[x265] [PATCH 101 of 307] x86: revoke some changes in ipfilter16

mythreyi at multicorewareinc.com mythreyi at multicorewareinc.com
Sat Apr 7 04:31:39 CEST 2018


# HG changeset patch
# User Aasaipriya Chandran <aasaipriya at multicorewareinc.com>
# Date 1522962868 25200
#      Thu Apr 05 14:14:28 2018 -0700
# Node ID 1a31df496144c526fd5eba9d960bb286a81ae2d5
# Parent  562c00d2153193eec85ab907b60eeb5aca7cc609
x86: revoke some changes in ipfilter16

diff -r 562c00d21531 -r 1a31df496144 source/common/x86/h-ipfilter16.asm
--- a/source/common/x86/h-ipfilter16.asm	Mon Aug 28 14:59:38 2017 +0530
+++ b/source/common/x86/h-ipfilter16.asm	Thu Apr 05 14:14:28 2018 -0700
@@ -47,7 +47,7 @@
 
 h_pd_524800:        times 8 dd 524800
                                     
-tab_LumaCoeff:    dw   0, 0,  0,  64,  0,   0,  0,  0
+h_tab_LumaCoeff:    dw   0, 0,  0,  64,  0,   0,  0,  0
                   dw  -1, 4, -10, 58,  17, -5,  1,  0
                   dw  -1, 4, -11, 40,  40, -11, 4, -1
                   dw   0, 1, -5,  17,  58, -10, 4, -1
@@ -207,10 +207,10 @@
     add         r3d,    r3d
 
 %ifdef PIC
-    lea         r6,     [tab_LumaCoeff]
+    lea         r6,     [h_tab_LumaCoeff]
     mova        m0,     [r6 + r4]
 %else
-    mova        m0,     [tab_LumaCoeff + r4]
+    mova        m0,     [h_tab_LumaCoeff + r4]
 %endif
 
 %ifidn %3, pp
@@ -625,10 +625,10 @@
     add         r3, r3
 
 %ifdef PIC
-    lea         r6, [tab_LumaCoeff]
+    lea         r6, [h_tab_LumaCoeff]
     mova        m0, [r6 + r4]
 %else
-    mova        m0, [tab_LumaCoeff + r4]
+    mova        m0, [h_tab_LumaCoeff + r4]
 %endif
 
 %ifidn %3, pp
@@ -712,10 +712,10 @@
     shl         r4d, 4
 
 %ifdef PIC
-    lea         r6, [tab_LumaCoeff]
+    lea         r6, [h_tab_LumaCoeff]
     mova        m0, [r6 + r4]
 %else
-    mova        m0, [tab_LumaCoeff + r4]
+    mova        m0, [h_tab_LumaCoeff + r4]
 %endif
 
 %ifidn %3, pp
@@ -815,10 +815,10 @@
     shl         r4d, 4
 
 %ifdef PIC
-    lea         r6, [tab_LumaCoeff]
+    lea         r6, [h_tab_LumaCoeff]
     mova        m0, [r6 + r4]
 %else
-    mova        m0, [tab_LumaCoeff + r4]
+    mova        m0, [h_tab_LumaCoeff + r4]
 %endif
 %ifidn %3, pp
     mova        m1, [INTERP_OFFSET_PP]
@@ -936,10 +936,10 @@
     shl         r4d, 4
 
 %ifdef PIC
-    lea         r6, [tab_LumaCoeff]
+    lea         r6, [h_tab_LumaCoeff]
     mova        m0, [r6 + r4]
 %else
-    mova        m0, [tab_LumaCoeff + r4]
+    mova        m0, [h_tab_LumaCoeff + r4]
 %endif
 
