vec generation in memset-vec-unaligned-erms.S
Revert usage of 'pshufb' in broadcast logic as it is an SSSE3
instruction and memset.S is restricted to only SSE2 instructions.
---
sysdeps/x86_64/memset.S | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/sysdeps/x86_64/memset.S b/sysdeps/x86_64/memset.S
index ccf036be53..3f0517bbfc 100644
--- a/sysdeps/x86_64/memset.S
+++ b/sysdeps/x86_64/memset.S
@@ -30,9 +30,10 @@
# define MEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
movd d, %xmm0; \
- pxor %xmm1, %xmm1; \
- pshufb %xmm1, %xmm0; \
- movq r, %rax
+ movq r, %rax; \
+ punpcklbw %xmm0, %xmm0; \
+ punpcklwd %xmm0, %xmm0; \
+ pshufd $0, %xmm0, %xmm0
# define WMEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
movd d, %xmm0; \
--
2.25.1
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v2] x86: Remove SSSE3 instruction for broadcast in memset.S (SSE2 Only)
2022-02-07 19:39 ` [PATCH v2] x86: Remove SSSE3 " Noah Goldstein
@ 2022-02-07 19:49 ` H.J. Lu
2022-02-07 20:56 ` Noah Goldstein
0 siblings, 1 reply; 8+ messages in thread
From: H.J. Lu @ 2022-02-07 19:49 UTC (permalink / raw)
To: Noah Goldstein; +Cc: GNU C Library, Carlos O'Donell
On Mon, Feb 7, 2022 at 11:39 AM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
>
> commit b62ace2740a106222e124cc86956448fa07abf4d
> Author: Noah Goldstein <goldstein.w.n@gmail.com>
> Date: Sun Feb 6 00:54:18 2022 -0600
>
> x86: Improve vec generation in memset-vec-unaligned-erms.S
>
> Revert usage of 'pshufb' in broadcast logic as it is an SSSE3
> instruction and memset.S is restricted to only SSE2 instructions.
> ---
> sysdeps/x86_64/memset.S | 7 ++++---
> 1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/sysdeps/x86_64/memset.S b/sysdeps/x86_64/memset.S
> index ccf036be53..3f0517bbfc 100644
> --- a/sysdeps/x86_64/memset.S
> +++ b/sysdeps/x86_64/memset.S
> @@ -30,9 +30,10 @@
>
> # define MEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> movd d, %xmm0; \
> - pxor %xmm1, %xmm1; \
> - pshufb %xmm1, %xmm0; \
> - movq r, %rax
> + movq r, %rax; \
> + punpcklbw %xmm0, %xmm0; \
> + punpcklwd %xmm0, %xmm0; \
> + pshufd $0, %xmm0, %xmm0
>
> # define WMEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> movd d, %xmm0; \
> --
> 2.25.1
>
LGTM.
Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
Thanks.
--
H.J.
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v2] x86: Remove SSSE3 instruction for broadcast in memset.S (SSE2 Only)
2022-02-07 19:49 ` H.J. Lu
@ 2022-02-07 20:56 ` Noah Goldstein
2022-05-04 5:48 ` Sunil Pandey
0 siblings, 1 reply; 8+ messages in thread
From: Noah Goldstein @ 2022-02-07 20:56 UTC (permalink / raw)
To: H.J. Lu; +Cc: GNU C Library, Carlos O'Donell
On Mon, Feb 7, 2022 at 1:49 PM H.J. Lu <hjl.tools@gmail.com> wrote:
>
> On Mon, Feb 7, 2022 at 11:39 AM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
> >
> > commit b62ace2740a106222e124cc86956448fa07abf4d
> > Author: Noah Goldstein <goldstein.w.n@gmail.com>
> > Date: Sun Feb 6 00:54:18 2022 -0600
> >
> > x86: Improve vec generation in memset-vec-unaligned-erms.S
> >
> > Revert usage of 'pshufb' in broadcast logic as it is an SSSE3
> > instruction and memset.S is restricted to only SSE2 instructions.
