Add MSA optimized ARGB scaling functions

R=fbarchard@google.com
BUG=libyuv:634

Performance Gain (vs C vectorized)
ScaleARGBRowDown2_MSA           - ~2.6x
ScaleARGBRowDown2Linear_MSA     - ~7.9x
ScaleARGBRowDown2Box_MSA        - ~3.7x
ScaleARGBRowDownEven_MSA        - ~1.2x
ScaleARGBRowDownEvenBox_MSA     - ~3.5x

ScaleARGBRowDown2_Any_MSA       - ~2.6x
ScaleARGBRowDown2Linear_Any_MSA - ~7.9x
ScaleARGBRowDown2Box_Any_MSA    - ~3.6x
ScaleARGBRowDownEven_Any_MSA    - ~1.2x
ScaleARGBRowDownEvenBox_Any_MSA - ~3.5x

Performance Gain (vs C non-vectorized)
ScaleARGBRowDown2_MSA           - 2.6x
ScaleARGBRowDown2Linear_MSA     - 13.5x
ScaleARGBRowDown2Box_MSA        - 5.8x
ScaleARGBRowDownEven_MSA        - 1.2x
ScaleARGBRowDownEvenBox_MSA     - 3.7x

ScaleARGBRowDown2_Any_MSA       - 2.6x
ScaleARGBRowDown2Linear_Any_MSA - 13.5x
ScaleARGBRowDown2Box_Any_MSA    - 5.3x
ScaleARGBRowDownEven_Any_MSA    - 1.2x
ScaleARGBRowDownEvenBox_Any_MSA - 3.7x

Review URL: https://codereview.chromium.org/2527983002 .

Android.mk

@@ -29,7 +29,7 @@ LOCAL_SRC_FILES := \
     source/row_common.cc        \
     source/row_mips.cc          \
     source/row_neon64.cc        \
-    source/row_gcc.cc	        \
+    source/row_gcc.cc           \
     source/scale.cc             \
     source/scale_any.cc         \
     source/scale_argb.cc        \
@@ -56,7 +56,8 @@ endif
 ifeq ($(TARGET_ARCH_ABI),mips)
     LOCAL_CFLAGS += -DLIBYUV_MSA
     LOCAL_SRC_FILES += \
-        source/row_msa.cc
+        source/row_msa.cc    \
+        source/scale_msa.cc
 endif
 
 LOCAL_EXPORT_C_INCLUDES := $(LOCAL_PATH)/include

BUILD.gn

@@ -164,6 +164,7 @@ if (libyuv_use_msa) {
     sources = [
       # MSA Source Files
       "source/row_msa.cc",
+      "source/scale_msa.cc",
     ]
 
     public_configs = [ ":libyuv_config" ]

CMakeLists.txt

@@ -50,6 +50,7 @@ set(ly_source_files
   ${ly_src_dir}/scale_argb.cc
   ${ly_src_dir}/scale_common.cc
   ${ly_src_dir}/scale_mips.cc
+  ${ly_src_dir}/scale_msa.cc
   ${ly_src_dir}/scale_neon.cc
   ${ly_src_dir}/scale_neon64.cc
   ${ly_src_dir}/scale_gcc.cc
@@ -130,13 +131,13 @@ if(TEST)
   if (JPEG_FOUND)
     target_link_libraries(libyuv_unittest ${JPEG_LIBRARY})
   endif()
-  
+
   if(NACL AND NACL_LIBC STREQUAL "newlib")
     target_link_libraries(libyuv_unittest glibc-compat)
   endif()
 
   target_link_libraries(libyuv_unittest gflags)
-  
+
 endif()
 
 install(TARGETS ${ly_lib_name} DESTINATION lib)

