scale test

source/row_win.cc

 /*
  *  Copyright 2011 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.
  */
 
@@ skipping 2154 matching lines @@
 
     pop        ebp
     pop        edi
     pop        esi
     vzeroupper
     ret
   }
 }
 #endif  // HAS_I422TOARGBMATRIXROW_AVX2
 
-#ifdef HAS_I444TOARGBROW_AVX2
+#ifdef HAS_I444TOARGBMATRIXROW_AVX2
 // 16 pixels
 // 16 UV values with 16 Y producing 16 ARGB (64 bytes).
 __declspec(naked)
-void I444ToARGBRow_AVX2(const uint8* y_buf,
-                        const uint8* u_buf,
-                        const uint8* v_buf,
-                        uint8* dst_argb,
-                        int width) {
+void I444ToARGBMatrixRow_AVX2(const uint8* y_buf,
+                              const uint8* u_buf,
+                              const uint8* v_buf,
+                              uint8* dst_argb,
+                              struct YuvConstants* YuvConstants,
+                              int width) {
   __asm {
     push       esi
     push       edi
-    mov        eax, [esp + 8 + 4]   // Y
-    mov        esi, [esp + 8 + 8]   // U
-    mov        edi, [esp + 8 + 12]  // V
-    mov        edx, [esp + 8 + 16]  // argb
-    mov        ecx, [esp + 8 + 20]  // width
+    push       ebp
+    mov        eax, [esp + 12 + 4]   // Y
+    mov        esi, [esp + 12 + 8]   // U
+    mov        edi, [esp + 12 + 12]  // V
+    mov        edx, [esp + 12 + 16]  // argb
+    mov        ebp, [esp + 12 + 20]  // YuvConstants
+    mov        ecx, [esp + 12 + 24]  // width
     sub        edi, esi
     vpcmpeqb   ymm5, ymm5, ymm5     // generate 0xffffffffffffffff for alpha
-
  convertloop:
     READYUV444_AVX2
-    YUVTORGB_AVX2(kYuvConstants)
+    YUVTORGB_AVX2(ebp)
     STOREARGB_AVX2
 
     sub        ecx, 16
     jg         convertloop
 
+    pop        ebp
     pop        edi
     pop        esi
     vzeroupper
     ret
   }
 }
-#endif  // HAS_I444TOARGBROW_AVX2
+#endif  // HAS_I444TOARGBMATRIXROW_AVX2
+
+#ifdef HAS_I444TOABGRMATRIXROW_AVX2
+// 16 pixels
+// 16 UV values with 16 Y producing 16 ABGR (64 bytes).
+__declspec(naked)
+void I444ToABGRMatrixRow_AVX2(const uint8* y_buf,
+                              const uint8* u_buf,
+                              const uint8* v_buf,
+                              uint8* dst_abgr,
+                              struct YuvConstants* YuvConstants,
+                              int width) {
+  __asm {
+    push       esi
+    push       edi
+    push       ebp
+    mov        eax, [esp + 12 + 4]   // Y
+    mov        esi, [esp + 12 + 8]   // U
+    mov        edi, [esp + 12 + 12]  // V
+    mov        edx, [esp + 12 + 16]  // abgr
+    mov        ebp, [esp + 12 + 20]  // YuvConstants
+    mov        ecx, [esp + 12 + 24]  // width
+    sub        edi, esi
+    vpcmpeqb   ymm5, ymm5, ymm5     // generate 0xffffffffffffffff for alpha
+ convertloop:
+    READYUV444_AVX2
+    YUVTORGB_AVX2(ebp)
+    STOREABGR_AVX2
+
+    sub        ecx, 16
+    jg         convertloop
+
+    pop        ebp
+    pop        edi
+    pop        esi
+    vzeroupper
+    ret
+  }
+}
+#endif  // HAS_I444TOABGRMATRIXROW_AVX2
 
