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 2451241126 for ; Fri, 15 Jul 2022 08:03:51 +0000 (UTC) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id BE67F68B9D7; Fri, 15 Jul 2022 11:03:12 +0300 (EEST) Received: from mail-lj1-f172.google.com (mail-lj1-f172.google.com [209.85.208.172]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id AC9A268B9C5 for ; Fri, 15 Jul 2022 11:03:06 +0300 (EEST) Received: by mail-lj1-f172.google.com with SMTP id w17so4873554ljh.6 for ; Fri, 15 Jul 2022 01:03:06 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=semihalf.com; s=google; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=n6jw87FEiCsAYLGYWZOMEx4dHyRKN1RYEAHyzZFV/hg=; b=ar8hlWAMFvoRk1RMEXgyGTFgKULE716ijaxTFcTFAwqYkUi8aRuqutK9ILuXYtippR sx0K3bf67h4XS+SOqq2TMXi8MWX0CZszqhhOUDtl3sF6XpHKQaPkuLRwSYBiLSdVeJFV t0gWAj6heQfAmHBU/gk79LJRsQTQRXS7FMpVoLaoV7eAFOTQXgff3e89HswULWnhrs2u RQhPBzNSlQ4qqaM86Tom/Ed/zjp7tjZVqCHdy5yn7N55gx2/6np9VZe7ujbR5UukjQme pdQJ6J2ClC8LtQFnwmNbIUfSAuqdMg4M0s31Cs3yJFKdFwNvXIYvoccLL9eWVEhZpaB4 R/Gg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=n6jw87FEiCsAYLGYWZOMEx4dHyRKN1RYEAHyzZFV/hg=; b=L2NrHgzhrBMfyC7/AFJo8Zu1yAOVQsx8+hmw+zNsXUau5OKfVoBy2sLAc3d5B8GjZr PLnO7nBD2+ej3ocZkKCyOTnxqUsvD5CfUEtqSxFpd9wmuAJYBwvls8jbj434X9EsSqBl 2WroWF5LQp3CT0tfT5ijoV1ueVAYmnRKR+/fUFL2GOdvbe0qUwoCCtbW483cgoSY43Yb OLsfmKBDZcIhy+NCcrxPp0ds9YywpGLN5M+INnrkLOXyOx6WroedYzYA7Arq1Giw48sv JCNjGAhhku8GhI9YNC6y0+DoJIonT9OioPll4jagUjNCuuZTSV44/sfAJYXfqs8ux2PN 95lg== X-Gm-Message-State: AJIora8wT9LcLDiHk/GzbiIwt7Cf26YU/GcmPFjnW/Q3XViNl4J22ffK fyFVUCXuwqm4d+wU/8xlWx4VlasRORFI4g== X-Google-Smtp-Source: AGRyM1u90W3zpmEIGYvy7vWFlujlQyrbNHe72iY6FH04y7nFSHH7VlwAYPAGUXTVqblqNLA+Vm9YzQ== X-Received: by 2002:a2e:be04:0:b0:25d:5e4c:9028 with SMTP id z4-20020a2ebe04000000b0025d5e4c9028mr6117114ljq.464.1657872185769; Fri, 15 Jul 2022 01:03:05 -0700 (PDT) Received: from hum-HP-ProBook-440-G7.semihalf.net (adrf243.neoplus.adsl.tpnet.pl. [79.185.165.243]) by smtp.gmail.com with ESMTPSA id k1-20020a2eb741000000b0025d542731ffsm645548ljo.5.2022.07.15.01.03.04 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 15 Jul 2022 01:03:05 -0700 (PDT) From: Hubert Mazur To: ffmpeg-devel@ffmpeg.org Date: Fri, 15 Jul 2022 10:02:27 +0200 Message-Id: <20220715080228.686736-5-hum@semihalf.com> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20220715080228.686736-1-hum@semihalf.com> References: <20220715080228.686736-1-hum@semihalf.com> MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH 4/5] lavc/aarch64: Add neon implementation for sse8 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 Cc: gjb@semihalf.com, upstream@semihalf.com, jswinney@amazon.com, Hubert Mazur , martin@martin.st, mw@semihalf.com, spop@amazon.com 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: Provide optimized implementation