include/libyuv/scale_row.h

@@ -103,6 +103,11 @@ extern "C" {
 #define HAS_SCALEROWDOWN38_DSPR2
 #endif
 
+#if !defined(LIBYUV_DISABLE_MSA) && defined(__mips_msa)
+#define HAS_SCALEARGBROWDOWN2_MSA
+#define HAS_SCALEARGBROWDOWNEVEN_MSA
+#endif
+
 // Scale ARGB vertically with bilinear interpolation.
 void ScalePlaneVertical(int src_height,
                         int dst_width,
@@ -562,6 +567,18 @@ void ScaleARGBRowDown2Box_NEON(const uint8* src_ptr,
                                ptrdiff_t src_stride,
                                uint8* dst,
                                int dst_width);
+void ScaleARGBRowDown2_MSA(const uint8_t* src_argb,
+                           ptrdiff_t src_stride,
+                           uint8_t* dst_argb,
+                           int dst_width);
+void ScaleARGBRowDown2Linear_MSA(const uint8_t* src_argb,
+                                 ptrdiff_t src_stride,
+                                 uint8_t* dst_argb,
+                                 int dst_width);
+void ScaleARGBRowDown2Box_MSA(const uint8_t* src_argb,
+                              ptrdiff_t src_stride,
+                              uint8_t* dst_argb,
+                              int dst_width);
 void ScaleARGBRowDown2_Any_SSE2(const uint8* src_argb,
                                 ptrdiff_t src_stride,
                                 uint8* dst_argb,
@@ -586,6 +603,18 @@ void ScaleARGBRowDown2Box_Any_NEON(const uint8* src_ptr,
                                    ptrdiff_t src_stride,
                                    uint8* dst,
                                    int dst_width);
+void ScaleARGBRowDown2_Any_MSA(const uint8_t* src_argb,
+                               ptrdiff_t src_stride,
+                               uint8_t* dst_argb,
+                               int dst_width);
+void ScaleARGBRowDown2Linear_Any_MSA(const uint8_t* src_argb,
+                                     ptrdiff_t src_stride,
+                                     uint8_t* dst_argb,
+                                     int dst_width);
+void ScaleARGBRowDown2Box_Any_MSA(const uint8_t* src_argb,
+                                  ptrdiff_t src_stride,
+                                  uint8_t* dst_argb,
+                                  int dst_width);
 
 void ScaleARGBRowDownEven_SSE2(const uint8* src_argb,
                                ptrdiff_t src_stride,
@@ -607,6 +636,16 @@ void ScaleARGBRowDownEvenBox_NEON(const uint8* src_argb,
                                   int src_stepx,
                                   uint8* dst_argb,
                                   int dst_width);
+void ScaleARGBRowDownEven_MSA(const uint8_t* src_argb,
+                              ptrdiff_t src_stride,
+                              int32_t src_stepx,
+                              uint8_t* dst_argb,
+                              int dst_width);
+void ScaleARGBRowDownEvenBox_MSA(const uint8* src_argb,
+                                 ptrdiff_t src_stride,
+                                 int src_stepx,
+                                 uint8* dst_argb,
+                                 int dst_width);
 void ScaleARGBRowDownEven_Any_SSE2(const uint8* src_argb,
                                    ptrdiff_t src_stride,
                                    int src_stepx,
@@ -627,6 +666,16 @@ void ScaleARGBRowDownEvenBox_Any_NEON(const uint8* src_argb,
                                       int src_stepx,
                                       uint8* dst_argb,
                                       int dst_width);
+void ScaleARGBRowDownEven_Any_MSA(const uint8_t* src_argb,
+                                  ptrdiff_t src_stride,
+                                  int32_t src_stepx,
+                                  uint8_t* dst_argb,
+                                  int dst_width);
+void ScaleARGBRowDownEvenBox_Any_MSA(const uint8* src_argb,
+                                     ptrdiff_t src_stride,
+                                     int src_stepx,
+                                     uint8* dst_argb,
+                                     int dst_width);
 
 // ScaleRowDown2Box also used by planar functions
 // NEON downscalers with interpolation.