 %ifidn %3, pp
@@ -1132,10 +1132,10 @@
     shl         r4d, 4
 
 %ifdef PIC
-    lea         r6, [tab_LumaCoeff]
+    lea         r6, [h_tab_LumaCoeff]
     mova        m0, [r6 + r4]
 %else
-    mova        m0, [tab_LumaCoeff + r4]
+    mova        m0, [h_tab_LumaCoeff + r4]
 %endif
 %ifidn %3, pp
     mova        m1, [pd_32]
@@ -1307,12 +1307,12 @@
     mov              r4d, r4m
     shl              r4d, 4
 %ifdef PIC
-    lea              r5, [tab_LumaCoeff]
+    lea              r5, [h_tab_LumaCoeff]
     vpbroadcastq     m0, [r5 + r4]
     vpbroadcastq     m1, [r5 + r4 + 8]
 %else
-    vpbroadcastq     m0, [tab_LumaCoeff + r4]
-    vpbroadcastq     m1, [tab_LumaCoeff + r4 + 8]
+    vpbroadcastq     m0, [h_tab_LumaCoeff + r4]
+    vpbroadcastq     m1, [h_tab_LumaCoeff + r4 + 8]
 %endif
     lea              r6, [pw_pixel_max]
     mova             m3, [interp8_hpp_shuf]
@@ -1385,11 +1385,11 @@
     mov              r4d, r4m
     shl              r4d, 4
 %ifdef PIC
-    lea              r5, [tab_LumaCoeff]
+    lea              r5, [h_tab_LumaCoeff]
     vpbroadcastq     m0, [r5 + r4]
     vpbroadcastq     m1, [r5 + r4 + 8]
 %else
-    vpbroadcastq     m0, [tab_LumaCoeff + r4]
+    vpbroadcastq     m0, [h_tab_LumaCoeff + r4]
     vpbroadcastq     m1, [h_ab_LumaCoeff + r4 + 8]
 %endif
     mova             m3, [interp8_hpp_shuf]
@@ -1481,12 +1481,12 @@
     mov              r4d, r4m
     shl              r4d, 4
 %ifdef PIC
-    lea              r5, [tab_LumaCoeff]
+    lea              r5, [h_tab_LumaCoeff]
     vpbroadcastq     m0, [r5 + r4]
     vpbroadcastq     m1, [r5 + r4 + 8]
 %else
-    vpbroadcastq     m0, [tab_LumaCoeff + r4]
-    vpbroadcastq     m1, [tab_LumaCoeff + r4 + 8]
+    vpbroadcastq     m0, [h_tab_LumaCoeff + r4]
+    vpbroadcastq     m1, [h_tab_LumaCoeff + r4 + 8]
 %endif
     mova             m3, [interp8_hpp_shuf]
     mova             m7, [pd_32]
@@ -1579,12 +1579,12 @@
     mov              r4d, r4m
     shl              r4d, 4
 %ifdef PIC
-    lea              r5, [tab_LumaCoeff]
+    lea              r5, [h_tab_LumaCoeff]
     vpbroadcastq     m0, [r5 + r4]
     vpbroadcastq     m1, [r5 + r4 + 8]
 %else
-    vpbroadcastq     m0, [tab_LumaCoeff + r4]
-    vpbroadcastq     m1, [tab_LumaCoeff + r4 + 8]
+    vpbroadcastq     m0, [h_tab_LumaCoeff + r4]
+    vpbroadcastq     m1, [h_tab_LumaCoeff + r4 + 8]
 %endif
     mova             m3, [interp8_hpp_shuf]
     mova             m7, [pd_32]
@@ -1684,12 +1684,12 @@
     mov              r4d, r4m
     shl              r4d, 4
 %ifdef PIC
-    lea              r5, [tab_LumaCoeff]
+    lea              r5, [h_tab_LumaCoeff]
     vpbroadcastq     m0, [r5 + r4]
     vpbroadcastq     m1, [r5 + r4 + 8]
 %else
-    vpbroadcastq     m0, [tab_LumaCoeff + r4]
-    vpbroadcastq     m1, [tab_LumaCoeff + r4 + 8]
+    vpbroadcastq     m0, [h_tab_LumaCoeff + r4]
+    vpbroadcastq     m1, [h_tab_LumaCoeff + r4 + 8]
 %endif
     mova             m3, [interp8_hpp_shuf]
     mova             m7, [pd_32]
@@ -1774,12 +1774,12 @@
     mov              r4d, r4m
     shl              r4d, 4
 %ifdef PIC
-    lea              r5, [tab_LumaCoeff]
+    lea              r5, [h_tab_LumaCoeff]
     vpbroadcastq     m0, [r5 + r4]
     vpbroadcastq     m1, [r5 + r4 + 8]
 %else
-    vpbroadcastq     m0, [tab_LumaCoeff + r4]
-    vpbroadcastq     m1, [tab_LumaCoeff + r4 + 8]
+    vpbroadcastq     m0, [h_tab_LumaCoeff + r4]
+    vpbroadcastq     m1, [h_tab_LumaCoeff + r4 + 8]
 %endif
     mova             m3, [interp8_hpp_shuf]
     mova             m7, [pd_32]
@@ -1892,12 +1892,12 @@
     mov              r4d, r4m
     shl              r4d, 4
 %ifdef PIC
-    lea              r5, [tab_LumaCoeff]
+    lea              r5, [h_tab_LumaCoeff]
     vpbroadcastq     m0, [r5 + r4]
     vpbroadcastq     m1, [r5 + r4 + 8]
 %else
-    vpbroadcastq     m0, [tab_LumaCoeff + r4]
-    vpbroadcastq     m1, [tab_LumaCoeff + r4 + 8]
+    vpbroadcastq     m0, [h_tab_LumaCoeff + r4]
+    vpbroadcastq     m1, [h_tab_LumaCoeff + r4 + 8]
 %endif
     mova             m3, [interp8_hpp_shuf]
     mova             m7, [pd_32]
@@ -2018,12 +2018,12 @@
     add                         r3d,               r3d
 
