ARGBExtractAlpha 16 pixels at a time for ARM

source/row_neon64.cc

 /*
  *  Copyright 2014 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 432 matching lines @@
     : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v20",
       "v21", "v22", "v23", "v24", "v25", "v26", "v27", "v28", "v29", "v30"
   );
 }
 #endif  // HAS_I422TOARGB4444ROW_NEON
 
 #ifdef HAS_I400TOARGBROW_NEON
 void I400ToARGBRow_NEON(const uint8* src_y,
                         uint8* dst_argb,
                         int width) {
-  int64 width64 = (int64)(width);
   asm volatile (
     YUVTORGB_SETUP
     "movi       v23.8b, #255                   \n"
   "1:                                          \n"
     READYUV400
     YUVTORGB(v22, v21, v20)
     "subs       %w2, %w2, #8                   \n"
     MEMACCESS(1)
     "st4        {v20.8b,v21.8b,v22.8b,v23.8b}, [%1], #32     \n"
     "b.gt       1b                             \n"
     : "+r"(src_y),     // %0
       "+r"(dst_argb),  // %1
-      "+r"(width64)    // %2
+      "+r"(width)      // %2
     : [kUVToRB]"r"(&kYuvI601Constants.kUVToRB),
       [kUVToG]"r"(&kYuvI601Constants.kUVToG),
       [kUVBiasBGR]"r"(&kYuvI601Constants.kUVBiasBGR),
       [kYToRgb]"r"(&kYuvI601Constants.kYToRgb)
     : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v20",
       "v21", "v22", "v23", "v24", "v25", "v26", "v27", "v28", "v29", "v30"
   );
 }
 #endif  // HAS_I400TOARGBROW_NEON
 
@@ skipping 920 matching lines @@
   : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7"
   );
 }
 #endif  // HAS_ARGBTOYROW_NEON
 
 #ifdef HAS_ARGBEXTRACTALPHAROW_NEON
 void ARGBExtractAlphaRow_NEON(const uint8* src_argb, uint8* dst_a, int width) {
   asm volatile (
   "1:                                          \n"
     MEMACCESS(0)
-    "ld4        {v0.8b,v1.8b,v2.8b,v3.8b}, [%0], #32 \n"  // load row 8 pixels
-    "subs       %w2, %w2, #8                   \n"  // 8 processed per loop
+    "ld4        {v0.16b,v1.16b,v2.16b,v3.16b}, [%0], #64 \n"  // load row 16 pixels
+    "subs       %w2, %w2, #16                  \n"  // 16 processed per loop
     MEMACCESS(1)
-    "st1        {v3.8b}, [%1], #8              \n"  // store 8 A's.
+    "st1        {v3.16b}, [%1], #16            \n"  // store 16 A's.
     "b.gt       1b                             \n"
   : "+r"(src_argb),   // %0
     "+r"(dst_a),      // %1
     "+r"(width)       // %2
   :
   : "cc", "memory", "v0", "v1", "v2", "v3"  // Clobber List
   );
 }
 #endif  // HAS_ARGBEXTRACTALPHAROW_NEON
 
@@ skipping 1550 matching lines @@
   : "cc", "memory", "v0", "v1", "v2", "v3"  // Clobber List
   );
 }
 #endif  // HAS_SOBELYROW_NEON
 #endif  // !defined(LIBYUV_DISABLE_NEON) && defined(__aarch64__)
 
 #ifdef __cplusplus
 }  // extern "C"
 }  // namespace libyuv
 #endif