Merge "Reinstate tests for wedge partition selection optimizations." into nextgenv2

test/test.mk

@@ -185,6 +185,7 @@ ifeq ($(CONFIG_EXT_INTER),yes)
 LIBVPX_TEST_SRCS-$(HAVE_SSSE3) += masked_variance_test.cc
 LIBVPX_TEST_SRCS-$(HAVE_SSSE3) += masked_sad_test.cc
 LIBVPX_TEST_SRCS-$(CONFIG_VP10_ENCODER) += blend_mask6_test.cc
+LIBVPX_TEST_SRCS-$(CONFIG_VP10_ENCODER) += vp10_wedge_utils_test.cc
 endif
 
 ifeq ($(CONFIG_OBMC),yes)

test/test_libvpx.cc

@@ -22,6 +22,9 @@ extern void vp8_rtcd();
 #if CONFIG_VP9
 extern void vp9_rtcd();
 #endif  // CONFIG_VP9
+#if CONFIG_VP10
+extern void vp10_rtcd();
+#endif  // CONFIG_VP10
 extern void vpx_dsp_rtcd();
 extern void vpx_scale_rtcd();
 }
@@ -69,6 +72,9 @@ int main(int argc, char **argv) {
 #if CONFIG_VP9
   vp9_rtcd();
 #endif  // CONFIG_VP9
+#if CONFIG_VP10
+  vp10_rtcd();
+#endif  // CONFIG_VP10
   vpx_dsp_rtcd();
   vpx_scale_rtcd();
 #endif  // !CONFIG_SHARED

