- Masks can only be -1 or 0 based on Deblock::edgeFilterChroma.
- Also fix the input buffer offsets of Chroma and LumaStrong in the
speedup calculation.
---
source/test/pixelharness.cpp | 109 +++++++++++++++++++++--------------
1 file changed, 67 insertions(+), 42 deletions(-)
diff --git a/source/test/pixelharness.cpp b/source/test/pixelharness.cpp
index 311985d83..f3d68238a 100644
--- a/source/test/pixelharness.cpp
+++ b/source/test/pixelharness.cpp
@@ -1749,7 +1749,6 @@ bool PixelHarness::check_scanPosLast(scanPosLast_t ref,
scanPosLast_t opt)
{
ref_src[32 * 32 + i] = 0x1234;
}
-
memset(ref_coeffNum, 0xCD, sizeof(ref_coeffNum));
memset(ref_coeffSign, 0xCD, sizeof(ref_coeffSign));
@@ -2096,7 +2095,7 @@ bool
PixelHarness::check_pelFilterLumaStrong_H(pelFilterLumaStrong_t ref, pelFil
{
intptr_t srcStep = 1, offset = 64;
int32_t tcP, tcQ, maskP, maskQ, tc;
- int j = 0;
+ intptr_t j = 0;
pixel pixel_test_buff1[TEST_CASES][BUFFSIZE];
for (int i = 0; i < TEST_CASES; i++)
@@ -2112,7 +2111,7 @@ bool
PixelHarness::check_pelFilterLumaStrong_H(pelFilterLumaStrong_t ref, pelFil
int index = rand() % 3;
- ref(pixel_test_buff[index] + 4 * offset + j, srcStep, offset, tcP,
tcQ);
+ ref(pixel_test_buff[index] + 4 * offset + j, srcStep, offset, tcP,
tcQ);
checked(opt, pixel_test_buff1[index] + 4 * offset + j, srcStep,
offset, tcP, tcQ);
if (memcmp(pixel_test_buff[index], pixel_test_buff1[index],
sizeof(pixel) * BUFFSIZE))
@@ -2129,7 +2128,7 @@ bool
PixelHarness::check_pelFilterLumaStrong_V(pelFilterLumaStrong_t ref, pelFil
{
intptr_t srcStep = 64, offset = 1;
int32_t tcP, tcQ, maskP, maskQ, tc;
- int j = 0;
+ intptr_t j = 0;
pixel pixel_test_buff1[TEST_CASES][BUFFSIZE];
for (int i = 0; i < TEST_CASES; i++)
@@ -2145,8 +2144,8 @@ bool
PixelHarness::check_pelFilterLumaStrong_V(pelFilterLumaStrong_t ref, pelFil
int index = rand() % 3;
- ref(pixel_test_buff[index] + 4 + j, srcStep, offset, tcP, tcQ);
- checked(opt, pixel_test_buff1[index] + 4 + j, srcStep, offset, tcP,
tcQ);
+ ref(pixel_test_buff[index] + 4 * offset + j, srcStep, offset, tcP,
tcQ);
+ checked(opt, pixel_test_buff1[index] + 4 * offset + j, srcStep,
offset, tcP, tcQ);
if (memcmp(pixel_test_buff[index], pixel_test_buff1[index],
sizeof(pixel) * BUFFSIZE))
return false;
@@ -2162,28 +2161,35 @@ bool
PixelHarness::check_pelFilterChroma_H(pelFilterChroma_t ref, pelFilterChrom
{
intptr_t srcStep = 1, offset = 64;
int32_t maskP, maskQ, tc;
- int j = 0;
+ intptr_t j = 0;
+ const int NUM_MASKS = 3;
pixel pixel_test_buff1[TEST_CASES][BUFFSIZE];
for (int i = 0; i < TEST_CASES; i++)
memcpy(pixel_test_buff1[i], pixel_test_buff[i], sizeof(pixel)*
BUFFSIZE);
- for (int i = 0; i < ITERS; i++)
+ int32_t masks[NUM_MASKS][2] = {{-1, -1}, {-1, 0}, {0, -1}};
+
+ for (int i = 0; i < NUM_MASKS; i++)
{
- tc = rand() % PIXEL_MAX;
