From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org [79.124.17.100]) by master.gitmailbox.com (Postfix) with ESMTP id 1D9914C600 for ; Tue, 30 Jul 2024 17:50:30 +0000 (UTC) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id B841B68D76A; Tue, 30 Jul 2024 20:50:27 +0300 (EEST) Received: from ursule.remlab.net (vps-a2bccee9.vps.ovh.net [51.75.19.47]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id A6C8368C73D for ; Tue, 30 Jul 2024 20:50:20 +0300 (EEST) Received: from basile.remlab.net (localhost [IPv6:::1]) by ursule.remlab.net (Postfix) with ESMTP id 39DA9C02F9 for ; Tue, 30 Jul 2024 20:50:20 +0300 (EEST) From: =?UTF-8?q?R=C3=A9mi=20Denis-Courmont?= To: ffmpeg-devel@ffmpeg.org Date: Tue, 30 Jul 2024 20:50:20 +0300 Message-ID: <20240730175020.182886-1-remi@remlab.net> X-Mailer: git-send-email 2.45.2 MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH] lavc/h264dsp: stick R-V V weight to 16-bit precision X-BeenThere: ffmpeg-devel@ffmpeg.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: FFmpeg development discussions and patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Reply-To: FFmpeg development discussions and patches Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" Archived-At: List-Archive: List-Post: T-Head C908 (ns): h264_weight2_8_c: 1607.8 h264_weight2_8_rvv_i32: 515.0 (before) h264_weight2_8_rvv_i32: 348.5 (after) h264_weight4_8_c: 2255.8 h264_weight4_8_rvv_i32: 1015.0 (before) h264_weight4_8_rvv_i32: 691.0 (after) h264_weight8_8_c: 3857.5 h264_weight8_8_rvv_i32: 2218.8 (before) h264_weight8_8_rvv_i32: 1561.3 (after) h264_weight16_8_c: 7431.5 h264_weight16_8_rvv_i32: 2737.3 (before) h264_weight16_8_rvv_i32: 1848.3 (after) SpacemiT X60 (ns): h264_weight2_8_c: 1624.1 h264_weight2_8_rvv_i32: 352.6 (before) h264_weight2_8_rvv_i32: 259.3 (after) h264_weight4_8_c: 2259.3 h264_weight4_8_rvv_i32: 685.8 (before) h264_weight4_8_rvv_i32: 530.3 (after) h264_weight8_8_c: 4103.3 h264_weight8_8_rvv_i32: 1581.8 (before) h264_weight8_8_rvv_i32: 1238.6 (after) h264_weight16_8_c: 7624.3 h264_weight16_8_rvv_i32: 2738.1 (before) h264_weight16_8_rvv_i32: 1853.3 (after) --- libavcodec/riscv/h264dsp_rvv.S | 37 +++++++++++++++------------------- 1 file changed, 16 insertions(+), 21 deletions(-) diff --git a/libavcodec/riscv/h264dsp_rvv.S b/libavcodec/riscv/h264dsp_rvv.S index 0e08de43e4..0d641008cb 100644 --- a/libavcodec/riscv/h264dsp_rvv.S +++ b/libavcodec/riscv/h264dsp_rvv.S @@ -32,17 +32,15 @@ func ff_h264_weight_pixels_simple_8_rvv, zve32x csrwi vxrm, 0 sll a5, a5, a3 1: - vsetvli zero, a6, e32, m4, ta, ma + vsetvli zero, a6, e16, m2, ta, ma vle8.v v8, (a0) addi a2, a2, -1 - vmv.v.x v16, a5 - vsetvli zero, zero, e16, m2, ta, ma vzext.vf2 v24, v8 - vwmaccsu.vx v16, a4, v24 - vnclip.wx v16, v16, a3 + vmul.vx v16, v24, a4 + vsadd.vx v16, v16, a5 vmax.vx v16, v16, zero vsetvli zero, zero, e8, m1, ta, ma - vnclipu.wi v8, v16, 0 + vnclipu.wx v8, v16, a3 vse8.v v8, (a0) add a0, a0, a1 bnez a2, 1b @@ -85,23 +83,20 @@ func ff_h264_weight_pixels_8_rvv, zve32x mv t0, a0 mv t6, a6 2: - vsetvli t2, a2, e32, m8, ta, ma + vsetvli t2, a2, e16, m8, ta, ma vlsseg2e8.v v0, (t0), a1 addi t6, t6, -2 - vmv.v.x v16, a5 - vmv.v.x v24, a5 - vsetvli zero, zero, e16, m4, ta, ma - vzext.vf2 v8, v0 - vzext.vf2 v12, v2 - vwmaccsu.vx v16, a4, v8 - vwmaccsu.vx v24, a4, v12 - vnclip.wx v8, v16, a3 - vnclip.wx v12, v24, a3 - vmax.vx v8, v8, zero - vmax.vx v12, v12, zero - vsetvli zero, zero, e8, m2, ta, ma - vnclipu.wi v0, v8, 0 - vnclipu.wi v2, v12, 0 + vzext.vf2 v16, v0 + vzext.vf2 v24, v4 + vmul.vx v16, v16, a4 + vmul.vx v24, v24, a4 + vsadd.vx v16, v16, a5 + vsadd.vx v24, v24, a5 + vmax.vx v16, v16, zero + vmax.vx v24, v24, zero + vsetvli zero, zero, e8, m4, ta, ma + vnclipu.wx v0, v16, a3 + vnclipu.wx v4, v24, a3 vssseg2e8.v v0, (t0), a1 addi t0, t0, 2 bnez t6, 2b -- 2.45.2 _______________________________________________ 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".