[x265] [PATCH] copy_cnt_4 avx2 asm code: nit, same speedup by sse version

[praveen at multicorewareinc.com]

# HG changeset patch
# User Praveen Tiwari
# Date 1410437024 -19800
# Node ID 9241634204a12babf8a2a90dc4f776646a9ddfb3
# Parent  c7d9506fe681e29b6315bf42ac898b46adb9b6a5
copy_cnt_4 avx2 asm code: nit, same speedup by sse version

diff -r c7d9506fe681 -r 9241634204a1 source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp	Thu Sep 11 17:20:05 2014 +0530
+++ b/source/common/x86/asm-primitives.cpp	Thu Sep 11 17:33:44 2014 +0530
@@ -1730,7 +1730,6 @@
         /* Need to update assembly code as per changed interface of the copy_cnt primitive, once
          * code is updated, avx2 version will be enabled */
 
-        // p.copy_cnt[BLOCK_4x4] = x265_copy_cnt_4_avx2;
         p.copy_cnt[BLOCK_8x8] = x265_copy_cnt_8_avx2;
         // p.copy_cnt[BLOCK_16x16] = x265_copy_cnt_16_avx2;
         // p.copy_cnt[BLOCK_32x32] = x265_copy_cnt_32_avx2;
diff -r c7d9506fe681 -r 9241634204a1 source/common/x86/blockcopy8.asm
--- a/source/common/x86/blockcopy8.asm	Thu Sep 11 17:20:05 2014 +0530
+++ b/source/common/x86/blockcopy8.asm	Thu Sep 11 17:33:44 2014 +0530
@@ -3988,34 +3988,6 @@
     RET
 
 
-INIT_YMM avx2
-cglobal copy_cnt_4, 3,3,3
-    add         r2d, r2d
-    xorpd       xm2, xm2
-
-    ; row 0 & 1
-    movq        xm0, [r1]
-    movhps      xm0, [r1 + r2]
-
-    ; row 2 & 3
-    movq        xm1, [r1 + r2 * 2]
-    lea         r2, [r2 * 3]
-    movhps      xm1, [r1 + r2]
-
-    vinserti128 m0, m0, xm1, 1
-    movu    [r0], m0
-
-    vextractf128 xm1, m0, 1
-    packsswb     xm0, xm1
-    pcmpeqb      xm0, xm2
-
-    ; get count
-    pmovmskb    eax, xm0
-    not         ax
-    popcnt      ax, ax
-    RET
-
-
 ;--------------------------------------------------------------------------------------
 ; uint32_t copy_cnt(int16_t *dst, int16_t *src, intptr_t stride);
 ;--------------------------------------------------------------------------------------