- maskP = (rand() % PIXEL_MAX) - 1;
- maskQ = (rand() % PIXEL_MAX) - 1;
+ int32_t maskP = masks[i][0];
+ int32_t maskQ = masks[i][1];
- int index = rand() % 3;
+ for (int i = 0; i < ITERS; i++)
+ {
+ tc = rand() % PIXEL_MAX;
- ref(pixel_test_buff[index] + 4 * offset + j, srcStep, offset, tc,
maskP, maskQ);
- checked(opt, pixel_test_buff1[index] + 4 * offset + j, srcStep,
offset, tc, maskP, maskQ);
+ int index = rand() % 3;
- if (memcmp(pixel_test_buff[index], pixel_test_buff1[index],
sizeof(pixel)* BUFFSIZE))
- return false;
+ ref(pixel_test_buff[index] + 2 * offset + j, srcStep, offset, tc,
maskP, maskQ);
+ checked(opt, pixel_test_buff1[index] + 2 * offset + j, srcStep,
offset, tc, maskP, maskQ);
- reportfail()
- j += INCR;
+ if (memcmp(pixel_test_buff[index], pixel_test_buff1[index],
sizeof(pixel) * BUFFSIZE))
+ return false;
+
+ reportfail()
+ j += INCR;
+ }
}
return true;
@@ -2193,28 +2199,35 @@ bool
PixelHarness::check_pelFilterChroma_V(pelFilterChroma_t ref, pelFilterChrom
{
intptr_t srcStep = 64, offset = 1;
int32_t maskP, maskQ, tc;
- int j = 0;
+ intptr_t j = 0;
+ const int NUM_MASKS = 3;
pixel pixel_test_buff1[TEST_CASES][BUFFSIZE];
for (int i = 0; i < TEST_CASES; i++)
memcpy(pixel_test_buff1[i], pixel_test_buff[i], sizeof(pixel)*
BUFFSIZE);
- for (int i = 0; i < ITERS; i++)
+ int32_t masks[NUM_MASKS][2] = {{-1, -1}, {-1, 0}, {0, -1}};
+
+ for (int i = 0; i < NUM_MASKS; i++)
{
- tc = rand() % PIXEL_MAX;
- maskP = (rand() % PIXEL_MAX) - 1;
- maskQ = (rand() % PIXEL_MAX) - 1;
+ int32_t maskP = masks[i][0];
+ int32_t maskQ = masks[i][1];
- int index = rand() % 3;
+ for (int i = 0; i < ITERS; i++)
+ {
+ tc = rand() % PIXEL_MAX;
- ref(pixel_test_buff[index] + 4 + j, srcStep, offset, tc, maskP, maskQ);
- checked(opt, pixel_test_buff1[index] + 4 + j, srcStep, offset, tc,
maskP, maskQ);
+ int index = rand() % 3;
- if (memcmp(pixel_test_buff[index], pixel_test_buff1[index],
sizeof(pixel)* BUFFSIZE))
- return false;
+ ref(pixel_test_buff[index] + 2 * offset + j, srcStep, offset, tc,
maskP, maskQ);
+ checked(opt, pixel_test_buff1[index] + 2 * offset + j, srcStep,
offset, tc, maskP, maskQ);
- reportfail()
- j += INCR;
+ if (memcmp(pixel_test_buff[index], pixel_test_buff1[index],
sizeof(pixel) * BUFFSIZE))
+ return false;
+
+ reportfail()
+ j += INCR;
+ }
}
return true;
@@ -3085,7 +3098,6 @@ bool PixelHarness::testCorrectness(const
EncoderPrimitives& ref, const EncoderPr
return false;
}
}
-
if (opt.pelFilterLumaStrong[0])
{
@@ -3152,7 +3164,6 @@ bool PixelHarness::testCorrectness(const
EncoderPrimitives& ref, const EncoderPr
}
}
-
for (int k = 0; k < NUM_INTEGRAL_SIZE; k++)
{
if (opt.integral_inith[k] &&
!check_integral_inith(ref.integral_inith[k], opt.integral_inith[k]))
@@ -3749,7 +3760,8 @@ void PixelHarness::measureSpeed(const EncoderPrimitives&