 %ifdef PIC
-    lea                         r6,                [tab_LumaCoeff]
+    lea                         r6,                [h_tab_LumaCoeff]
     lea                         r4,                [r4 * 8]
     vbroadcasti128              m0,                [r6 + r4 * 2]
 %else
     lea                         r4,                [r4 * 8]
-    vbroadcasti128              m0,                [tab_LumaCoeff + r4 * 2]
+    vbroadcasti128              m0,                [h_tab_LumaCoeff + r4 * 2]
 %endif
 
     vbroadcasti128              m2,                [INTERP_OFFSET_PS]
@@ -2129,12 +2129,12 @@
     mov                 r5d, r5m
     shl                 r4d, 4
 %ifdef PIC
-    lea                 r6, [tab_LumaCoeff]
+    lea                 r6, [h_tab_LumaCoeff]
     vpbroadcastq        m0, [r6 + r4]
     vpbroadcastq        m1, [r6 + r4 + 8]
 %else
-    vpbroadcastq        m0, [tab_LumaCoeff + r4]
-    vpbroadcastq        m1, [tab_LumaCoeff + r4 + 8]
+    vpbroadcastq        m0, [h_tab_LumaCoeff + r4]
+    vpbroadcastq        m1, [h_tab_LumaCoeff + r4 + 8]
 %endif
     mova                m3, [interp8_hpp_shuf]
     vbroadcasti128      m2, [INTERP_OFFSET_PS]
@@ -2197,12 +2197,12 @@
     mov                 r5d, r5m
     shl                 r4d, 4
 %ifdef PIC
-    lea                 r6, [tab_LumaCoeff]
+    lea                 r6, [h_tab_LumaCoeff]
     vpbroadcastq        m0, [r6 + r4]
     vpbroadcastq        m1, [r6 + r4 + 8]
 %else
-    vpbroadcastq        m0, [tab_LumaCoeff + r4]
-    vpbroadcastq        m1, [tab_LumaCoeff + r4 + 8]
+    vpbroadcastq        m0, [h_tab_LumaCoeff + r4]
+    vpbroadcastq        m1, [h_tab_LumaCoeff + r4 + 8]
 %endif
     mova                m3, [interp8_hpp_shuf]
     vbroadcasti128      m2, [INTERP_OFFSET_PS]
@@ -2376,12 +2376,12 @@
     mov                 r5d, r5m
     shl                 r4d, 4
 %ifdef PIC
-    lea                 r6, [tab_LumaCoeff]
+    lea                 r6, [h_tab_LumaCoeff]
     vpbroadcastq        m0, [r6 + r4]
     vpbroadcastq        m1, [r6 + r4 + 8]
 %else
-    vpbroadcastq        m0, [tab_LumaCoeff + r4]
-    vpbroadcastq        m1, [tab_LumaCoeff + r4 + 8]
+    vpbroadcastq        m0, [h_tab_LumaCoeff + r4]
+    vpbroadcastq        m1, [h_tab_LumaCoeff + r4 + 8]
 %endif
     mova                m3, [interp8_hpp_shuf]
     vbroadcasti128      m2, [INTERP_OFFSET_PS]
@@ -2469,12 +2469,12 @@
     mov                 r5d, r5m
     shl                 r4d, 4
 %ifdef PIC
-    lea                 r6, [tab_LumaCoeff]
+    lea                 r6, [h_tab_LumaCoeff]
     vpbroadcastq        m0, [r6 + r4]
     vpbroadcastq        m1, [r6 + r4 + 8]
 %else
-    vpbroadcastq        m0, [tab_LumaCoeff + r4]
-    vpbroadcastq        m1, [tab_LumaCoeff + r4 + 8]
+    vpbroadcastq        m0, [h_tab_LumaCoeff + r4]
+    vpbroadcastq        m1, [h_tab_LumaCoeff + r4 + 8]
 %endif
     mova                m3, [interp8_hpp_shuf]
     vbroadcasti128      m2, [INTERP_OFFSET_PS]
diff -r 562c00d21531 -r 1a31df496144 source/common/x86/h4-ipfilter16.asm
--- a/source/common/x86/h4-ipfilter16.asm	Mon Aug 28 14:59:38 2017 +0530
+++ b/source/common/x86/h4-ipfilter16.asm	Thu Apr 05 14:14:28 2018 -0700
@@ -52,7 +52,7 @@
 
