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 ESMTPS id F3DAF490B2 for ; Tue, 28 Jan 2025 15:55:09 +0000 (UTC) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id EF43E68B84B; Tue, 28 Jan 2025 17:55:05 +0200 (EET) Received: from mail-pl1-f180.google.com (mail-pl1-f180.google.com [209.85.214.180]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id 72A5F68AD70 for ; Tue, 28 Jan 2025 17:54:59 +0200 (EET) Received: by mail-pl1-f180.google.com with SMTP id d9443c01a7336-2162c0f6a39so123119975ad.0 for ; Tue, 28 Jan 2025 07:54:59 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20230601; t=1738079697; x=1738684497; darn=ffmpeg.org; h=content-transfer-encoding:mime-version:message-id:date:subject:to :from:from:to:cc:subject:date:message-id:reply-to; bh=8a6IN5O+JF0QUE/d41APumttOsS+SaVDPmpC/g+aXAc=; b=NM7TUo3iEkDnyrRaj0dHxU4dlfUmXlF0k6/roay1axEC8dT/fxAvtgiXCO7ZGWDteq ZaD+2RmuWPGYFE05eCBP1djCRL+3CcPL8cezkGRE2xrTzyJYQj2DjEYTxWIvxeoraIYO cWO37KYl8rrbEt0/o2RUvRZ0Bj/9n9HB+rt/SKhyyS8iDkNsEW6UvCjYoctAPCkbindK WPVQ0KqE6FVFWgrXoEkFAq/Ef174QD+FBWmfLd9eR4/wxxQ/3HTgE0Q2S6KYirxUeMlM plcJ+EHRlK6ey9MDcxkfG5fkkHD3ytfXpgnAm1kgBkrtWjr6jcwJBEZTMoXH4ezRo8e8 9P8w== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1738079697; x=1738684497; h=content-transfer-encoding:mime-version:message-id:date:subject:to :from:x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=8a6IN5O+JF0QUE/d41APumttOsS+SaVDPmpC/g+aXAc=; b=ZpnSQwcTpt9Uee4ej1ICil4aRO8P4Bp7b36FbJJaOXtGVm8ZqOgKooQ0w0tYsNYefB PUh8N5i6ouVpY1hzUODTDpWPbhaDCtiDYFlZCd2SG6o/4DdZW5ekNK4fAK/6/qvxUgoO rN0C/26DqoUMlyXbHdqrvGGAJJmWT2TDceuz0MFu3pSbpfXs9fkRDbFzMfOTtpE8V8c4 sZGL9OmxOdWPCuXrNUwFnVlWLj4aA7iC23ge3BX3GgpOZ3y8NoubO5FNVHBpMqRruMCb ssB/JkrKaPnRRUj8ygd1qxfDOqL8Hm/CucMIAIZn99wOowmazJ00urcJ9klgCe6NsQFH TgnA== X-Gm-Message-State: AOJu0YyzC459wBwBpFkJ8S/mxiRfOYQwzoUVjLJOGUnlpv4vBRn08efC PgpdqF/XQF5v8qTawR2zPhtg+x9wQ3KV1f3OcIL3VTFmt1gaPo/wb2g8fA== X-Gm-Gg: ASbGncsSHemkiGKdYzTdRtZJYYXS100oNlizndN2c1kWUO1iPzvTJcK2Ri32OztyRFO MV6RAk53HQ9Rmk5FyK/4NwtVNclLq9pOtJTnzwVrahqUbp+HTtEGGa/9Q/qJ60sTk7M8rWxcqXv dDncsk78KE/zuk4R0Ureu7SFLacSjETk0+Bg4wveatRhtTUbmUhlYiTTcvr1MNqucrUbIYv/fNw /wEfGEveba3p3zltRLVhfpz7BD9AFEChkq22GbGXMaRxCL8egquCKpT6HDEppM8aO/GjywbDm3t gss3EwpdRMuh+TgGD8PupDm+ZyaDjb8sxmyf6g== X-Google-Smtp-Source: AGHT+IGiwMJLnsm6kZTPRVlipQ7AamkQK67CVHXSJtN75k7hx9T8TYtsi9Qs0fyGlWHXDwZytYVaAg== X-Received: by 2002:a17:902:ecc7:b0:215:a808:61cf with SMTP id d9443c01a7336-21dcc56dd23mr60601215ad.25.1738079696681; Tue, 28 Jan 2025 07:54:56 -0800 (PST) Received: from localhost.localdomain ([106.51.30.120]) by smtp.gmail.com with ESMTPSA id d9443c01a7336-21da424ef28sm83107075ad.258.2025.01.28.07.54.55 for (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Tue, 