ref, const EncoderPrimi
int32_t tcP = (rand() % PIXEL_MAX) - 1;
int32_t tcQ = (rand() % PIXEL_MAX) - 1;
HEADER0("pelFilterLumaStrong_Vertical");
- REPORT_SPEEDUP(opt.pelFilterLumaStrong[0], ref.pelFilterLumaStrong[0],
pbuf1, STRIDE, 1, tcP, tcQ);
+ REPORT_SPEEDUP(opt.pelFilterLumaStrong[0], ref.pelFilterLumaStrong[0],
pbuf1 + 4,
+ STRIDE, 1, tcP, tcQ);
}
if (opt.pelFilterLumaStrong[1])
@@ -3757,25 +3769,38 @@ void PixelHarness::measureSpeed(const
EncoderPrimitives& ref, const EncoderPrimi
int32_t tcP = (rand() % PIXEL_MAX) - 1;
int32_t tcQ = (rand() % PIXEL_MAX) - 1;
HEADER0("pelFilterLumaStrong_Horizontal");
- REPORT_SPEEDUP(opt.pelFilterLumaStrong[1], ref.pelFilterLumaStrong[1],
pbuf1, 1, STRIDE, tcP, tcQ);
+ REPORT_SPEEDUP(opt.pelFilterLumaStrong[1], ref.pelFilterLumaStrong[1],
pbuf1 + 4 * STRIDE,
+ 1, STRIDE, tcP, tcQ);
}
if (opt.pelFilterChroma[0])
{
+ const int NUM_MASKS = 3;
+ int32_t masks[NUM_MASKS][2] = {{-1, -1}, {-1, 0}, {0, -1}};
int32_t tc = (rand() % PIXEL_MAX);
- int32_t maskP = (rand() % PIXEL_MAX) - 1;
- int32_t maskQ = (rand() % PIXEL_MAX) - 1;
- HEADER0("pelFilterChroma_Vertical");
- REPORT_SPEEDUP(opt.pelFilterChroma[0], ref.pelFilterChroma[0], pbuf1,
STRIDE, 1, tc, maskP, maskQ);
+ for (int i = 0; i < NUM_MASKS; i++)
+ {
+ int32_t maskP = masks[i][0];
+ int32_t maskQ = masks[i][1];
+ HEADER("pelFilterChroma_Vertical[MaskP: %d, MaskQ: %d]", maskP,
maskQ);
+ REPORT_SPEEDUP(opt.pelFilterChroma[0], ref.pelFilterChroma[0],
pbuf1 + 2,
+ STRIDE, 1, tc, maskP, maskQ);
+ }
}
if (opt.pelFilterChroma[1])
{
+ const int NUM_MASKS = 3;
+ int32_t masks[NUM_MASKS][2] = {{-1, -1}, {-1, 0}, {0, -1}};
int32_t tc = (rand() % PIXEL_MAX);
- int32_t maskP = (rand() % PIXEL_MAX) - 1;
- int32_t maskQ = (rand() % PIXEL_MAX) - 1;
- HEADER0("pelFilterChroma_Horizontal");
- REPORT_SPEEDUP(opt.pelFilterChroma[1], ref.pelFilterChroma[1], pbuf1,
1, STRIDE, tc, maskP, maskQ);
+ for (int i = 0; i < NUM_MASKS; i++)
+ {
+ int32_t maskP = masks[i][0];
+ int32_t maskQ = masks[i][1];
+ HEADER("pelFilterChroma_Horizontal[MaskP: %d, MaskQ: %d]", maskP,
maskQ);
+ REPORT_SPEEDUP(opt.pelFilterChroma[1], ref.pelFilterChroma[1],
pbuf1 + 2 * STRIDE,
+ 1, STRIDE, tc, maskP, maskQ);
+ }
}
for (int k = 0; k < NUM_INTEGRAL_SIZE; k++)
--
2.34.1
>From cde8fd021e7dd2d805b648ddf65609391172defd Mon Sep 17 00:00:00 2001
Message-Id: <cde8fd021e7dd2d805b648ddf65609391172defd.1739282617.git.microdaryl.rob...@arm.com>
In-Reply-To: <cover.1739282617.git.microdaryl.rob...@arm.com>
References: <cover.1739282617.git.microdaryl.rob...@arm.com>
From: Micro Daryl Robles <microdaryl.rob...@arm.com>
Date: Thu, 22 Aug 2024 16:49:17 +0100
Subject: [PATCH 3/5] Change the test conditions for pelFilterChroma in
PixelHarness
- Masks can only be -1 or 0 based on Deblock::edgeFilterChroma.