> > ---
> > sysdeps/x86_64/memset.S | 7 ++++---
> > 1 file changed, 4 insertions(+), 3 deletions(-)
> >
> > diff --git a/sysdeps/x86_64/memset.S b/sysdeps/x86_64/memset.S
> > index ccf036be53..3f0517bbfc 100644
> > --- a/sysdeps/x86_64/memset.S
> > +++ b/sysdeps/x86_64/memset.S
> > @@ -30,9 +30,10 @@
> >
> > # define MEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> > movd d, %xmm0; \
> > - pxor %xmm1, %xmm1; \
> > - pshufb %xmm1, %xmm0; \
> > - movq r, %rax
> > + movq r, %rax; \
> > + punpcklbw %xmm0, %xmm0; \
> > + punpcklwd %xmm0, %xmm0; \
> > + pshufd $0, %xmm0, %xmm0
> >
> > # define WMEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> > movd d, %xmm0; \
> > --
> > 2.25.1
> >
>
> LGTM.
>
> Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
>
> Thanks.
Thanks pushed.
>
> --
> H.J.
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v2] x86: Remove SSSE3 instruction for broadcast in memset.S (SSE2 Only)
2022-02-07 20:56 ` Noah Goldstein
@ 2022-05-04 5:48 ` Sunil Pandey
0 siblings, 0 replies; 8+ messages in thread
From: Sunil Pandey @ 2022-05-04 5:48 UTC (permalink / raw)
To: Noah Goldstein, Libc-stable Mailing List; +Cc: H.J. Lu, GNU C Library
On Mon, Feb 7, 2022 at 12:56 PM Noah Goldstein via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
> On Mon, Feb 7, 2022 at 1:49 PM H.J. Lu <hjl.tools@gmail.com> wrote:
> >
> > On Mon, Feb 7, 2022 at 11:39 AM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
> > >
> > > commit b62ace2740a106222e124cc86956448fa07abf4d
> > > Author: Noah Goldstein <goldstein.w.n@gmail.com>
> > > Date: Sun Feb 6 00:54:18 2022 -0600
> > >
> > > x86: Improve vec generation in memset-vec-unaligned-erms.S
> > >
> > > Revert usage of 'pshufb' in broadcast logic as it is an SSSE3
> > > instruction and memset.S is restricted to only SSE2 instructions.
> > > ---
> > > sysdeps/x86_64/memset.S | 7 ++++---
> > > 1 file changed, 4 insertions(+), 3 deletions(-)
> > >
> > > diff --git a/sysdeps/x86_64/memset.S b/sysdeps/x86_64/memset.S
> > > index ccf036be53..3f0517bbfc 100644
> > > --- a/sysdeps/x86_64/memset.S
> > > +++ b/sysdeps/x86_64/memset.S
> > > @@ -30,9 +30,10 @@
> > >
> > > # define MEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> > > movd d, %xmm0; \
> > > - pxor %xmm1, %xmm1; \
> > > - pshufb %xmm1, %xmm0; \
> > > - movq r, %rax
> > > + movq r, %rax; \
> > > + punpcklbw %xmm0, %xmm0; \
> > > + punpcklwd %xmm0, %xmm0; \
> > > + pshufd $0, %xmm0, %xmm0
> > >
> > > # define WMEMSET_SET_VEC0_AND_SET_RETURN(d, r) \
> > > movd d, %xmm0; \
> > > --
> > > 2.25.1
> > >
> >
> > LGTM.
> >
> > Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
> >
> > Thanks.
>
> Thanks pushed.
> >
> > --
> > H.J.
I would like to backport this patch to release branches.
Any comments or objections?
--Sunil
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2022-05-04 5:48 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-02-07 6:38 [PATCH v1] x86: Remove SSE3 instruction for broadcast in memset.S (SSE2 Only) Noah Goldstein
2022-02-07 12:54 ` H.J. Lu
2022-02-07 14:32 ` H.J. Lu
2022-02-07 19:40 ` Noah Goldstein
2022-02-07 19:39 ` [PATCH v2] x86: Remove SSSE3 " Noah Goldstein
2022-02-07 19:49 ` H.J. Lu
2022-02-07 20:56 ` Noah Goldstein
2022-05-04 5:48 ` Sunil Pandey
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).