Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
 help / color / mirror / Atom feed
From: flow gg <hlefthleft@gmail.com>
To: FFmpeg development discussions and patches <ffmpeg-devel@ffmpeg.org>
Subject: Re: [FFmpeg-devel] [PATCH v3 5/9] lavc/vp9dsp: R-V V mc avg
Date: Sat, 18 May 2024 00:06:13 +0800
Message-ID: <CAEa-L+uPPeU9MdsP=aib0xa-+DvVc2s_=_hd=aXmREamnk3eEw@mail.gmail.com> (raw)
In-Reply-To: <9276774.vF19fDgCsf@basile.remlab.net>

yeah, updated it in the reply

Rémi Denis-Courmont <remi@remlab.net> 于2024年5月17日周五 23:11写道:

> Le maanantaina 13. toukokuuta 2024, 19.59.22 EEST uk7b@foxmail.com a
> écrit :
> > From: sunyuechi <sunyuechi@iscas.ac.cn>
> >
> > C908:
> > vp9_avg4_8bpp_c: 1.2
> > vp9_avg4_8bpp_rvv_i64: 1.0
> > vp9_avg8_8bpp_c: 3.7
> > vp9_avg8_8bpp_rvv_i64: 1.5
> > vp9_avg16_8bpp_c: 14.7
> > vp9_avg16_8bpp_rvv_i64: 3.5
> > vp9_avg32_8bpp_c: 57.7
> > vp9_avg32_8bpp_rvv_i64: 10.0
> > vp9_avg64_8bpp_c: 229.0
> > vp9_avg64_8bpp_rvv_i64: 31.7
> > ---
> >  libavcodec/riscv/Makefile      |  3 +-
> >  libavcodec/riscv/vp9_mc_rvv.S  | 58 ++++++++++++++++++++++++++++++++++
> >  libavcodec/riscv/vp9dsp_init.c | 18 +++++++++++
> >  3 files changed, 78 insertions(+), 1 deletion(-)
> >  create mode 100644 libavcodec/riscv/vp9_mc_rvv.S
> >
> > diff --git a/libavcodec/riscv/Makefile b/libavcodec/riscv/Makefile
> > index 0cd900104f..1183357b37 100644
> > --- a/libavcodec/riscv/Makefile
> > +++ b/libavcodec/riscv/Makefile
> > @@ -64,6 +64,7 @@ RVV-OBJS-$(CONFIG_VP8DSP) += riscv/vp8dsp_rvv.o
> >  OBJS-$(CONFIG_VP9_DECODER) += riscv/vp9dsp_init.o
> >  RV-OBJS-$(CONFIG_VP9_DECODER) += riscv/vp9_intra_rvi.o \
> >                                   riscv/vp9_mc_rvi.o
> > -RVV-OBJS-$(CONFIG_VP9_DECODER) += riscv/vp9_intra_rvv.o
> > +RVV-OBJS-$(CONFIG_VP9_DECODER) += riscv/vp9_intra_rvv.o \
> > +                                  riscv/vp9_mc_rvv.o
> >  OBJS-$(CONFIG_VORBIS_DECODER) += riscv/vorbisdsp_init.o
> >  RVV-OBJS-$(CONFIG_VORBIS_DECODER) += riscv/vorbisdsp_rvv.o
> > diff --git a/libavcodec/riscv/vp9_mc_rvv.S
> b/libavcodec/riscv/vp9_mc_rvv.S
> > new file mode 100644
> > index 0000000000..5d917e7b98
> > --- /dev/null
> > +++ b/libavcodec/riscv/vp9_mc_rvv.S
> > @@ -0,0 +1,58 @@
> > +/*
> > + * Copyright (c) 2024 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"
> > +
> > +.macro vsetvlstatic8 len an maxlen mn=m4
> > +.if \len == 4
> > +        vsetivli        zero, \len, e8, mf4, ta, ma
> > +.elseif \len == 8
> > +        vsetivli        zero, \len, e8, mf2, ta, ma
> > +.elseif \len == 16
> > +        vsetivli        zero, \len, e8, m1, ta, ma
> > +.elseif \len == 32
> > +        li              \an, \len
> > +        vsetvli         zero, \an, e8, m2, ta, ma
> > +.elseif \len == 64
> > +        li              \an, \maxlen
> > +        vsetvli         zero, \an, e8, \mn, ta, ma
> > +.endif
> > +.endm
> > +
> > +.macro copy_avg len
> > +func ff_avg\len\()_rvv, zve32x
> > +        csrwi           vxrm, 0
> > +        vsetvlstatic8   \len t0 64
> > +1:
> > +        addi            a4, a4, -1
> > +        vle8.v          v8, (a2)
> > +        vle8.v          v16, (a0)
> > +        vaaddu.vv       v8, v8, v16
> > +        vse8.v          v8, (a0)
> > +        add             a2, a2, a3
> > +        add             a0, a0, a1
> > +        bnez            a4, 1b
> > +        ret
>
> Doesn't this get slightly faster by interleaving scalar and vector
> instructions?
>
> > +endfunc
> > +.endm
> > +
> > +.irp len 64, 32, 16, 8, 4
> > +        copy_avg \len
> > +.endr
> > diff --git a/libavcodec/riscv/vp9dsp_init.c
> b/libavcodec/riscv/vp9dsp_init.c
> > index 184fadbaf7..1922484a1d 100644
> > --- a/libavcodec/riscv/vp9dsp_init.c
> > +++ b/libavcodec/riscv/vp9dsp_init.c
> > @@ -48,6 +48,24 @@ static av_cold void vp9dsp_mc_init_riscv(VP9DSPContext
> > *dsp, int bpp) }
> >  # endif
> >
> > +#if HAVE_RVV
> > +    if (bpp == 8 && (flags & AV_CPU_FLAG_RVV_I32) &&
> ff_rv_vlen_least(128))
> > { +
> > +#define init_fpel(idx1, sz)                                           \
> > +    dsp->mc[idx1][FILTER_8TAP_SMOOTH ][1][0][0] = ff_avg##sz##_rvv;  \
> > +    dsp->mc[idx1][FILTER_8TAP_REGULAR][1][0][0] = ff_avg##sz##_rvv;  \
> > +    dsp->mc[idx1][FILTER_8TAP_SHARP  ][1][0][0] = ff_avg##sz##_rvv;  \
> > +    dsp->mc[idx1][FILTER_BILINEAR    ][1][0][0] = ff_avg##sz##_rvv
> > +
> > +    init_fpel(0, 64);
> > +    init_fpel(1, 32);
> > +    init_fpel(2, 16);
> > +    init_fpel(3, 8);
> > +    init_fpel(4, 4);
> > +
> > +#undef init_fpel
> > +    }
> > +#endif
> >  #endif
> >  }
>
>
> --
> 雷米‧德尼-库尔蒙
> 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-request@ffmpeg.org with subject "unsubscribe".
>
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

