Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(34)

Issues Search
    My Issues | Starred     Open | Closed | All

Unified Diff: source/libvpx/vp9/common/arm/neon/vp9_idct16x16_1_add_neon.asm

Issue 168343002: libvpx: Pull from upstream (Closed) Base URL: svn://svn.chromium.org/chrome/trunk/deps/third_party/libvpx/
Patch Set: libvpx: Pull from upstream Created 6 years, 10 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View side-by-side diff with in-line comments
Download patch
« no previous file with comments | « source/libvpx/vp8_scalable_patterns.c ('k') | source/libvpx/vp9/common/arm/neon/vp9_idct16x16_add_neon.asm » ('j') | no next file with comments »
Expand Comments ('e') | Collapse Comments ('c') | Show Comments Hide Comments ('s')
Index: source/libvpx/vp9/common/arm/neon/vp9_idct16x16_1_add_neon.asm
===================================================================
--- source/libvpx/vp9/common/arm/neon/vp9_idct16x16_1_add_neon.asm (revision 0)
+++ source/libvpx/vp9/common/arm/neon/vp9_idct16x16_1_add_neon.asm (revision 0)
@@ -0,0 +1,198 @@
+;
+; Copyright (c) 2013 The WebM project authors. All Rights Reserved.
+;
+; Use of this source code is governed by a BSD-style license and patent
+; grant that can be found in the LICENSE file in the root of the source
+; tree. All contributing project authors may be found in the AUTHORS
+; file in the root of the source tree.
+;
+
+
+ EXPORT |vp9_idct16x16_1_add_neon|
+ ARM
+ REQUIRE8
+ PRESERVE8
+
+ AREA ||.text||, CODE, READONLY, ALIGN=2
+
+;void vp9_idct16x16_1_add_neon(int16_t *input, uint8_t *dest,
+; int dest_stride)
+;
+; r0 int16_t input
+; r1 uint8_t *dest
+; r2 int dest_stride)
+
+|vp9_idct16x16_1_add_neon| PROC
+ ldrsh r0, [r0]
+
+ ; generate cospi_16_64 = 11585
+ mov r12, #0x2d00
+ add r12, #0x41
+
+ ; out = dct_const_round_shift(input[0] * cospi_16_64)
+ mul r0, r0, r12 ; input[0] * cospi_16_64
+ add r0, r0, #0x2000 ; +(1 << ((DCT_CONST_BITS) - 1))
+ asr r0, r0, #14 ; >> DCT_CONST_BITS
+
+ ; out = dct_const_round_shift(out * cospi_16_64)
+ mul r0, r0, r12 ; out * cospi_16_64
+ mov r12, r1 ; save dest
+ add r0, r0, #0x2000 ; +(1 << ((DCT_CONST_BITS) - 1))
+ asr r0, r0, #14 ; >> DCT_CONST_BITS
+
+ ; a1 = ROUND_POWER_OF_TWO(out, 6)
+ add r0, r0, #32 ; + (1 <<((6) - 1))
+ asr r0, r0, #6 ; >> 6
+
+ vdup.s16 q0, r0 ; duplicate a1
+ mov r0, #8
+ sub r2, #8
+
+ ; load destination data row0 - row3
+ vld1.64 {d2}, [r1], r0
+ vld1.64 {d3}, [r1], r2
+ vld1.64 {d4}, [r1], r0
+ vld1.64 {d5}, [r1], r2
+ vld1.64 {d6}, [r1], r0
+ vld1.64 {d7}, [r1], r2
+ vld1.64 {d16}, [r1], r0
+ vld1.64 {d17}, [r1], r2
+
+ vaddw.u8 q9, q0, d2 ; dest[x] + a1
+ vaddw.u8 q10, q0, d3 ; dest[x] + a1
+ vaddw.u8 q11, q0, d4 ; dest[x] + a1
+ vaddw.u8 q12, q0, d5 ; dest[x] + a1
+ vqmovun.s16 d2, q9 ; clip_pixel