 tab_Tm16:         db 0, 1, 2, 3, 4,  5,  6, 7, 2, 3, 4,  5, 6, 7, 8, 9
 
-tab_ChromaCoeff:  dw  0, 64,  0,  0
+h4_tab_ChromaCoeff:  dw  0, 64,  0,  0
                   dw -2, 58, 10, -2
                   dw -4, 54, 16, -2
                   dw -6, 46, 28, -4
@@ -279,10 +279,10 @@
     add         r4d,    r4d
 
 %ifdef PIC
-    lea         r6,     [tab_ChromaCoeff]
+    lea         r6,     [h4_tab_ChromaCoeff]
     movddup     m0,     [r6 + r4 * 4]
 %else
-    movddup     m0,     [tab_ChromaCoeff + r4 * 4]
+    movddup     m0,     [h4_tab_ChromaCoeff + r4 * 4]
 %endif
 
 %ifidn %3, ps
@@ -530,10 +530,10 @@
     add         r4d,      r4d
 
 %ifdef PIC
-    lea         r%6,      [tab_ChromaCoeff]
+    lea         r%6,      [h4_tab_ChromaCoeff]
     movh        m0,       [r%6 + r4 * 4]
 %else
-    movh        m0,       [tab_ChromaCoeff + r4 * 4]
+    movh        m0,       [h4_tab_ChromaCoeff + r4 * 4]
 %endif
 
     punpcklqdq  m0,       m0
@@ -1129,10 +1129,10 @@
     add         r4d,        r4d
 
 %ifdef PIC
-    lea         r%4,       [tab_ChromaCoeff]
+    lea         r%4,       [h4_tab_ChromaCoeff]
     movh        m0,       [r%4 + r4 * 4]
 %else
-    movh        m0,       [tab_ChromaCoeff + r4 * 4]
+    movh        m0,       [h4_tab_ChromaCoeff + r4 * 4]
 %endif
 