  parent reply	other threads:[~2024-05-17 16:06 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20240513165926.1467967-1-uk7b@foxmail.com>
2024-05-13 16:59 ` [FFmpeg-devel] [PATCH v3 2/9] lavc/vp9dsp: R-V mc copy uk7b
2024-05-13 19:55   ` Rémi Denis-Courmont
2024-05-14  4:44     ` flow gg
2024-05-14 17:02       ` Rémi Denis-Courmont
2024-05-14 17:20         ` flow gg
2024-05-13 16:59 ` [FFmpeg-devel] [PATCH v3 3/9] lavc/vp9dsp: R-V V ipred hor uk7b
2024-05-13 16:59 ` [FFmpeg-devel] [PATCH v3 4/9] lavc/vp9dsp: R-V V ipred tm uk7b
2024-05-14 17:45   ` Rémi Denis-Courmont
2024-05-14 17:57     ` flow gg
2024-05-14 18:08       ` Rémi Denis-Courmont
2024-05-14 18:33         ` uk7b
2024-05-14 18:34         ` flow gg
2024-05-15  3:55           ` [FFmpeg-devel] [PATCH " uk7b
2024-05-15  3:56             ` flow gg
2024-05-13 16:59 ` [FFmpeg-devel] [PATCH v3 5/9] lavc/vp9dsp: R-V V mc avg uk7b
2024-05-17 15:11   ` Rémi Denis-Courmont
2024-05-17 16:05     ` [FFmpeg-devel] [PATCH 1/5] " uk7b
2024-05-17 16:06     ` flow gg [this message]
2024-05-13 16:59 ` [FFmpeg-devel] [PATCH v3 6/9] lavc/vp9dsp: R-V V mc bilin h v uk7b
2024-05-18 15:56   ` Rémi Denis-Courmont
2024-05-18 18:18     ` flow gg
2024-05-13 16:59 ` [FFmpeg-devel] [PATCH v3 7/9] lavc/vp9dsp: R-V V mc tap " uk7b
2024-05-13 16:59 ` [FFmpeg-devel] [PATCH v3 8/9] lavc/vp9dsp: R-V V mc bilin hv uk7b
2024-05-13 16:59 ` [FFmpeg-devel] [PATCH v3 9/9] lavc/vp9dsp: R-V V mc tap hv uk7b
     [not found] <20240512100331.995415-1-uk7b@foxmail.com>
2024-05-12 10:03 ` [FFmpeg-devel] [PATCH v3 5/9] lavc/vp9dsp: R-V V mc avg uk7b

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='CAEa-L+uPPeU9MdsP=aib0xa-+DvVc2s_=_hd=aXmREamnk3eEw@mail.gmail.com' \
    --to=hlefthleft@gmail.com \
    --cc=ffmpeg-devel@ffmpeg.org \
    /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

Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

This inbox may be cloned and mirrored by anyone:

	git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \
		ffmpegdev@gitmailbox.com
	public-inbox-index ffmpegdev

Example config snippet for mirrors.


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git