test/vp10_wedge_utils_test.cc

@@ -0,0 +1,415 @@
+/*
+ *  Copyright (c) 2016 The WebM project authors. All Rights Reserved.
+ *
+ *  Use of this source code is governed by a BSD-style license
+ *  that can be found in the LICENSE file in the root of the source
+ *  tree. An additional intellectual property rights grant can be found
+ *  in the file PATENTS.  All contributing project authors may
+ *  be found in the AUTHORS file in the root of the source tree.
+ */
+
+#include "third_party/googletest/src/include/gtest/gtest.h"
+
+#include "./vpx_config.h"
+
+#include "./vpx_dsp_rtcd.h"
+#include "./vp10_rtcd.h"
+
+#include "vpx_dsp/vpx_dsp_common.h"
+
+#include "vp10/common/enums.h"
+
+#include "test/acm_random.h"
+#include "test/function_equivalence_test.h"
+
+#define WEDGE_WEIGHT_BITS 6
+#define MAX_MASK_VALUE  (1 << (WEDGE_WEIGHT_BITS))
+
+using std::tr1::make_tuple;
+using libvpx_test::ACMRandom;
+using libvpx_test::FunctionEquivalenceTest;
+
+namespace {
+
+static const int16_t kInt13Max = (1 << 12) - 1;
+
+//////////////////////////////////////////////////////////////////////////////
+// vp10_wedge_sse_from_residuals - functionality
+//////////////////////////////////////////////////////////////////////////////
+
+class WedgeUtilsSSEFuncTest : public testing::Test {
+ protected:
+  WedgeUtilsSSEFuncTest() : rng_(ACMRandom::DeterministicSeed()) {}
+
+  static const int kIterations = 1000;
+
+  ACMRandom rng_;
+};
+
+static void equiv_blend_residuals(int16_t *r,
+                                  const int16_t *r0,
+                                  const int16_t *r1,
+                                  const uint8_t *m,
+                                  int N) {
+  for (int i = 0 ; i < N ; i++) {
+    const int32_t m0 = m[i];
+    const int32_t m1 = MAX_MASK_VALUE - m0;
+    const int16_t R = m0 * r0[i] + m1 * r1[i];
+    // Note that this rounding is designed to match the result
+    // you would get when actually blending the 2 predictors and computing
+    // the residuals.
+    r[i] = ROUND_POWER_OF_TWO(R - 1, WEDGE_WEIGHT_BITS);
+  }
+}
+
+static uint64_t equiv_sse_from_residuals(const int16_t *r0,
+                                         const int16_t *r1,
+                                         const uint8_t *m,
+                                         int N) {
+  uint64_t acc = 0;
+  for (int i = 0 ; i < N ; i++) {
+    const int32_t m0 = m[i];
+    const int32_t m1 = MAX_MASK_VALUE - m0;
+    const int16_t R = m0 * r0[i] + m1 * r1[i];
+    const int32_t r = ROUND_POWER_OF_TWO(R - 1, WEDGE_WEIGHT_BITS);
+    acc += r * r;
+  }
+  return acc;
+}
+
+TEST_F(WedgeUtilsSSEFuncTest, ResidualBlendingEquiv) {
+  DECLARE_ALIGNED(32, uint8_t, s[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, uint8_t, p0[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, uint8_t, p1[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, uint8_t, p[MAX_SB_SQUARE]);
+
+  DECLARE_ALIGNED(32, int16_t, r0[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, int16_t, r1[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, int16_t, r_ref[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, int16_t, r_tst[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, uint8_t, m[MAX_SB_SQUARE]);
+
+  for (int iter = 0 ; iter < kIterations && !HasFatalFailure(); ++iter) {
+    for (int i = 0 ; i < MAX_SB_SQUARE ; ++i) {
+      s[i] = rng_.Rand8();
+      m[i] = rng_(MAX_MASK_VALUE + 1);
+    }
+
+    const int w = 1 << (rng_(MAX_SB_SIZE_LOG2 + 1 - 3) + 3);
+    const int h = 1 << (rng_(MAX_SB_SIZE_LOG2 + 1 - 3) + 3);
+    const int N = w * h;
+
+    for (int j = 0 ; j < N ; j++) {
+      p0[j] = clamp(s[j] + rng_(33) - 16, 0, UINT8_MAX);
+      p1[j] = clamp(s[j] + rng_(33) - 16, 0, UINT8_MAX);
+    }
+
+    vpx_blend_mask6(p, w, p0, w, p1, w, m, w, h, w, 0, 0);
+
+    vpx_subtract_block(h, w, r0, w, s, w, p0, w);
+    vpx_subtract_block(h, w, r1, w, s, w, p1, w);
+
+    vpx_subtract_block(h, w, r_ref, w, s, w, p, w);
+    equiv_blend_residuals(r_tst, r0, r1, m, N);
+
+    for (int i = 0 ; i < N ; ++i)
+      ASSERT_EQ(r_ref[i], r_tst[i]);
+
+    uint64_t ref_sse = vpx_sum_squares_i16(r_ref, N);
+    uint64_t tst_sse = equiv_sse_from_residuals(r0, r1, m, N);
+
+    ASSERT_EQ(ref_sse, tst_sse);
+  }
+}
+
+static uint64_t sse_from_residuals(const int16_t *r0,
+                                   const int16_t *r1,
+                                   const uint8_t *m,
+                                   int N) {