- Also fix the input buffer offsets of Chroma and LumaStrong in the
speedup calculation.
---
source/test/pixelharness.cpp | 109 +++++++++++++++++++++--------------
1 file changed, 67 insertions(+), 42 deletions(-)
diff --git a/source/test/pixelharness.cpp b/source/test/pixelharness.cpp
index 311985d83..f3d68238a 100644
--- a/source/test/pixelharness.cpp
+++ b/source/test/pixelharness.cpp
@@ -1749,7 +1749,6 @@ bool PixelHarness::check_scanPosLast(scanPosLast_t ref, scanPosLast_t opt)
{
ref_src[32 * 32 + i] = 0x1234;
}
-
memset(ref_coeffNum, 0xCD, sizeof(ref_coeffNum));
memset(ref_coeffSign, 0xCD, sizeof(ref_coeffSign));
@@ -2096,7 +2095,7 @@ bool PixelHarness::check_pelFilterLumaStrong_H(pelFilterLumaStrong_t ref, pelFil
{
intptr_t srcStep = 1, offset = 64;
int32_t tcP, tcQ, maskP, maskQ, tc;
- int j = 0;
+ intptr_t j = 0;
pixel pixel_test_buff1[TEST_CASES][BUFFSIZE];
for (int i = 0; i < TEST_CASES; i++)
@@ -2112,7 +2111,7 @@ bool PixelHarness::check_pelFilterLumaStrong_H(pelFilterLumaStrong_t ref, pelFil
int index = rand() % 3;
- ref(pixel_test_buff[index] + 4 * offset + j, srcStep, offset, tcP, tcQ);
+ ref(pixel_test_buff[index] + 4 * offset + j, srcStep, offset, tcP, tcQ);
checked(opt, pixel_test_buff1[index] + 4 * offset + j, srcStep, offset, tcP, tcQ);
if (memcmp(pixel_test_buff[index], pixel_test_buff1[index], sizeof(pixel) * BUFFSIZE))
@@ -2129,7 +2128,7 @@ bool PixelHarness::check_pelFilterLumaStrong_V(pelFilterLumaStrong_t ref, pelFil
{
intptr_t srcStep = 64, offset = 1;
int32_t tcP, tcQ, maskP, maskQ, tc;
- int j = 0;
+ intptr_t j = 0;
pixel pixel_test_buff1[TEST_CASES][BUFFSIZE];
for (int i = 0; i < TEST_CASES; i++)
@@ -2145,8 +2144,8 @@ bool PixelHarness::check_pelFilterLumaStrong_V(pelFilterLumaStrong_t ref, pelFil
int index = rand() % 3;
- ref(pixel_test_buff[index] + 4 + j, srcStep, offset, tcP, tcQ);
- checked(opt, pixel_test_buff1[index] + 4 + j, srcStep, offset, tcP, tcQ);
+ ref(pixel_test_buff[index] + 4 * offset + j, srcStep, offset, tcP, tcQ);
+ checked(opt, pixel_test_buff1[index] + 4 * offset + j, srcStep, offset, tcP, tcQ);
if (memcmp(pixel_test_buff[index], pixel_test_buff1[index], sizeof(pixel) * BUFFSIZE))
return false;
@@ -2162,28 +2161,35 @@ bool PixelHarness::check_pelFilterChroma_H(pelFilterChroma_t ref, pelFilterChrom
{
intptr_t srcStep = 1, offset = 64;
int32_t maskP, maskQ, tc;
- int j = 0;
+ intptr_t j = 0;
+ const int NUM_MASKS = 3;
pixel pixel_test_buff1[TEST_CASES][BUFFSIZE];
for (int i = 0; i < TEST_CASES; i++)
memcpy(pixel_test_buff1[i], pixel_test_buff[i], sizeof(pixel)* BUFFSIZE);
- for (int i = 0; i < ITERS; i++)
+ int32_t masks[NUM_MASKS][2] = {{-1, -1}, {-1, 0}, {0, -1}};