libyuv.gypi

@@ -72,6 +72,7 @@
       'source/scale_common.cc',
       'source/scale_gcc.cc',
       'source/scale_mips.cc',
+      'source/scale_msa.cc',
       'source/scale_neon.cc',
       'source/scale_neon64.cc',
       'source/scale_win.cc',

source/scale_any.cc

@@ -283,6 +283,26 @@ SDANY(ScaleARGBRowDown2Box_Any_NEON,
       4,
       7)
 #endif
+#ifdef HAS_SCALEARGBROWDOWN2_MSA
+SDANY(ScaleARGBRowDown2_Any_MSA,
+      ScaleARGBRowDown2_MSA,
+      ScaleARGBRowDown2_C,
+      2,
+      4,
+      3)
+SDANY(ScaleARGBRowDown2Linear_Any_MSA,
+      ScaleARGBRowDown2Linear_MSA,
+      ScaleARGBRowDown2Linear_C,
+      2,
+      4,
+      3)
+SDANY(ScaleARGBRowDown2Box_Any_MSA,
+      ScaleARGBRowDown2Box_MSA,
+      ScaleARGBRowDown2Box_C,
+      2,
+      4,
+      3)
+#endif
 #undef SDANY
 
 // Scale down by even scale factor.
@@ -322,6 +342,18 @@ SDAANY(ScaleARGBRowDownEvenBox_Any_NEON,
        4,
        3)
 #endif
+#ifdef HAS_SCALEARGBROWDOWNEVEN_MSA
+SDAANY(ScaleARGBRowDownEven_Any_MSA,
+       ScaleARGBRowDownEven_MSA,
+       ScaleARGBRowDownEven_C,
+       4,
+       3)
+SDAANY(ScaleARGBRowDownEvenBox_Any_MSA,
+       ScaleARGBRowDownEvenBox_MSA,
+       ScaleARGBRowDownEvenBox_C,
+       4,
+       3)
+#endif
 
 // Add rows box filter scale down.
 #define SAANY(NAMEANY, SCALEADDROW_SIMD, SCALEADDROW_C, MASK)          \

source/scale_argb.cc

@@ -92,6 +92,22 @@ static void ScaleARGBDown2(int src_width,
     }
   }
 #endif
+#if defined(HAS_SCALEARGBROWDOWN2_MSA)
+  if (TestCpuFlag(kCpuHasMSA)) {
+    ScaleARGBRowDown2 =
+        filtering == kFilterNone
+            ? ScaleARGBRowDown2_Any_MSA
+            : (filtering == kFilterLinear ? ScaleARGBRowDown2Linear_Any_MSA
+                                          : ScaleARGBRowDown2Box_Any_MSA);
+    if (IS_ALIGNED(dst_width, 4)) {
+      ScaleARGBRowDown2 =
+          filtering == kFilterNone
+              ? ScaleARGBRowDown2_MSA
+              : (filtering == kFilterLinear ? ScaleARGBRowDown2Linear_MSA
+                                            : ScaleARGBRowDown2Box_MSA);
+    }
+  }
+#endif
 
   if (filtering == kFilterLinear) {
     src_stride = 0;
@@ -203,6 +219,16 @@ static void ScaleARGBDownEven(int src_width,
     }
   }
 #endif
+#if defined(HAS_SCALEARGBROWDOWNEVEN_MSA)
+  if (TestCpuFlag(kCpuHasMSA)) {
+    ScaleARGBRowDownEven = filtering ? ScaleARGBRowDownEvenBox_Any_MSA
+                                     : ScaleARGBRowDownEven_Any_MSA;
+    if (IS_ALIGNED(dst_width, 4)) {
+      ScaleARGBRowDownEven =
+          filtering ? ScaleARGBRowDownEvenBox_MSA : ScaleARGBRowDownEven_MSA;
+    }
+  }
+#endif
 
   if (filtering == kFilterLinear) {
     src_stride = 0;

