[x265] [PATCH] asm: 16bpp asm code for intra_pred_ang4_8 and intra_pred_ang4_9

yuvaraj at multicorewareinc.com yuvaraj at multicorewareinc.com
Mon Dec 9 13:17:31 CET 2013


# HG changeset patch
# User Yuvaraj Venkatesh <yuvaraj at multicorewareinc.com>
# Date 1386591441 -19800
#      Mon Dec 09 17:47:21 2013 +0530
# Node ID 9f68d1d7b3d903d4f72aa3451a83c2363a114dd9
# Parent  9ceeba76f4714972bd93bda3d1513e8bb68a2384
asm: 16bpp asm code for intra_pred_ang4_8 and intra_pred_ang4_9

diff -r 9ceeba76f471 -r 9f68d1d7b3d9 source/common/x86/asm-primitives.cpp
--- a/source/common/x86/asm-primitives.cpp	Mon Dec 09 17:08:00 2013 +0530
+++ b/source/common/x86/asm-primitives.cpp	Mon Dec 09 17:47:21 2013 +0530
@@ -691,6 +691,10 @@
         SETUP_INTRA_ANG4(5, 5, sse4);
         SETUP_INTRA_ANG4(6, 6, sse4);
         SETUP_INTRA_ANG4(7, 7, sse4);
+        SETUP_INTRA_ANG4(8, 8, sse4);
+        SETUP_INTRA_ANG4(9, 9, sse4);
+        SETUP_INTRA_ANG4(27, 9, sse4);
+        SETUP_INTRA_ANG4(28, 8, sse4);
         SETUP_INTRA_ANG4(29, 7, sse4);
         SETUP_INTRA_ANG4(30, 6, sse4);
         SETUP_INTRA_ANG4(31, 5, sse4);
diff -r 9ceeba76f471 -r 9f68d1d7b3d9 source/common/x86/intrapred16.asm
--- a/source/common/x86/intrapred16.asm	Mon Dec 09 17:08:00 2013 +0530
+++ b/source/common/x86/intrapred16.asm	Mon Dec 09 17:47:21 2013 +0530
@@ -557,3 +557,38 @@
     mova        m6, [r3 +  7 * 16]  ; [27]
     mova        m7, [r3 - 16 * 16]  ; [ 4]
     jmp         mangle(private_prefix %+ _ %+ intra_pred_ang4_3 %+ SUFFIX %+ .do_filter4x4)
+
+cglobal intra_pred_ang4_8, 3,4,8
+    cmp         r4m, byte 28
+    cmove       r2, r3mp
+    lea         r3, [ang_table + 13 * 16]
+    movu        m0, [r2 + 2]    ; [8 7 6 5 4 3 2 1]
+    palignr     m1, m0, 2       ; [x 8 7 6 5 4 3 2]
+    punpcklwd   m2, m0, m1      ; [5 4 4 3 3 2 2 1]
+    mova        m3, m2
+    mova        m4, m2
+    mova        m5, m2
+
+    mova        m0, [r3 -  8 * 16]  ; [ 5]
+    mova        m1, [r3 -  3 * 16]  ; [10]
+    mova        m6, [r3 +  2 * 16]  ; [15]
+    mova        m7, [r3 +  7 * 16]  ; [20]
+    jmp         mangle(private_prefix %+ _ %+ intra_pred_ang4_3 %+ SUFFIX %+ .do_filter4x4)
+
+
+cglobal intra_pred_ang4_9, 3,4,8
+    cmp         r4m, byte 27
+    cmove       r2, r3mp
+    lea         r3, [ang_table + 4 * 16]
+    movu        m0, [r2 + 2]    ; [8 7 6 5 4 3 2 1]
+    palignr     m1, m0, 2       ; [x 8 7 6 5 4 3 2]
+    punpcklwd   m2, m0, m1      ; [5 4 4 3 3 2 2 1]
+    mova        m3, m2
+    mova        m4, m2
+    mova        m5, m2
+
+    mova        m0, [r3 -  2 * 16]  ; [ 2]
+    mova        m1, [r3 -  0 * 16]  ; [ 4]
+    mova        m6, [r3 +  2 * 16]  ; [ 6]
+    mova        m7, [r3 +  4 * 16]  ; [ 8]
+    jmp         mangle(private_prefix %+ _ %+ intra_pred_ang4_3 %+ SUFFIX %+ .do_filter4x4)


More information about the x265-devel mailing list