 #ifdef HAS_I411TOARGBROW_AVX2
 // 16 pixels
 // 4 UV values upsampled to 16 UV, mixed with 16 Y producing 16 ARGB (64 bytes).
 __declspec(naked)
 void I411ToARGBRow_AVX2(const uint8* y_buf,
                         const uint8* u_buf,
                         const uint8* v_buf,
                         uint8* dst_argb,
                         int width) {
@@ skipping 381 matching lines @@
     __asm por        xmm3, xmm2    /* BG */                                    \
     __asm por        xmm1, xmm3    /* BGR */                                   \
     __asm packssdw   xmm0, xmm1                                                \
     __asm movdqu     0[edx], xmm0  /* store 8 pixels of RGB565 */              \
     __asm lea        edx, [edx + 16]                                           \
   }
 
 // 8 pixels.
 // 8 UV values, mixed with 8 Y producing 8 ARGB (32 bytes).
 __declspec(naked)
-void I444ToARGBRow_SSSE3(const uint8* y_buf,
-                         const uint8* u_buf,
-                         const uint8* v_buf,
-                         uint8* dst_argb,
-                         int width) {
+void I444ToARGBMatrixRow_SSSE3(const uint8* y_buf,
+                               const uint8* u_buf,
+                               const uint8* v_buf,
+                               uint8* dst_argb,
+                               struct YuvConstants* YuvConstants,
+                               int width) {
   __asm {
     push       esi
     push       edi
-    mov        eax, [esp + 8 + 4]   // Y
-    mov        esi, [esp + 8 + 8]   // U
-    mov        edi, [esp + 8 + 12]  // V
-    mov        edx, [esp + 8 + 16]  // argb
-    mov        ecx, [esp + 8 + 20]  // width
+    push       ebp
+    mov        eax, [esp + 12 + 4]   // Y
+    mov        esi, [esp + 12 + 8]   // U
+    mov        edi, [esp + 12 + 12]  // V
+    mov        edx, [esp + 12 + 16]  // argb
+    mov        ebp, [esp + 12 + 20]  // YuvConstants
+    mov        ecx, [esp + 12 + 24]  // width
     sub        edi, esi
-    pcmpeqb    xmm5, xmm5           // generate 0xffffffff for alpha
+    pcmpeqb    xmm5, xmm5            // generate 0xffffffff for alpha
 
  convertloop:
     READYUV444
-    YUVTORGB(kYuvConstants)
+    YUVTORGB(ebp)
     STOREARGB
 
     sub        ecx, 8
     jg         convertloop
 
+    pop        ebp
     pop        edi
     pop        esi
     ret
+  }
+}
+
+// 8 pixels.
+// 8 UV values, mixed with 8 Y producing 8 ABGR (32 bytes).
+__declspec(naked)
+void I444ToABGRMatrixRow_SSSE3(const uint8* y_buf,
+                               const uint8* u_buf,
+                               const uint8* v_buf,
+                               uint8* dst_abgr,
+                               struct YuvConstants* YuvConstants,
+                               int width) {
+  __asm {
+    push       esi
+    push       edi
+    push       ebp
+    mov        eax, [esp + 12 + 4]   // Y
+    mov        esi, [esp + 12 + 8]   // U
+    mov        edi, [esp + 12 + 12]  // V
+    mov        edx, [esp + 12 + 16]  // abgr
+    mov        ebp, [esp + 12 + 20]  // YuvConstants
+    mov        ecx, [esp + 12 + 24]  // width
+    sub        edi, esi
+    pcmpeqb    xmm5, xmm5            // generate 0xffffffff for alpha
+
+ convertloop:
+    READYUV444
+    YUVTORGB(ebp)
+    STOREABGR
+
+    sub        ecx, 8
+    jg         convertloop
+
+    pop        ebp
+    pop        edi
+    pop        esi
+    ret
   }
 }
 
 // 8 pixels.
 // 4 UV values upsampled to 8 UV, mixed with 8 Y producing 8 RGB24 (24 bytes).
 __declspec(naked)
 void I422ToRGB24Row_SSSE3(const uint8* y_buf,
                           const uint8* u_buf,
                           const uint8* v_buf,
                           uint8* dst_rgb24,
@@ skipping 3716 matching lines @@
 }
 #endif  // HAS_ARGBLUMACOLORTABLEROW_SSSE3
 
 #endif  // defined(_M_X64)
 #endif  // !defined(LIBYUV_DISABLE_X86) && (defined(_M_IX86) || defined(_M_X64))
 
 #ifdef __cplusplus
 }  // extern "C"
 }  // namespace libyuv
 #endif