I tested it, and indeed using vwsub is faster. Updated it in the reply.
---
I have a question: if I tweak the load order a bit, using one less vset, it
leads to being slower (the patch I submitted is 13.2, if I make the
following change, the time would be 15.2).
But I thought it would be faster.
- vsetvli t0, a2, e8, m2, tu, ma
- vle8.v v0, (a0)
- sub a2, a2, t0
- vsetvli zero, t0, e16, m4, tu, ma
- vle16.v v8, (a1)
- vsetvli zero, t0, e8, m2, tu, ma
- vwsub.wv v16, v8, v0
+ vsetvli t0, a2, e16, m4, tu, ma
+ vle16.v v8, (a1)
+ sub a2, a2, t0
+ vsetvli zero, t0, e8, m2, tu, ma
+ vle8.v v0, (a0)
+ vwsub.wv v16, v8, v0
Rémi Denis-Courmont <r...@remlab.net> 于2024年1月6日周六 23:05写道：
> Le perjantaina 5. tammikuuta 2024, 2.56.18 EET flow gg a écrit :
> > One vset can be reduced, but vwsub should not be used in this case. I
> > modified it in this reply.
>
> Fair enough, but are you sure that that's faster than keeping the vsetvli
> and
> removing the sign extension?
>
> > Rémi Denis-Courmont <r...@remlab.net> 于2024年1月5日周五 00:00写道：
> >
> > > Le lauantaina 30. joulukuuta 2023, 18.20.15 EET flow gg a écrit :
> > > > I mistook it, seeing the vector length as the length of the vector
> > >
> > > register
> > >
> > > > ..
> > > > I have modified it in this reply.
> > >
> > > Setting element size to 8-bit is unnecessary, and a widening
> subtraction
> > > can
> > > presumably avoid the sign extension.
> > >
> > > --
> > > レミ・デニ-クールモン
> > > http://www.remlab.net/
> > >
> > >
> > >
> > > _______________________________________________
> > > ffmpeg-devel mailing list
> > > ffmpeg-devel@ffmpeg.org
> > > https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
> > >
> > > To unsubscribe, visit link above, or email
> > > ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
>
>
> --
> Rémi Denis-Courmont
> http://www.remlab.net/
>
>
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
>
From e028a42bf7a63eb7f0851d03dbca471bbfecbc7b Mon Sep 17 00:00:00 2001
From: sunyuechi <sunyue...@iscas.ac.cn>
Date: Fri, 29 Dec 2023 13:27:31 +0800
Subject: [PATCH 3/3] lavc/svq1enc: R-V V ssd_int8_vs_int16
C908
ssd_int8_vs_int16_c: 207.7
ssd_int8_vs_int16_rvv_i32: 13.2
---
libavcodec/riscv/Makefile | 2 ++
libavcodec/riscv/svqenc_init.c | 41 +++++++++++++++++++++++++++++++
libavcodec/riscv/svqenc_rvv.S | 45 ++++++++++++++++++++++++++++++++++
libavcodec/svq1enc.c | 2 ++
libavcodec/svq1encdsp.h | 1 +
5 files changed, 91 insertions(+)
create mode 100644 libavcodec/riscv/svqenc_init.c
create mode 100644 libavcodec/riscv/svqenc_rvv.S
diff --git a/libavcodec/riscv/Makefile b/libavcodec/riscv/Makefile
index 7f253bba12..4e14c3d094 100644
--- a/libavcodec/riscv/Makefile
+++ b/libavcodec/riscv/Makefile
@@ -46,6 +46,8 @@ RVV-OBJS-$(CONFIG_OPUS_DECODER) += riscv/opusdsp_rvv.o
OBJS-$(CONFIG_PIXBLOCKDSP) += riscv/pixblockdsp_init.o
RV-OBJS-$(CONFIG_PIXBLOCKDSP) += riscv/pixblockdsp_rvi.o
RVV-OBJS-$(CONFIG_PIXBLOCKDSP) += riscv/pixblockdsp_rvv.o
+OBJS-$(CONFIG_SVQ1_ENCODER) += riscv/svqenc_init.o
+RVV-OBJS-$(CONFIG_SVQ1_ENCODER) += riscv/svqenc_rvv.o
OBJS-$(CONFIG_TAK_DECODER) += riscv/takdsp_init.o
RVV-OBJS-$(CONFIG_TAK_DECODER) += riscv/takdsp_rvv.o
OBJS-$(CONFIG_UTVIDEO_DECODER) += riscv/utvideodsp_init.o
diff --git a/libavcodec/riscv/svqenc_init.c b/libavcodec/riscv/svqenc_init.c
new file mode 100644
index 0000000000..f4c398960c
--- /dev/null
+++ b/libavcodec/riscv/svqenc_init.c
@@ -0,0 +1,41 @@
+/*
+ * Copyright (c) 2023 Institue of Software Chinese Academy of Sciences (ISCAS).