source/scale_msa.cc

+/*
+ *  Copyright 2016 The LibYuv 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 "libyuv/scale_row.h"
+
+// This module is for GCC MSA
+#if !defined(LIBYUV_DISABLE_MSA) && defined(__mips_msa)
+#include "libyuv/macros_msa.h"
+
+#ifdef __cplusplus
+namespace libyuv {
+extern "C" {
+#endif
+
+void ScaleARGBRowDown2_MSA(const uint8_t* src_argb,
+                           ptrdiff_t src_stride,
+                           uint8_t* dst_argb,
+                           int dst_width) {
+  int x;
+  v16u8 src0, src1, dst0;
+
+  for (x = 0; x < dst_width; x += 4) {
+    src0 = (v16u8)__msa_ld_b((v16i8*)src_argb, 0);
+    src1 = (v16u8)__msa_ld_b((v16i8*)src_argb, 16);
+    dst0 = (v16u8)__msa_pckod_w((v4i32)src1, (v4i32)src0);
+    ST_UB(dst0, dst_argb);
+    src_argb += 32;
+    dst_argb += 16;
+  }
+}
+
+void ScaleARGBRowDown2Linear_MSA(const uint8_t* src_argb,
+                                 ptrdiff_t src_stride,
+                                 uint8_t* dst_argb,
+                                 int dst_width) {
+  int x;
+  v16u8 src0, src1, vec0, vec1, dst0;
+
+  for (x = 0; x < dst_width; x += 4) {
+    src0 = (v16u8)__msa_ld_b((v16i8*)src_argb, 0);
+    src1 = (v16u8)__msa_ld_b((v16i8*)src_argb, 16);
+    vec0 = (v16u8)__msa_pckev_w((v4i32)src1, (v4i32)src0);
+    vec1 = (v16u8)__msa_pckod_w((v4i32)src1, (v4i32)src0);
+    dst0 = (v16u8)__msa_aver_u_b((v16u8)vec0, (v16u8)vec1);
+    ST_UB(dst0, dst_argb);
+    src_argb += 32;
+    dst_argb += 16;
+  }
+}
+
+void ScaleARGBRowDown2Box_MSA(const uint8_t* src_argb,
+                              ptrdiff_t src_stride,
+                              uint8_t* dst_argb,
+                              int dst_width) {
+  int x;
+  const uint8_t* s = src_argb;
+  const uint8_t* t = src_argb + src_stride;
+  v16u8 src0, src1, src2, src3, vec0, vec1, vec2, vec3, dst0;
+  v8u16 reg0, reg1, reg2, reg3;
+  v16i8 shuffler = {0, 4, 1, 5, 2, 6, 3, 7, 8, 12, 9, 13, 10, 14, 11, 15};
+
+  for (x = 0; x < dst_width; x += 4) {
+    src0 = (v16u8)__msa_ld_b((v16i8*)s, 0);
+    src1 = (v16u8)__msa_ld_b((v16i8*)s, 16);
+    src2 = (v16u8)__msa_ld_b((v16i8*)t, 0);
+    src3 = (v16u8)__msa_ld_b((v16i8*)t, 16);
+    vec0 = (v16u8)__msa_vshf_b(shuffler, (v16i8)src0, (v16i8)src0);
+    vec1 = (v16u8)__msa_vshf_b(shuffler, (v16i8)src1, (v16i8)src1);
+    vec2 = (v16u8)__msa_vshf_b(shuffler, (v16i8)src2, (v16i8)src2);
+    vec3 = (v16u8)__msa_vshf_b(shuffler, (v16i8)src3, (v16i8)src3);
+    reg0 = __msa_hadd_u_h(vec0, vec0);
+    reg1 = __msa_hadd_u_h(vec1, vec1);
+    reg2 = __msa_hadd_u_h(vec2, vec2);
+    reg3 = __msa_hadd_u_h(vec3, vec3);
+    reg0 += reg2;
+    reg1 += reg3;
+    reg0 = (v8u16)__msa_srari_h((v8i16)reg0, 2);
+    reg1 = (v8u16)__msa_srari_h((v8i16)reg1, 2);
+    dst0 = (v16u8)__msa_pckev_b((v16i8)reg1, (v16i8)reg0);
+    ST_UB(dst0, dst_argb);
+    s += 32;
+    t += 32;
+    dst_argb += 16;
+  }
+}
+
+void ScaleARGBRowDownEven_MSA(const uint8_t* src_argb,
+                              ptrdiff_t src_stride,
