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 1805A48245 for ; Tue, 14 Nov 2023 19:06:09 +0000 (UTC) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 7A81F68CD3A; Tue, 14 Nov 2023 21:06:07 +0200 (EET) Received: from ursule.remlab.net (vps-a2bccee9.vps.ovh.net [51.75.19.47]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 58B1468CC06 for ; Tue, 14 Nov 2023 21:06:01 +0200 (EET) Received: from basile.remlab.net (localhost [IPv6:::1]) by ursule.remlab.net (Postfix) with ESMTP id D9E33C00A2 for ; Tue, 14 Nov 2023 21:06:00 +0200 (EET) From: =?UTF-8?q?R=C3=A9mi=20Denis-Courmont?= To: ffmpeg-devel@ffmpeg.org Date: Tue, 14 Nov 2023 21:06:00 +0200 Message-ID: <20231114190600.10084-1-remi@remlab.net> X-Mailer: git-send-email 2.42.0 MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH] lavc/flacdsp: R-V V decorrelate_indep 32-bit 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: flac_decorrelate_indep2_32_c: 981.7 flac_decorrelate_indep2_32_rvv_i32: 183.7 flac_decorrelate_indep4_32_c: 1749.7 flac_decorrelate_indep4_32_rvv_i32: 362.5 flac_decorrelate_indep6_32_c: 2517.7 flac_decorrelate_indep6_32_rvv_i32: 715.2 flac_decorrelate_indep8_32_c: 3285.7 flac_decorrelate_indep8_32_rvv_i32: 909.0 --- libavcodec/riscv/flacdsp_init.c | 22 ++++++ libavcodec/riscv/flacdsp_rvv.S | 132 ++++++++++++++++++++++++++++++++ 2 files changed, 154 insertions(+) diff --git a/libavcodec/riscv/flacdsp_init.c b/libavcodec/riscv/flacdsp_init.c index 0e7be25d98..5bfaad7ca8 100644 --- a/libavcodec/riscv/flacdsp_init.c +++ b/libavcodec/riscv/flacdsp_init.c @@ -30,6 +30,14 @@ void ff_flac_decorrelate_rs_16_rvv(uint8_t **out, int32_t **in, int channels, int len, int shift); void ff_flac_decorrelate_ms_16_rvv(uint8_t **out, int32_t **in, int channels, int len, int shift); +void ff_flac_decorrelate_indep2_32_rvv(uint8_t **out, int32_t **in, + int channels, int len, int shift); +void ff_flac_decorrelate_indep4_32_rvv(uint8_t **out, int32_t **in, + int channels, int len, int shift); +void ff_flac_decorrelate_indep6_32_rvv(uint8_t **out, int32_t **in, + int channels, int len, int shift); +void ff_flac_decorrelate_indep8_32_rvv(uint8_t **out, int32_t **in, + int channels, int len, int shift); void ff_flac_decorrelate_ls_32_rvv(uint8_t **out, int32_t **in, int channels, int len, int shift); void ff_flac_decorrelate_rs_32_rvv(uint8_t **out, int32_t **in, @@ -51,6 +59,20 @@ av_cold void ff_flacdsp_init_riscv(FLACDSPContext *c, enum AVSampleFormat fmt, c->decorrelate[3] = ff_flac_decorrelate_ms_16_rvv; break; case AV_SAMPLE_FMT_S32: + switch (channels) { + case 2: + c->decorrelate[0] = ff_flac_decorrelate_indep2_32_rvv; + break; + case 4: + c->decorrelate[0] = ff_flac_decorrelate_indep4_32_rvv; + break; + case 6: + c->decorrelate[0] = ff_flac_decorrelate_indep6_32_rvv; + break; + case 8: + c->decorrelate[0] = ff_flac_decorrelate_indep8_32_rvv; + break; + } c->decorrelate[1] = ff_flac_decorrelate_ls_32_rvv; c->decorrelate[2] = ff_flac_decorrelate_rs_32_rvv; c->decorrelate[3] = ff_flac_decorrelate_ms_32_rvv; diff --git a/libavcodec/riscv/flacdsp_rvv.S b/libavcodec/riscv/flacdsp_rvv.S index 616565ed7e..5a0607e044 100644 --- a/libavcodec/riscv/flacdsp_rvv.S +++ b/libavcodec/riscv/flacdsp_rvv.S @@ -95,6 +95,138 @@ func ff_flac_decorrelate_ms_16_rvv, zve32x ret endfunc +func ff_flac_decorrelate_indep2_32_rvv, zve32x + ld a0, (a0) + ld a2, 8(a1) + ld a1, (a1) +1: + vsetvli t0, a3, e32, m4, ta, ma + vle32.v v0, (a1) + sub a3, a3, t0 + vle32.v v4, (a2) + sh2add a1, t0, a1 + vsll.vx v0, v0, a4 + sh2add a2, t0, a2 + vsll.vx v4, v4, a4 + vsseg2e32.v v0, (a0) + sh3add a0, t0, a0 + bnez a3, 1b + + ret +endfunc + +func ff_flac_decorrelate_indep4_32_rvv, zve32x + ld a0, (a0) + ld a2, 8(a1) + ld t1, 16(a1) + ld t2, 24(a1) + ld a1, (a1) +1: + vsetvli t0, a3, e32, m2, ta, ma + vle32.v v0, (a1) + sub a3, a3, t0 + vle32.v v2, (a2) + sh2add a1, t0, a1 + vsll.vx v0, v0, a4 + sh2add a2, t0, a2 + vle32.v v4, (t1) + sh2add t1, t0, t1 + vsll.vx v2, v2, a4 + vle32.v v6, (t2) + sh2add t2, t0, t2 + vsll.vx v4, v4, a4 + slli t0, t0, 4 + vsll.vx v6, v6, a4 + vsseg4e32.v v0, (a0) + add a0, t0, a0 + bnez a3, 1b + + ret +endfunc + +func ff_flac_decorrelate_indep6_32_rvv, zve32x + ld a0, (a0) + ld a2, 8(a1) + ld t1, 16(a1) + ld t2, 24(a1) + ld t3, 32(a1) + ld t4, 40(a1) + ld a1, (a1) +1: + vsetvli t0, a3, e32, m1, ta, ma + vle32.v v0, (a1) + sub a3, a3, t0 + vle32.v v1, (a2) + sh2add a1, t0, a1 + vsll.vx v0, v0, a4 + sh2add a2, t0, a2 + vle32.v v2, (t1) + sh2add t1, t0, t1 + vsll.vx v1, v1, a4 + vle32.v v3, (t2) + sh2add t2, t0, t2 + vsll.vx v2, v2, a4 + vle32.v v4, (t3) + sh2add t3, t0, t3 + vsll.vx v3, v3, a4 + vle32.v v5, (t4) + sh2add t4, t0, t4 + vsll.vx v4, v4, a4 + slli t0, t0, 3 + vsll.vx v5, v5, a4 + sh1add t0, t0, t0 // t0 *= 3 + vsseg6e32.v v0, (a0) + add a0, t0, a0 + bnez a3, 1b + + ret +endfunc + +func ff_flac_decorrelate_indep8_32_rvv, zve32x + ld a0, (a0) + ld a2, 8(a1) + ld t1, 16(a1) + ld t2, 24(a1) + ld t3, 32(a1) + ld t4, 40(a1) + ld t5, 48(a1) + ld t6, 56(a1) + ld a1, (a1) +1: + vsetvli t0, a3, e32, m1, ta, ma + vle32.v v0, (a1) + sub a3, a3, t0 + vle32.v v1, (a2) + sh2add a1, t0, a1 + vsll.vx v0, v0, a4 + vle32.v v2, (t1) + sh2add a2, t0, a2 + vsll.vx v1, v1, a4 + sh2add t1, t0, t1 + vle32.v v3, (t2) + vsll.vx v2, v2, a4 + sh2add t2, t0, t2 + vle32.v v4, (t3) + sh2add t3, t0, t3 + vsll.vx v3, v3, a4 + vle32.v v5, (t4) + sh2add t4, t0, t4 + vsll.vx v4, v4, a4 + vle32.v v6, (t5) + sh2add t5, t0, t5 + vsll.vx v5, v5, a4 + vle32.v v7, (t6) + sh2add t6, t0, t6 + vsll.vx v6, v6, a4 + slli t0, t0, 5 + vsll.vx v7, v7, a4 + vsseg8e32.v v0, (a0) + add a0, t0, a0 + bnez a3, 1b + + ret +endfunc + func ff_flac_decorrelate_ls_32_rvv, zve32x ld a0, (a0) ld a2, 8(a1) -- 2.42.0 _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".