     punpcklqdq  m0,       m0
@@ -1246,10 +1246,10 @@
     sub             r0, 2
     mov             r4d, r4m
 %ifdef PIC
-    lea             r5, [tab_ChromaCoeff]
+    lea             r5, [h4_tab_ChromaCoeff]
     vpbroadcastq    m0, [r5 + r4 * 8]
 %else
-    vpbroadcastq    m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq    m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova            m1, [h4_interp8_hpp_shuf]
     vpbroadcastd    m2, [pd_32]
@@ -1314,10 +1314,10 @@
     sub             r0, 2
     mov             r4d, r4m
 %ifdef PIC
-    lea             r5, [tab_ChromaCoeff]
+    lea             r5, [h4_tab_ChromaCoeff]
     vpbroadcastq    m0, [r5 + r4 * 8]
 %else
-    vpbroadcastq    m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq    m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova            m1, [h4_interp8_hpp_shuf]
     vpbroadcastd    m2, [pd_32]
@@ -1370,10 +1370,10 @@
     sub             r0, 2
     mov             r4d, r4m
 %ifdef PIC
-    lea             r5, [tab_ChromaCoeff]
+    lea             r5, [h4_tab_ChromaCoeff]
     vpbroadcastq    m0, [r5 + r4 * 8]
 %else
-    vpbroadcastq    m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq    m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova            m1, [h4_interp8_hpp_shuf]
     vpbroadcastd    m2, [pd_32]
@@ -1432,10 +1432,10 @@
     sub             r0, 2
     mov             r4d, r4m
 %ifdef PIC
-    lea             r5, [tab_ChromaCoeff]
+    lea             r5, [h4_tab_ChromaCoeff]
     vpbroadcastq    m0, [r5 + r4 * 8]
 %else
-    vpbroadcastq    m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq    m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova            m1, [h4_interp8_hpp_shuf]
     vpbroadcastd    m2, [pd_32]
@@ -1504,10 +1504,10 @@
     sub             r0, 2
     mov             r4d, r4m
 %ifdef PIC
-    lea             r5, [tab_ChromaCoeff]
+    lea             r5, [h4_tab_ChromaCoeff]
     vpbroadcastq    m0, [r5 + r4 * 8]
 %else
-    vpbroadcastq    m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq    m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova            m1, [h4_interp8_hpp_shuf]
     vpbroadcastd    m2, [pd_32]
@@ -1579,10 +1579,10 @@
     sub             r0, 2
     mov             r4d, r4m
 %ifdef PIC
-    lea             r5, [tab_ChromaCoeff]
+    lea             r5, [h4_tab_ChromaCoeff]
     vpbroadcastq    m0, [r5 + r4 * 8]
 %else
-    vpbroadcastq    m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq    m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova            m1, [h4_interp8_hpp_shuf]
     vpbroadcastd    m2, [pd_32]
@@ -1655,10 +1655,10 @@
     sub             r0, 2
     mov             r4d, r4m
 %ifdef PIC
-    lea             r5, [tab_ChromaCoeff]
+    lea             r5, [h4_tab_ChromaCoeff]
     vpbroadcastq    m0, [r5 + r4 * 8]
 %else
-    vpbroadcastq    m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq    m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova            m1, [h4_interp8_hpp_shuf]
     vpbroadcastd    m2, [pd_32]
@@ -1724,10 +1724,10 @@
     sub             r0, 2
     mov             r4d, r4m
 %ifdef PIC
-    lea             r5, [tab_ChromaCoeff]
+    lea             r5, [h4_tab_ChromaCoeff]
     vpbroadcastq    m0, [r5 + r4 * 8]
 %else
-    vpbroadcastq    m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq    m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova            m1, [h4_interp8_hpp_shuf]
     vpbroadcastd    m2, [pd_32]
@@ -1804,10 +1804,10 @@
     sub             r0, 2
     mov             r4d, r4m
 %ifdef PIC
-    lea             r5, [tab_ChromaCoeff]
+    lea             r5, [h4_tab_ChromaCoeff]
     vpbroadcastq    m0, [r5 + r4 * 8]
 %else
-    vpbroadcastq    m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq    m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova            m1, [h4_interp8_hpp_shuf]
     vpbroadcastd    m2, [pd_32]
@@ -1872,10 +1872,10 @@
     sub             r0, 2
     mov             r4d, r4m
 %ifdef PIC
-    lea             r5, [tab_ChromaCoeff]
+    lea             r5, [h4_tab_ChromaCoeff]
     vpbroadcastq    m0, [r5 + r4 * 8]
 %else
-    vpbroadcastq    m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq    m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova            m1, [h4_interp8_hpp_shuf]
     vpbroadcastd    m2, [pd_32]
@@ -1934,10 +1934,10 @@
     mov                 r5d, r5m
 
 %ifdef PIC
-    lea                 r6, [tab_ChromaCoeff]
+    lea                 r6, [h4_tab_ChromaCoeff]
     vpbroadcastq        m0, [r6 + r4 * 8]
 %else
-    vpbroadcastq        m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq        m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova                m3, [h4_interp8_hpp_shuf]
     vbroadcasti128      m2, [INTERP_OFFSET_PS]
@@ -1993,10 +1993,10 @@
     mov                 r5d, r5m
 
 %ifdef PIC
-    lea                 r6, [tab_ChromaCoeff]
+    lea                 r6, [h4_tab_ChromaCoeff]
     vpbroadcastq        m0, [r6 + r4 * 8]
 %else
-    vpbroadcastq        m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq        m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova                m3, [h4_interp8_hpp_shuf]
     vbroadcasti128      m2, [INTERP_OFFSET_PS]
@@ -2066,10 +2066,10 @@
     mov                 r5d, r5m
 