+                              int32_t src_stepx,
+                              uint8_t* dst_argb,
+                              int dst_width) {
+  int x;
+  int32_t stepx = src_stepx * 4;
+  int32_t data0, data1, data2, data3;
+
+  for (x = 0; x < dst_width; x += 4) {
+    data0 = LW(src_argb);
+    data1 = LW(src_argb + stepx);
+    data2 = LW(src_argb + stepx * 2);
+    data3 = LW(src_argb + stepx * 3);
+    SW(data0, dst_argb);
+    SW(data1, dst_argb + 4);
+    SW(data2, dst_argb + 8);
+    SW(data3, dst_argb + 12);
+    src_argb += stepx * 4;
+    dst_argb += 16;
+  }
+}
+
+void ScaleARGBRowDownEvenBox_MSA(const uint8* src_argb,
+                                 ptrdiff_t src_stride,
+                                 int src_stepx,
+                                 uint8* dst_argb,
+                                 int dst_width) {
+  int x;
+  const uint8* nxt_argb = src_argb + src_stride;
+  int32_t stepx = src_stepx * 4;
+  int64_t data0, data1, data2, data3;
+  v16u8 src0 = {0}, src1 = {0}, src2 = {0}, src3 = {0};
+  v16u8 vec0, vec1, vec2, vec3;
+  v8u16 reg0, reg1, reg2, reg3, reg4, reg5, reg6, reg7;
+  v16u8 dst0;
+
+  for (x = 0; x < dst_width; x += 4) {
+    data0 = LD(src_argb);
+    data1 = LD(src_argb + stepx);
+    data2 = LD(src_argb + stepx * 2);
+    data3 = LD(src_argb + stepx * 3);
+    src0 = (v16u8)__msa_insert_d((v2i64)src0, 0, data0);
+    src0 = (v16u8)__msa_insert_d((v2i64)src0, 1, data1);
+    src1 = (v16u8)__msa_insert_d((v2i64)src1, 0, data2);
+    src1 = (v16u8)__msa_insert_d((v2i64)src1, 1, data3);
+    data0 = LD(nxt_argb);
+    data1 = LD(nxt_argb + stepx);
+    data2 = LD(nxt_argb + stepx * 2);
+    data3 = LD(nxt_argb + stepx * 3);
+    src2 = (v16u8)__msa_insert_d((v2i64)src2, 0, data0);
+    src2 = (v16u8)__msa_insert_d((v2i64)src2, 1, data1);
+    src3 = (v16u8)__msa_insert_d((v2i64)src3, 0, data2);
+    src3 = (v16u8)__msa_insert_d((v2i64)src3, 1, data3);
+    vec0 = (v16u8)__msa_ilvr_b((v16i8)src2, (v16i8)src0);
+    vec1 = (v16u8)__msa_ilvr_b((v16i8)src3, (v16i8)src1);
+    vec2 = (v16u8)__msa_ilvl_b((v16i8)src2, (v16i8)src0);
+    vec3 = (v16u8)__msa_ilvl_b((v16i8)src3, (v16i8)src1);
+    reg0 = __msa_hadd_u_h(vec0, vec0);
+    reg1 = __msa_hadd_u_h(vec1, vec1);
+    reg2 = __msa_hadd_u_h(vec2, vec2);
+    reg3 = __msa_hadd_u_h(vec3, vec3);
+    reg4 = (v8u16)__msa_pckev_d((v2i64)reg2, (v2i64)reg0);
+    reg5 = (v8u16)__msa_pckev_d((v2i64)reg3, (v2i64)reg1);
+    reg6 = (v8u16)__msa_pckod_d((v2i64)reg2, (v2i64)reg0);
+    reg7 = (v8u16)__msa_pckod_d((v2i64)reg3, (v2i64)reg1);
+    reg4 += reg6;
+    reg5 += reg7;
+    reg4 = (v8u16)__msa_srari_h((v8i16)reg4, 2);
+    reg5 = (v8u16)__msa_srari_h((v8i16)reg5, 2);
+    dst0 = (v16u8)__msa_pckev_b((v16i8)reg5, (v16i8)reg4);
+    ST_UB(dst0, dst_argb);
+    src_argb += stepx * 4;
+    nxt_argb += stepx * 4;
+    dst_argb += 16;
+  }
+}
+
+#ifdef __cplusplus
+}  // extern "C"
+}  // namespace libyuv
+#endif
+
+#endif  // !defined(LIBYUV_DISABLE_MSA) && defined(__mips_msa)