of sse8 function for arm64. Performance comparison tests are shown below. - sse_1_c: 133.0 - sse_1_neon: 36.7 Benchmarks and tests run with checkasm tool on AWS Graviton 3. Signed-off-by: Hubert Mazur --- libavcodec/aarch64/me_cmp_init_aarch64.c | 3 + libavcodec/aarch64/me_cmp_neon.S | 72 ++++++++++++++++++++++++ 2 files changed, 75 insertions(+) diff --git a/libavcodec/aarch64/me_cmp_init_aarch64.c b/libavcodec/aarch64/me_cmp_init_aarch64.c index 07d62cc1e5..89c817990c 100644 --- a/libavcodec/aarch64/me_cmp_init_aarch64.c +++ b/libavcodec/aarch64/me_cmp_init_aarch64.c @@ -34,6 +34,8 @@ int ff_pix_abs16_y2_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, int sse16_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t stride, int h); +int sse8_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, + ptrdiff_t stride, int h); int sse4_neon(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, ptrdiff_t stride, int h); @@ -49,6 +51,7 @@ av_cold void ff_me_cmp_init_aarch64(MECmpContext *c, AVCodecContext *avctx) c->sad[0] = ff_pix_abs16_neon; c->sse[0] = sse16_neon; + c->sse[1] = sse8_neon; c->sse[2] = sse4_neon; } } diff --git a/libavcodec/aarch64/me_cmp_neon.S b/libavcodec/aarch64/me_cmp_neon.S index 858833b0ae..c78e26df4b 100644 --- a/libavcodec/aarch64/me_cmp_neon.S +++ b/libavcodec/aarch64/me_cmp_neon.S @@ -361,6 +361,78 @@ function sse16_neon, export=1 endfunc +function sse8_neon, export=1 + // x0 - unused + // x1 - pix1 + // x2 - pix2 + // x3 - stride + // w4 - h + + movi d18, #0 + cmp w4, #4 + b.le 2f + +// make 4 iterations at once +1: + + // res = abs(pix1[0] - pix2[0]) + // res * res + + ld1 {v0.8b}, [x1], x3 + ld1 {v1.8b}, [x2], x3 + uabdl v30.8h, v0.8b, v1.8b + umull v21.4s, v30.4h, v30.4h + ld1 {v2.8b}, [x1], x3 + umull2 v20.4s, v30.8h, v30.8h + ld1 {v3.8b}, [x2], x3 + uabdl v29.8h, v2.8b, v3.8b + ld1 {v4.8b}, [x1], x3 + umlal v21.4s, v29.4h, v29.4h + ld1 {v5.8b}, [x2], x3 + umlal2 v20.4s, v29.8h, v29.8h + uabdl v28.8h, v4.8b, v5.8b + ld1 {v6.8b}, [x1], x3 + umlal v21.4s, v28.4h, v28.4h + ld1 {v7.8b}, [x2], x3 + umlal2 v20.4s, v28.8h, v28.8h + uabdl v27.8h, v6.8b, v7.8b + umlal v21.4s, v27.4h, v27.4h + umlal2 v20.4s, v27.8h, v27.8h + + add v21.4s, v21.4s, v20.4s + sub w4, w4, #4 + uaddlv d17, v21.4s + add d18, d18, d17 + cmp w4, #4 + b.ge 1b + + cbnz w4, 2f + fmov w0, s18 + + ret + +// iterate by one +2: + ld1 {v0.8b}, [x1], x3 + ld1 {v1.8b}, [x2], x3 + + uabdl v30.8h, v0.8b, v1.8b + umull v21.4s, v30.4h, v30.4h + umull2 v20.4s, v30.8h, v30.8h + + subs w4, w4, #1 + + uaddlv d17, v21.4s + add d18, d18, d17 + uaddlv d17, v20.4S + add d18, d18, d17 + + b.ne 2b + fmov w0, s18 + + ret +endfunc + function sse4_neon, export=1 // x0 - unused // x1 - pix1 -- 2.34.1 _______________________________________________ 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".