28 Jan 2025 07:54:56 -0800 (PST) From: Shreesh Adiga <16567adigashreesh@gmail.com> To: ffmpeg-devel@ffmpeg.org Date: Tue, 28 Jan 2025 21:18:57 +0530 Message-ID: <20250128155444.256068-1-16567adigashreesh@gmail.com> X-Mailer: git-send-email 2.45.3 MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH v3] swscale/x86/rgb2rgb: add AVX512ICL versions of shuffle_bytes 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: Signed-off-by: Shreesh Adiga <16567adigashreesh@gmail.com> --- v3: Fix build failure on older nasm by replacing "kmovw k, tmpw" with "kmov k, tmpd" which matches "kmovw k, r32" syntax. v2: Tried to align operands and improve indentation for ASM routine. libswscale/x86/rgb2rgb.c | 21 +++++++++ libswscale/x86/rgb_2_rgb.asm | 90 +++++++++++++++++++++++------------- 2 files changed, 80 insertions(+), 31 deletions(-) diff --git a/libswscale/x86/rgb2rgb.c b/libswscale/x86/rgb2rgb.c index 6790551a38..4cbed54b35 100644 --- a/libswscale/x86/rgb2rgb.c +++ b/libswscale/x86/rgb2rgb.c @@ -2364,6 +2364,16 @@ void ff_shuffle_bytes_2013_avx2(const uint8_t *src, uint8_t *dst, int src_size); void ff_shuffle_bytes_2130_avx2(const uint8_t *src, uint8_t *dst, int src_size); void ff_shuffle_bytes_1203_avx2(const uint8_t *src, uint8_t *dst, int src_size); +void ff_shuffle_bytes_2103_avx512icl(const uint8_t *src, uint8_t *dst, int src_size); +void ff_shuffle_bytes_0321_avx512icl(const uint8_t *src, uint8_t *dst, int src_size); +void ff_shuffle_bytes_1230_avx512icl(const uint8_t *src, uint8_t *dst, int src_size); +void ff_shuffle_bytes_3012_avx512icl(const uint8_t *src, uint8_t *dst, int src_size); +void ff_shuffle_bytes_3210_avx512icl(const uint8_t *src, uint8_t *dst, int src_size); +void ff_shuffle_bytes_3102_avx512icl(const uint8_t *src, uint8_t *dst, int src_size); +void ff_shuffle_bytes_2013_avx512icl(const uint8_t *src, uint8_t *dst, int src_size); +void ff_shuffle_bytes_2130_avx512icl(const uint8_t *src, uint8_t *dst, int src_size); +void ff_shuffle_bytes_1203_avx512icl(const uint8_t *src, uint8_t *dst, int src_size); + void ff_uyvytoyuv422_sse2(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src, int width, int height, int lumStride, int chromStride, int srcStride); @@ -2454,6 +2464,17 @@ av_cold void rgb2rgb_init_x86(void) shuffle_bytes_2130 = ff_shuffle_bytes_2130_avx2; shuffle_bytes_1203 = ff_shuffle_bytes_1203_avx2; } + if (EXTERNAL_AVX512ICL(cpu_flags)) { + shuffle_bytes_0321 = ff_shuffle_bytes_0321_avx512icl; + shuffle_bytes_2103 = ff_shuffle_bytes_2103_avx512icl; + shuffle_bytes_1230 = ff_shuffle_bytes_1230_avx512icl; + shuffle_bytes_3012 = ff_shuffle_bytes_3012_avx512icl; + shuffle_bytes_3210 = ff_shuffle_bytes_3210_avx512icl; + shuffle_bytes_3102 = ff_shuffle_bytes_3102_avx512icl; + shuffle_bytes_2013 = ff_shuffle_bytes_2013_avx512icl; + shuffle_bytes_2130 = ff_shuffle_bytes_2130_avx512icl; + shuffle_bytes_1203 = ff_shuffle_bytes_1203_avx512icl; + } if (EXTERNAL_AVX2_FAST(cpu_flags)) { uyvytoyuv422 = ff_uyvytoyuv422_avx2; #endif diff --git a/libswscale/x86/rgb_2_rgb.asm b/libswscale/x86/rgb_2_rgb.asm index b468beb12d..ca7a481255 100644 --- a/libswscale/x86/rgb_2_rgb.asm +++ b/libswscale/x86/rgb_2_rgb.asm @@ -57,40 +57,53 @@ SECTION .text %macro SHUFFLE_BYTES 4 cglobal shuffle_bytes_%1%2%3%4, 3, 5, 2, src, dst, w, tmp, x VBROADCASTI128 m0, [pb_shuffle%1%2%3%4] - movsxdifnidn wq, wd - mov xq, wq - - add srcq, wq - add dstq, wq - neg wq - -;calc scalar loop + movsxdifnidn wq, wd + mov xq, wq + + add srcq, wq + add dstq, wq + neg wq + +%if mmsize == 64 + and xq, mmsize - 4 + shr xq, 2 + mov tmpd, -1 + shlx tmpd, tmpd, xd + not tmpd + kmovw k7, tmpd + vmovdqu32 m1{k7}{z}, [srcq + wq] + pshufb m1, m0 + vmovdqu32 [dstq + wq]{k7}, m1 + lea wq, [wq + 4 * xq] +%else + ;calc scalar loop and xq, mmsize-4 je .loop_simd -.loop_scalar: - mov tmpb, [srcq + wq + %1] - mov [dstq+wq + 0], tmpb - mov tmpb, [srcq + wq + %2] - mov [dstq+wq + 1], tmpb - mov tmpb, [srcq + wq + %3] - mov [dstq+wq + 2], tmpb - mov tmpb, [srcq + wq + %4] - mov [dstq+wq + 3], tmpb - add wq, 4 - sub xq, 4 - jg .loop_scalar - -;check if src_size < mmsize -cmp wq, 0 -jge .end - -.loop_simd: - movu m1, [srcq+wq] - pshufb m1, m0 - movu [dstq+wq], m1 - add wq, mmsize - jl .loop_simd + .loop_scalar: + mov tmpb, [srcq + wq + %1] + mov [dstq+wq + 0], tmpb + mov tmpb, [srcq + wq + %2] + mov [dstq+wq + 1], tmpb + mov tmpb, [srcq + wq + %3] + mov [dstq+wq + 2], tmpb + mov tmpb, [srcq + wq + %4] + mov [dstq+wq + 3], tmpb + add wq, 4 + sub xq, 4 + jg .loop_scalar +%endif + + ;check if src_size < mmsize + cmp wq, 0 + jge .end + + .loop_simd: + movu m1, [srcq + wq] + pshufb m1, m0 + movu [dstq + wq], m1 + add wq, mmsize + jl .loop_simd .end: RET @@ -122,6 +135,21 @@ SHUFFLE_BYTES 1, 2, 0, 3 %endif %endif +%if ARCH_X86_64 +%if HAVE_AVX512ICL_EXTERNAL +INIT_ZMM avx512icl +SHUFFLE_BYTES 2, 1, 0, 3 +SHUFFLE_BYTES 0, 3, 2, 1 +SHUFFLE_BYTES 1, 2, 3, 0 +SHUFFLE_BYTES 3, 0, 1, 2 +SHUFFLE_BYTES 3, 2, 1, 0 +SHUFFLE_BYTES 3, 1, 0, 2 +SHUFFLE_BYTES 2, 0, 1, 3 +SHUFFLE_BYTES 2, 1, 3, 0 +SHUFFLE_BYTES 1, 2, 0, 3 +%endif +%endif + ;----------------------------------------------------------------------------------------------- ; uyvytoyuv422(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, ; const uint8_t *src, int width, int height, -- 2.45.3 _______________________________________________ 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".