From patchwork Mon Mar 9 14:54:59 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Fu, Linjie" X-Patchwork-Id: 18099 Return-Path: X-Original-To: patchwork@ffaux-bg.ffmpeg.org Delivered-To: patchwork@ffaux-bg.ffmpeg.org Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org [79.124.17.100]) by ffaux.localdomain (Postfix) with ESMTP id 8C5E844B4C4 for ; Mon, 9 Mar 2020 16:59:41 +0200 (EET) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 6AB16689AE3; Mon, 9 Mar 2020 16:59:41 +0200 (EET) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from mga02.intel.com (mga02.intel.com [134.134.136.20]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id B85AB6880EF for ; Mon, 9 Mar 2020 16:59:34 +0200 (EET) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga005.fm.intel.com ([10.253.24.32]) by orsmga101.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 09 Mar 2020 07:59:32 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.70,533,1574150400"; d="scan'208";a="440945779" Received: from icl-dev.sh.intel.com ([10.239.158.73]) by fmsmga005.fm.intel.com with ESMTP; 09 Mar 2020 07:59:30 -0700 From: Linjie Fu To: ffmpeg-devel@ffmpeg.org Date: Mon, 9 Mar 2020 22:54:59 +0800 Message-Id: <1583765699-27504-1-git-send-email-linjie.fu@intel.com> X-Mailer: git-send-email 2.7.4 Subject: [FFmpeg-devel] [PATCH 1/2] checkasm/hevc_add_res: prepare test data only if the fuction is not tested X-BeenThere: ffmpeg-devel@ffmpeg.org X-Mailman-Version: 2.1.20 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: Linjie Fu MIME-Version: 1.0 Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" check_func will return NULL for functions that have already been tested. If the func is tested and skipped (which happens several times), there is no need to prepare data(randomize_buffers and memcpy). Move relative code in compare_add_res(), prepare data and do check only if the function is not tested. Signed-off-by: Linjie Fu --- tests/checkasm/hevc_add_res.c | 33 ++++++++++++++++++++------------- 1 file changed, 20 insertions(+), 13 deletions(-) diff --git a/tests/checkasm/hevc_add_res.c b/tests/checkasm/hevc_add_res.c index e92c6b4..a68d574 100644 --- a/tests/checkasm/hevc_add_res.c +++ b/tests/checkasm/hevc_add_res.c @@ -42,31 +42,38 @@ AV_WN16A(buf + j * 2, rnd() & 0x3FF); \ } while (0) -static void check_add_res(HEVCDSPContext h, int bit_depth) +static void compare_add_res(int size, ptrdiff_t stride) { - int i; LOCAL_ALIGNED_32(int16_t, res0, [32 * 32]); LOCAL_ALIGNED_32(int16_t, res1, [32 * 32]); LOCAL_ALIGNED_32(uint8_t, dst0, [32 * 32 * 2]); LOCAL_ALIGNED_32(uint8_t, dst1, [32 * 32 * 2]); + declare_func_emms(AV_CPU_FLAG_MMX, void, uint8_t *dst, int16_t *res, ptrdiff_t stride); + + randomize_buffers(res0, size); + randomize_buffers2(dst0, size); + memcpy(res1, res0, sizeof(*res0) * size); + memcpy(dst1, dst0, sizeof(int16_t) * size); + + call_ref(dst0, res0, stride); + call_new(dst1, res1, stride); + if (memcmp(dst0, dst1, size)) + fail(); + bench_new(dst1, res1, stride); +} + +static void check_add_res(HEVCDSPContext h, int bit_depth) +{ + int i; + for (i = 2; i <= 5; i++) { int block_size = 1 << i; int size = block_size * block_size; ptrdiff_t stride = block_size << (bit_depth > 8); - declare_func_emms(AV_CPU_FLAG_MMX, void, uint8_t *dst, int16_t *res, ptrdiff_t stride); - - randomize_buffers(res0, size); - randomize_buffers2(dst0, size); - memcpy(res1, res0, sizeof(*res0) * size); - memcpy(dst1, dst0, sizeof(int16_t) * size); if (check_func(h.add_residual[i - 2], "hevc_add_res_%dx%d_%d", block_size, block_size, bit_depth)) { - call_ref(dst0, res0, stride); - call_new(dst1, res1, stride); - if (memcmp(dst0, dst1, size)) - fail(); - bench_new(dst1, res1, stride); + compare_add_res(size, stride); } } } From patchwork Mon Mar 9 14:55:28 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Fu, Linjie" X-Patchwork-Id: 18100 Return-Path: X-Original-To: patchwork@ffaux-bg.ffmpeg.org Delivered-To: patchwork@ffaux-bg.ffmpeg.org Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org [79.124.17.100]) by ffaux.localdomain (Postfix) with ESMTP id 9460744AAC0 for ; Mon, 9 Mar 2020 17:00:10 +0200 (EET) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 7A0B9689F20; Mon, 9 Mar 2020 17:00:10 +0200 (EET) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from mga11.intel.com (mga11.intel.com [192.55.52.93]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id 1B3FD689EBE for ; Mon, 9 Mar 2020 17:00:03 +0200 (EET) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga006.fm.intel.com ([10.253.24.20]) by fmsmga102.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 09 Mar 2020 08:00:02 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.70,533,1574150400"; d="scan'208";a="442774751" Received: from icl-dev.sh.intel.com ([10.239.158.73]) by fmsmga006.fm.intel.com with ESMTP; 09 Mar 2020 08:00:01 -0700 From: Linjie Fu To: ffmpeg-devel@ffmpeg.org Date: Mon, 9 Mar 2020 22:55:28 +0800 Message-Id: <1583765728-27680-1-git-send-email-linjie.fu@intel.com> X-Mailer: git-send-email 2.7.4 Subject: [FFmpeg-devel] [PATCH 2/2] tests/checkasm: add overflow test for hevc_add_res X-BeenThere: ffmpeg-devel@ffmpeg.org X-Mailman-Version: 2.1.20 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: Xu Guangxin , Linjie Fu MIME-Version: 1.0 Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" Add overflow test for hevc_add_res when int16_t coeff = -32768. The result of C is good, while ASM is not. To verify: make fate-checkasm-hevc_add_res ffmpeg/tests/checkasm/checkasm --test=hevc_add_res ./checkasm --test=hevc_add_res checkasm: using random seed 679391863 MMXEXT: hevc_add_res_4x4_8_mmxext (hevc_add_res.c:69) - hevc_add_res.add_residual [FAILED] SSE2: hevc_add_res_8x8_8_sse2 (hevc_add_res.c:69) hevc_add_res_16x16_8_sse2 (hevc_add_res.c:69) hevc_add_res_32x32_8_sse2 (hevc_add_res.c:69) - hevc_add_res.add_residual [FAILED] AVX: hevc_add_res_8x8_8_avx (hevc_add_res.c:69) hevc_add_res_16x16_8_avx (hevc_add_res.c:69) hevc_add_res_32x32_8_avx (hevc_add_res.c:69) - hevc_add_res.add_residual [FAILED] AVX2: hevc_add_res_32x32_8_avx2 (hevc_add_res.c:69) - hevc_add_res.add_residual [FAILED] checkasm: 8 of 14 tests have failed Signed-off-by: Xu Guangxin Signed-off-by: Linjie Fu --- [v2/v3]: test 2x cases to make sure enough random residuals Should be applied after the 3 fixing patch, otherwise this would break fate. [1] https://patchwork.ffmpeg.org/project/ffmpeg/patch/1583394457-21484-1-git-send-email-linjie.fu@intel.com/ [2] https://patchwork.ffmpeg.org/project/ffmpeg/patch/1583394474-21605-1-git-send-email-linjie.fu@intel.com/ [3] https://patchwork.ffmpeg.org/project/ffmpeg/patch/1583394489-21717-1-git-send-email-linjie.fu@intel.com/ tests/checkasm/hevc_add_res.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/tests/checkasm/hevc_add_res.c b/tests/checkasm/hevc_add_res.c index a68d574..0a3bcbb 100644 --- a/tests/checkasm/hevc_add_res.c +++ b/tests/checkasm/hevc_add_res.c @@ -42,7 +42,7 @@ AV_WN16A(buf + j * 2, rnd() & 0x3FF); \ } while (0) -static void compare_add_res(int size, ptrdiff_t stride) +static void compare_add_res(int size, ptrdiff_t stride, int overflow_test) { LOCAL_ALIGNED_32(int16_t, res0, [32 * 32]); LOCAL_ALIGNED_32(int16_t, res1, [32 * 32]); @@ -53,6 +53,8 @@ static void compare_add_res(int size, ptrdiff_t stride) randomize_buffers(res0, size); randomize_buffers2(dst0, size); + if (overflow_test) + res0[0] = 0x8000; memcpy(res1, res0, sizeof(*res0) * size); memcpy(dst1, dst0, sizeof(int16_t) * size); @@ -73,7 +75,9 @@ static void check_add_res(HEVCDSPContext h, int bit_depth) ptrdiff_t stride = block_size << (bit_depth > 8); if (check_func(h.add_residual[i - 2], "hevc_add_res_%dx%d_%d", block_size, block_size, bit_depth)) { - compare_add_res(size, stride); + compare_add_res(size, stride, 0); + // overflow test for res = -32768 + compare_add_res(size, stride, 1); } } }