+  uint64_t acc = 0;
+  for (int i = 0 ; i < N ; i++) {
+    const int32_t m0 = m[i];
+    const int32_t m1 = MAX_MASK_VALUE - m0;
+    const int32_t r = m0 * r0[i] + m1 * r1[i];
+    acc += r * r;
+  }
+  return ROUND_POWER_OF_TWO(acc, 2 * WEDGE_WEIGHT_BITS);
+}
+
+TEST_F(WedgeUtilsSSEFuncTest, ResidualBlendingMethod) {
+  DECLARE_ALIGNED(32, int16_t, r0[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, int16_t, r1[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, int16_t, d[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, uint8_t, m[MAX_SB_SQUARE]);
+
+  for (int iter = 0 ; iter < kIterations && !HasFatalFailure(); ++iter) {
+    for (int i = 0 ; i < MAX_SB_SQUARE ; ++i) {
+      r1[i] = rng_(2 * INT8_MAX - 2 * INT8_MIN + 1) + 2 * INT8_MIN;
+      d[i] = rng_(2 * INT8_MAX - 2 * INT8_MIN + 1) + 2 * INT8_MIN;
+      m[i] = rng_(MAX_MASK_VALUE + 1);
+    }
+
+    const int N = 64 * (rng_(MAX_SB_SQUARE/64) + 1);
+
+    for (int i = 0 ; i < N ; i++)
+      r0[i] = r1[i] + d[i];
+
+    uint64_t ref_res = sse_from_residuals(r0, r1, m, N);
+    uint64_t tst_res = vp10_wedge_sse_from_residuals(r1, d, m, N);
+
+    ASSERT_EQ(ref_res, tst_res);
+  }
+}
+
+//////////////////////////////////////////////////////////////////////////////
+// vp10_wedge_sse_from_residuals - optimizations
+//////////////////////////////////////////////////////////////////////////////
+
+typedef uint64_t (*FSSE)(const int16_t *r1,
+                         const int16_t *d,
+                         const uint8_t *m,
+                         int N);
+
+class WedgeUtilsSSEOptTest : public FunctionEquivalenceTest<FSSE> {
+ protected:
+  WedgeUtilsSSEOptTest() : rng_(ACMRandom::DeterministicSeed()) {}
+
+  static const int kIterations = 10000;
+
+  ACMRandom rng_;
+};
+
+TEST_P(WedgeUtilsSSEOptTest, RandomValues) {
+  DECLARE_ALIGNED(32, int16_t, r1[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, int16_t, d[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, uint8_t, m[MAX_SB_SQUARE]);
+
+  for (int iter = 0 ; iter < kIterations && !HasFatalFailure(); ++iter) {
+    for (int i = 0 ; i < MAX_SB_SQUARE ; ++i) {
+      r1[i] = rng_(2 * kInt13Max + 1) - kInt13Max;
+      d[i] = rng_(2 * kInt13Max + 1) - kInt13Max;
+      m[i] = rng_(MAX_MASK_VALUE + 1);
+    }
+
+    const int N = 64 * (rng_(MAX_SB_SQUARE/64) + 1);
+
+    const uint64_t ref_res = ref_func_(r1, d, m, N);
+    const uint64_t tst_res = tst_func_(r1, d, m, N);
+
+    ASSERT_EQ(ref_res, tst_res);
+  }
+}
+
+TEST_P(WedgeUtilsSSEOptTest, ExtremeValues) {
+  DECLARE_ALIGNED(32, int16_t, r1[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, int16_t, d[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, uint8_t, m[MAX_SB_SQUARE]);
+
+  for (int iter = 0 ; iter < kIterations && !HasFatalFailure(); ++iter) {
+    if (rng_(2)) {
+      for (int i = 0 ; i < MAX_SB_SQUARE ; ++i)
+        r1[i] = kInt13Max;
+    } else {
+      for (int i = 0 ; i < MAX_SB_SQUARE ; ++i)
+        r1[i] = -kInt13Max;
+    }
+
+    if (rng_(2)) {
+      for (int i = 0 ; i < MAX_SB_SQUARE ; ++i)
+        d[i] = kInt13Max;
+    } else {
+      for (int i = 0 ; i < MAX_SB_SQUARE ; ++i)
+        d[i] = -kInt13Max;
+    }
+
+    for (int i = 0 ; i < MAX_SB_SQUARE ; ++i)
+      m[i] = MAX_MASK_VALUE;
+
+    const int N = 64 * (rng_(MAX_SB_SQUARE/64) + 1);
+
+    const uint64_t ref_res = ref_func_(r1, d, m, N);
+    const uint64_t tst_res = tst_func_(r1, d, m, N);
+
+    ASSERT_EQ(ref_res, tst_res);
+  }
+}
+
+#if HAVE_SSE2
+INSTANTIATE_TEST_CASE_P(
+    SSE2, WedgeUtilsSSEOptTest,
+    ::testing::Values(
+        make_tuple(&vp10_wedge_sse_from_residuals_c,
+                   &vp10_wedge_sse_from_residuals_sse2)
+    )
+);
+#endif  // HAVE_SSE2
+
+//////////////////////////////////////////////////////////////////////////////
+// vp10_wedge_sign_from_residuals
+//////////////////////////////////////////////////////////////////////////////
+
+typedef int (*FSign)(const int16_t *ds,
+                     const uint8_t *m,
+                     int N,
+                     int64_t limit);
+
+class WedgeUtilsSignOptTest : public FunctionEquivalenceTest<FSign> {
+ protected:
+  WedgeUtilsSignOptTest() : rng_(ACMRandom::DeterministicSeed()) {}
+
+  static const int kIterations = 10000;
+  static const int kMaxSize = 8196;  // Size limited by SIMD implementation.
+
+  ACMRandom rng_;
+};
+
+TEST_P(WedgeUtilsSignOptTest, RandomValues) {