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "config.h"
+
+#include "libavutil/attributes.h"
+#include "libavutil/cpu.h"
+#include "libavcodec/svq1encdsp.h"
+
+int ff_ssd_int8_vs_int16_rvv(const int8_t *pix1, const int16_t *pix2,
+ intptr_t size);
+
+av_cold void ff_svq1enc_init_riscv(SVQ1EncDSPContext *c)
+{
+#if HAVE_RVV
+ int flags = av_get_cpu_flags();
+
+ if (flags & AV_CPU_FLAG_RVV_I32) {
+ if (flags & AV_CPU_FLAG_RVB_ADDR) {
+ c->ssd_int8_vs_int16 = ff_ssd_int8_vs_int16_rvv;
+ }
+ }
+#endif
+}
diff --git a/libavcodec/riscv/svqenc_rvv.S b/libavcodec/riscv/svqenc_rvv.S
new file mode 100644
index 0000000000..d5985cafa6
--- /dev/null
+++ b/libavcodec/riscv/svqenc_rvv.S
@@ -0,0 +1,45 @@
+/*
+ * Copyright (c) 2023 Institue of Software Chinese Academy of Sciences (ISCAS).
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavutil/riscv/asm.S"
+
+func ff_ssd_int8_vs_int16_rvv, zve32x
+ vsetvli t0, zero, e32, m8, ta, ma
+ vmv.v.x v24, zero
+1:
+ vsetvli t0, a2, e8, m2, tu, ma
+ vle8.v v0, (a0)
+ sub a2, a2, t0
+ vsetvli zero, t0, e16, m4, tu, ma
+ vle16.v v8, (a1)
+ vsetvli zero, t0, e8, m2, tu, ma
+ vwsub.wv v16, v8, v0
+ vsetvli zero, t0, e16, m4, tu, ma
+ add a0, a0, t0
+ vwmacc.vv v24, v16, v16
+ sh1add a1, t0, a1
+ bnez a2, 1b
+ vsetvli t0, zero, e32, m8, ta, ma
+ vmv.s.x v0, zero
+ vredsum.vs v0, v24, v0
+ vmv.x.s a0, v0
+
+ ret
+endfunc
diff --git a/libavcodec/svq1enc.c b/libavcodec/svq1enc.c
index 0dea405dec..6e7ea12aa7 100644
--- a/libavcodec/svq1enc.c
+++ b/libavcodec/svq1enc.c
@@ -766,6 +766,8 @@ void ff_svq1enc_init(SVQ1EncDSPContext *c)
#if ARCH_PPC
ff_svq1enc_init_ppc(c);
+#elif ARCH_RISCV
+ ff_svq1enc_init_riscv(c);
#elif ARCH_X86
ff_svq1enc_init_x86(c);
#endif
diff --git a/libavcodec/svq1encdsp.h b/libavcodec/svq1encdsp.h
index 618bf8463b..5dfa35cc62 100644
--- a/libavcodec/svq1encdsp.h
+++ b/libavcodec/svq1encdsp.h
@@ -30,6 +30,7 @@ typedef struct SVQ1EncDSPContext {
void ff_svq1enc_init(SVQ1EncDSPContext *c);
void ff_svq1enc_init_ppc(SVQ1EncDSPContext *c);
+void ff_svq1enc_init_riscv(SVQ1EncDSPContext *c);
void ff_svq1enc_init_x86(SVQ1EncDSPContext *c);
#endif /* AVCODEC_SVQ1ENCDSP_H */
--
2.43.0
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