+
+ for (int i = 0; i < NUM_MASKS; i++)
{
- tc = rand() % PIXEL_MAX;
- maskP = (rand() % PIXEL_MAX) - 1;
- maskQ = (rand() % PIXEL_MAX) - 1;
+ int32_t maskP = masks[i][0];
+ int32_t maskQ = masks[i][1];
- int index = rand() % 3;
+ for (int i = 0; i < ITERS; i++)
+ {
+ tc = rand() % PIXEL_MAX;
- ref(pixel_test_buff[index] + 4 * offset + j, srcStep, offset, tc, maskP, maskQ);
- checked(opt, pixel_test_buff1[index] + 4 * offset + j, srcStep, offset, tc, maskP, maskQ);
+ int index = rand() % 3;
- if (memcmp(pixel_test_buff[index], pixel_test_buff1[index], sizeof(pixel)* BUFFSIZE))
- return false;
+ ref(pixel_test_buff[index] + 2 * offset + j, srcStep, offset, tc, maskP, maskQ);
+ checked(opt, pixel_test_buff1[index] + 2 * offset + j, srcStep, offset, tc, maskP, maskQ);
- reportfail()
- j += INCR;
+ if (memcmp(pixel_test_buff[index], pixel_test_buff1[index], sizeof(pixel) * BUFFSIZE))
+ return false;
+
+ reportfail()
+ j += INCR;
+ }
}
return true;
@@ -2193,28 +2199,35 @@ bool PixelHarness::check_pelFilterChroma_V(pelFilterChroma_t ref, pelFilterChrom
{
intptr_t srcStep = 64, offset = 1;
int32_t maskP, maskQ, tc;
- int j = 0;
+ intptr_t j = 0;
+ const int NUM_MASKS = 3;
pixel pixel_test_buff1[TEST_CASES][BUFFSIZE];
for (int i = 0; i < TEST_CASES; i++)
memcpy(pixel_test_buff1[i], pixel_test_buff[i], sizeof(pixel)* BUFFSIZE);
- for (int i = 0; i < ITERS; i++)
+ int32_t masks[NUM_MASKS][2] = {{-1, -1}, {-1, 0}, {0, -1}};
+
+ for (int i = 0; i < NUM_MASKS; i++)
{
- tc = rand() % PIXEL_MAX;
- maskP = (rand() % PIXEL_MAX) - 1;
- maskQ = (rand() % PIXEL_MAX) - 1;
+ int32_t maskP = masks[i][0];
+ int32_t maskQ = masks[i][1];
- int index = rand() % 3;
+ for (int i = 0; i < ITERS; i++)
+ {
+ tc = rand() % PIXEL_MAX;
- ref(pixel_test_buff[index] + 4 + j, srcStep, offset, tc, maskP, maskQ);
- checked(opt, pixel_test_buff1[index] + 4 + j, srcStep, offset, tc, maskP, maskQ);
+ int index = rand() % 3;
- if (memcmp(pixel_test_buff[index], pixel_test_buff1[index], sizeof(pixel)* BUFFSIZE))
- return false;
+ ref(pixel_test_buff[index] + 2 * offset + j, srcStep, offset, tc, maskP, maskQ);
+ checked(opt, pixel_test_buff1[index] + 2 * offset + j, srcStep, offset, tc, maskP, maskQ);
- reportfail()
- j += INCR;
+ if (memcmp(pixel_test_buff[index], pixel_test_buff1[index], sizeof(pixel) * BUFFSIZE))
+ return false;
+
+ reportfail()
+ j += INCR;
+ }
}
return true;
@@ -3085,7 +3098,6 @@ bool PixelHarness::testCorrectness(const EncoderPrimitives& ref, const EncoderPr
return false;
}
}
-
if (opt.pelFilterLumaStrong[0])
{
@@ -3152,7 +3164,6 @@ bool PixelHarness::testCorrectness(const EncoderPrimitives& ref, const EncoderPr