+  DECLARE_ALIGNED(32, int16_t, r0[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, int16_t, r1[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, int16_t, ds[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, uint8_t, m[MAX_SB_SQUARE]);
+
+  for (int iter = 0 ; iter < kIterations && !HasFatalFailure(); ++iter) {
+    for (int i = 0 ; i < MAX_SB_SQUARE ; ++i) {
+      r0[i] = rng_(2 * kInt13Max + 1) - kInt13Max;
+      r1[i] = rng_(2 * kInt13Max + 1) - kInt13Max;
+      m[i] = rng_(MAX_MASK_VALUE + 1);
+    }
+
+    const int maxN = VPXMIN(kMaxSize, MAX_SB_SQUARE);
+    const int N = 64 * (rng_(maxN/64 - 1) + 1);
+
+    int64_t limit;
+    limit = (int64_t)vpx_sum_squares_i16(r0, N);
+    limit -= (int64_t)vpx_sum_squares_i16(r1, N);
+    limit *= (1 << WEDGE_WEIGHT_BITS) / 2;
+
+    for (int i = 0 ; i < N ; i++)
+      ds[i] = clamp(r0[i]*r0[i] - r1[i]*r1[i], INT16_MIN, INT16_MAX);
+
+    const int ref_res = ref_func_(ds, m, N, limit);
+    const int tst_res = tst_func_(ds, m, N, limit);
+
+    ASSERT_EQ(ref_res, tst_res);
+  }
+}
+
+TEST_P(WedgeUtilsSignOptTest, ExtremeValues) {
+  DECLARE_ALIGNED(32, int16_t, r0[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, int16_t, r1[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, int16_t, ds[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, uint8_t, m[MAX_SB_SQUARE]);
+
+  for (int iter = 0 ; iter < kIterations && !HasFatalFailure(); ++iter) {
+    switch (rng_(4)) {
+    case 0:
+      for (int i = 0 ; i < MAX_SB_SQUARE ; ++i) {
+        r0[i] = 0;
+        r1[i] = kInt13Max;
+      }
+      break;
+    case 1:
+      for (int i = 0 ; i < MAX_SB_SQUARE ; ++i) {
+        r0[i] = kInt13Max;
+        r1[i] = 0;
+      }
+      break;
+    case 2:
+      for (int i = 0 ; i < MAX_SB_SQUARE ; ++i) {
+        r0[i] = 0;
+        r1[i] = -kInt13Max;
+      }
+      break;
+    default:
+      for (int i = 0 ; i < MAX_SB_SQUARE ; ++i) {
+        r0[i] = -kInt13Max;
+        r1[i] = 0;
+      }
+      break;
+    }
+
+    for (int i = 0 ; i < MAX_SB_SQUARE ; ++i)
+      m[i] = MAX_MASK_VALUE;
+
+    const int maxN = VPXMIN(kMaxSize, MAX_SB_SQUARE);
+    const int N = 64 * (rng_(maxN/64 - 1) + 1);
+
+    int64_t limit;
+    limit = (int64_t)vpx_sum_squares_i16(r0, N);
+    limit -= (int64_t)vpx_sum_squares_i16(r1, N);
+    limit *= (1 << WEDGE_WEIGHT_BITS) / 2;
+
+    for (int i = 0 ; i < N ; i++)
+      ds[i] = clamp(r0[i]*r0[i] - r1[i]*r1[i], INT16_MIN, INT16_MAX);
+
+    const int ref_res = ref_func_(ds, m, N, limit);
+    const int tst_res = tst_func_(ds, m, N, limit);
+
+    ASSERT_EQ(ref_res, tst_res);
+  }
+}
+
+#if HAVE_SSE2
+INSTANTIATE_TEST_CASE_P(
+    SSE2, WedgeUtilsSignOptTest,
+    ::testing::Values(
+        make_tuple(&vp10_wedge_sign_from_residuals_c,
+                   &vp10_wedge_sign_from_residuals_sse2)
+    )
+);
+#endif  // HAVE_SSE2
+
+//////////////////////////////////////////////////////////////////////////////
+// vp10_wedge_compute_delta_squares
+//////////////////////////////////////////////////////////////////////////////
+
+typedef void (*FDS)(int16_t *d,
+                    const int16_t *a,
+                    const int16_t *b,
+                    int N);
+
+class WedgeUtilsDeltaSquaresOptTest : public FunctionEquivalenceTest<FDS> {
+ protected:
+  WedgeUtilsDeltaSquaresOptTest() : rng_(ACMRandom::DeterministicSeed()) {}
+
+  static const int kIterations = 10000;
+
+  ACMRandom rng_;
+};
+
+TEST_P(WedgeUtilsDeltaSquaresOptTest, RandomValues) {
+  DECLARE_ALIGNED(32, int16_t, a[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, int16_t, b[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, int16_t, d_ref[MAX_SB_SQUARE]);
+  DECLARE_ALIGNED(32, int16_t, d_tst[MAX_SB_SQUARE]);
+
+  for (int iter = 0 ; iter < kIterations && !HasFatalFailure(); ++iter) {
+    for (int i = 0 ; i < MAX_SB_SQUARE ; ++i) {
+      a[i] = rng_.Rand16();
+      b[i] = rng_(2 * INT16_MAX + 1) - INT16_MAX;
+    }
+
+    const int N = 64 * (rng_(MAX_SB_SQUARE/64) + 1);
+
+    memset(&d_ref, INT16_MAX, sizeof(d_ref));
+    memset(&d_tst, INT16_MAX, sizeof(d_tst));
+
+    ref_func_(d_ref, a, b, N);
+    tst_func_(d_tst, a, b, N);
+
+    for (int i = 0 ; i < MAX_SB_SQUARE ; ++i)
+      ASSERT_EQ(d_ref[i], d_tst[i]);
+  }
+}
+
+#if HAVE_SSE2
+INSTANTIATE_TEST_CASE_P(
+    SSE2, WedgeUtilsDeltaSquaresOptTest,
+    ::testing::Values(
+        make_tuple(&vp10_wedge_compute_delta_squares_c,
+                   &vp10_wedge_compute_delta_squares_sse2)
+    )
+);
+#endif  // HAVE_SSE2
+
+}  // namespace