 %ifdef PIC
-    lea                 r6, [tab_ChromaCoeff]
+    lea                 r6, [h4_tab_ChromaCoeff]
     vpbroadcastq        m0, [r6 + r4 * 8]
 %else
-    vpbroadcastq        m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq        m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova                m3, [h4_interp8_hpp_shuf]
     vbroadcasti128      m2, [INTERP_OFFSET_PS]
@@ -2148,10 +2148,10 @@
     mov                 r5d, r5m
 
 %ifdef PIC
-    lea                 r6, [tab_ChromaCoeff]
+    lea                 r6, [h4_tab_ChromaCoeff]
     vpbroadcastq        m0, [r6 + r4 * 8]
 %else
-    vpbroadcastq        m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq        m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova                m3, [h4_interp8_hpp_shuf]
     vbroadcasti128      m2, [INTERP_OFFSET_PS]
@@ -2213,10 +2213,10 @@
     mov                 r5d, r5m
 
 %ifdef PIC
-    lea                 r6, [tab_ChromaCoeff]
+    lea                 r6, [h4_tab_ChromaCoeff]
     vpbroadcastq        m0, [r6 + r4 * 8]
 %else
-    vpbroadcastq        m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq        m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova                m3, [h4_interp8_hpp_shuf]
     vbroadcasti128      m2, [INTERP_OFFSET_PS]
@@ -2314,10 +2314,10 @@
     mov                 r5d, r5m
 
 %ifdef PIC
-    lea                 r6, [tab_ChromaCoeff]
+    lea                 r6, [h4_tab_ChromaCoeff]
     vpbroadcastq        m0, [r6 + r4 * 8]
 %else
-    vpbroadcastq        m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq        m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova                m3, [h4_interp8_hpp_shuf]
     vbroadcasti128      m2, [INTERP_OFFSET_PS]
@@ -2467,10 +2467,10 @@
     mov                 r5d, r5m
 
 %ifdef PIC
-    lea                 r6, [tab_ChromaCoeff]
+    lea                 r6, [h4_tab_ChromaCoeff]
     vpbroadcastq        m0, [r6 + r4 * 8]
 %else
-    vpbroadcastq        m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq        m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova                m3, [h4_interp8_hpp_shuf]
     vbroadcasti128      m2, [INTERP_OFFSET_PS]
@@ -2587,10 +2587,10 @@
     mov                 r5d, r5m
 
 %ifdef PIC
-    lea                 r6, [tab_ChromaCoeff]
+    lea                 r6, [h4_tab_ChromaCoeff]
     vpbroadcastq        m0, [r6 + r4 * 8]
 %else
-    vpbroadcastq        m0, [tab_ChromaCoeff + r4 * 8]
+    vpbroadcastq        m0, [h4_tab_ChromaCoeff + r4 * 8]
 %endif
     mova                m3, [h4_interp8_hpp_shuf]
     vbroadcasti128      m2, [INTERP_OFFSET_PS]
diff -r 562c00d21531 -r 1a31df496144 source/common/x86/ipfilter16.asm
--- a/source/common/x86/ipfilter16.asm	Mon Aug 28 14:59:38 2017 +0530
+++ b/source/common/x86/ipfilter16.asm	Thu Apr 05 14:14:28 2018 -0700
@@ -51,6 +51,20 @@
 tab_c_n8192:      times 8 dw -8192
 pd_524800:        times 8 dd 524800
 
+tab_ChromaCoeff:  dw  0, 64,  0,  0
+                  dw -2, 58, 10, -2
+                  dw -4, 54, 16, -2
+                  dw -6, 46, 28, -4
+                  dw -4, 36, 36, -4
+                  dw -4, 28, 46, -6
+                  dw -2, 16, 54, -4
+                  dw -2, 10, 58, -2
+				
+tab_LumaCoeff:    dw   0, 0,  0,  64,  0,   0,  0,  0
+                  dw  -1, 4, -10, 58,  17, -5,  1,  0
+                  dw  -1, 4, -11, 40,  40, -11, 4, -1
+                  dw   0, 1, -5,  17,  58, -10, 4, -1
+
 ALIGN 32
 tab_LumaCoeffV:   times 4 dw 0, 0
                   times 4 dw 0, 64


More information about the x265-devel mailing list