}
}
-
for (int k = 0; k < NUM_INTEGRAL_SIZE; k++)
{
if (opt.integral_inith[k] && !check_integral_inith(ref.integral_inith[k], opt.integral_inith[k]))
@@ -3749,7 +3760,8 @@ void PixelHarness::measureSpeed(const EncoderPrimitives& ref, const EncoderPrimi
int32_t tcP = (rand() % PIXEL_MAX) - 1;
int32_t tcQ = (rand() % PIXEL_MAX) - 1;
HEADER0("pelFilterLumaStrong_Vertical");
- REPORT_SPEEDUP(opt.pelFilterLumaStrong[0], ref.pelFilterLumaStrong[0], pbuf1, STRIDE, 1, tcP, tcQ);
+ REPORT_SPEEDUP(opt.pelFilterLumaStrong[0], ref.pelFilterLumaStrong[0], pbuf1 + 4,
+ STRIDE, 1, tcP, tcQ);
}
if (opt.pelFilterLumaStrong[1])
@@ -3757,25 +3769,38 @@ void PixelHarness::measureSpeed(const EncoderPrimitives& ref, const EncoderPrimi
int32_t tcP = (rand() % PIXEL_MAX) - 1;
int32_t tcQ = (rand() % PIXEL_MAX) - 1;
HEADER0("pelFilterLumaStrong_Horizontal");
- REPORT_SPEEDUP(opt.pelFilterLumaStrong[1], ref.pelFilterLumaStrong[1], pbuf1, 1, STRIDE, tcP, tcQ);
+ REPORT_SPEEDUP(opt.pelFilterLumaStrong[1], ref.pelFilterLumaStrong[1], pbuf1 + 4 * STRIDE,
+ 1, STRIDE, tcP, tcQ);
}
if (opt.pelFilterChroma[0])
{
+ const int NUM_MASKS = 3;
+ int32_t masks[NUM_MASKS][2] = {{-1, -1}, {-1, 0}, {0, -1}};
int32_t tc = (rand() % PIXEL_MAX);
- int32_t maskP = (rand() % PIXEL_MAX) - 1;
- int32_t maskQ = (rand() % PIXEL_MAX) - 1;
- HEADER0("pelFilterChroma_Vertical");
- REPORT_SPEEDUP(opt.pelFilterChroma[0], ref.pelFilterChroma[0], pbuf1, STRIDE, 1, tc, maskP, maskQ);
+ for (int i = 0; i < NUM_MASKS; i++)
+ {
+ int32_t maskP = masks[i][0];
+ int32_t maskQ = masks[i][1];
+ HEADER("pelFilterChroma_Vertical[MaskP: %d, MaskQ: %d]", maskP, maskQ);
+ REPORT_SPEEDUP(opt.pelFilterChroma[0], ref.pelFilterChroma[0], pbuf1 + 2,
+ STRIDE, 1, tc, maskP, maskQ);
+ }
}
if (opt.pelFilterChroma[1])
{
+ const int NUM_MASKS = 3;
+ int32_t masks[NUM_MASKS][2] = {{-1, -1}, {-1, 0}, {0, -1}};
int32_t tc = (rand() % PIXEL_MAX);
- int32_t maskP = (rand() % PIXEL_MAX) - 1;
- int32_t maskQ = (rand() % PIXEL_MAX) - 1;
- HEADER0("pelFilterChroma_Horizontal");
- REPORT_SPEEDUP(opt.pelFilterChroma[1], ref.pelFilterChroma[1], pbuf1, 1, STRIDE, tc, maskP, maskQ);
+ for (int i = 0; i < NUM_MASKS; i++)
+ {
+ int32_t maskP = masks[i][0];
+ int32_t maskQ = masks[i][1];
+ HEADER("pelFilterChroma_Horizontal[MaskP: %d, MaskQ: %d]", maskP, maskQ);
+ REPORT_SPEEDUP(opt.pelFilterChroma[1], ref.pelFilterChroma[1], pbuf1 + 2 * STRIDE,
+ 1, STRIDE, tc, maskP, maskQ);
+ }
}
for (int k = 0; k < NUM_INTEGRAL_SIZE; k++)
--
2.34.1
_______________________________________________
x265-devel mailing list
x265-devel@videolan.org
https://mailman.videolan.org/listinfo/x265-devel