+ vqmovun.s16 d3, q10 ; clip_pixel
+ vqmovun.s16 d30, q11 ; clip_pixel
+ vqmovun.s16 d31, q12 ; clip_pixel
+ vst1.64 {d2}, [r12], r0
+ vst1.64 {d3}, [r12], r2
+ vst1.64 {d30}, [r12], r0
+ vst1.64 {d31}, [r12], r2
+
+ vaddw.u8 q9, q0, d6 ; dest[x] + a1
+ vaddw.u8 q10, q0, d7 ; dest[x] + a1
+ vaddw.u8 q11, q0, d16 ; dest[x] + a1
+ vaddw.u8 q12, q0, d17 ; dest[x] + a1
+ vqmovun.s16 d2, q9 ; clip_pixel
+ vqmovun.s16 d3, q10 ; clip_pixel
+ vqmovun.s16 d30, q11 ; clip_pixel
+ vqmovun.s16 d31, q12 ; clip_pixel
+ vst1.64 {d2}, [r12], r0
+ vst1.64 {d3}, [r12], r2
+ vst1.64 {d30}, [r12], r0
+ vst1.64 {d31}, [r12], r2
+
+ ; load destination data row4 - row7
+ vld1.64 {d2}, [r1], r0
+ vld1.64 {d3}, [r1], r2
+ vld1.64 {d4}, [r1], r0
+ vld1.64 {d5}, [r1], r2
+ vld1.64 {d6}, [r1], r0
+ vld1.64 {d7}, [r1], r2
+ vld1.64 {d16}, [r1], r0
+ vld1.64 {d17}, [r1], r2
+
+ vaddw.u8 q9, q0, d2 ; dest[x] + a1
+ vaddw.u8 q10, q0, d3 ; dest[x] + a1
+ vaddw.u8 q11, q0, d4 ; dest[x] + a1
+ vaddw.u8 q12, q0, d5 ; dest[x] + a1
+ vqmovun.s16 d2, q9 ; clip_pixel
+ vqmovun.s16 d3, q10 ; clip_pixel
+ vqmovun.s16 d30, q11 ; clip_pixel
+ vqmovun.s16 d31, q12 ; clip_pixel
+ vst1.64 {d2}, [r12], r0
+ vst1.64 {d3}, [r12], r2
+ vst1.64 {d30}, [r12], r0
+ vst1.64 {d31}, [r12], r2
+
+ vaddw.u8 q9, q0, d6 ; dest[x] + a1
+ vaddw.u8 q10, q0, d7 ; dest[x] + a1
+ vaddw.u8 q11, q0, d16 ; dest[x] + a1
+ vaddw.u8 q12, q0, d17 ; dest[x] + a1
+ vqmovun.s16 d2, q9 ; clip_pixel
+ vqmovun.s16 d3, q10 ; clip_pixel
+ vqmovun.s16 d30, q11 ; clip_pixel
+ vqmovun.s16 d31, q12 ; clip_pixel
+ vst1.64 {d2}, [r12], r0
+ vst1.64 {d3}, [r12], r2
+ vst1.64 {d30}, [r12], r0
+ vst1.64 {d31}, [r12], r2
+
+ ; load destination data row8 - row11
+ vld1.64 {d2}, [r1], r0
+ vld1.64 {d3}, [r1], r2
+ vld1.64 {d4}, [r1], r0
+ vld1.64 {d5}, [r1], r2
+ vld1.64 {d6}, [r1], r0
+ vld1.64 {d7}, [r1], r2
+ vld1.64 {d16}, [r1], r0
+ vld1.64 {d17}, [r1], r2
+
+ vaddw.u8 q9, q0, d2 ; dest[x] + a1
+ vaddw.u8 q10, q0, d3 ; dest[x] + a1
+ vaddw.u8 q11, q0, d4 ; dest[x] + a1
+ vaddw.u8 q12, q0, d5 ; dest[x] + a1
+ vqmovun.s16 d2, q9 ; clip_pixel
+ vqmovun.s16 d3, q10 ; clip_pixel
+ vqmovun.s16 d30, q11 ; clip_pixel
+ vqmovun.s16 d31, q12 ; clip_pixel
+ vst1.64 {d2}, [r12], r0
+ vst1.64 {d3}, [r12], r2
+ vst1.64 {d30}, [r12], r0
+ vst1.64 {d31}, [r12], r2
+
+ vaddw.u8 q9, q0, d6 ; dest[x] + a1
+ vaddw.u8 q10, q0, d7 ; dest[x] + a1
+ vaddw.u8 q11, q0, d16 ; dest[x] + a1
+ vaddw.u8 q12, q0, d17 ; dest[x] + a1
+ vqmovun.s16 d2, q9 ; clip_pixel
+ vqmovun.s16 d3, q10 ; clip_pixel
+ vqmovun.s16 d30, q11 ; clip_pixel
+ vqmovun.s16 d31, q12 ; clip_pixel
+ vst1.64 {d2}, [r12], r0
+ vst1.64 {d3}, [r12], r2
+ vst1.64 {d30}, [r12], r0
+ vst1.64 {d31}, [r12], r2
+
+ ; load destination data row12 - row15
+ vld1.64 {d2}, [r1], r0
+ vld1.64 {d3}, [r1], r2
+ vld1.64 {d4}, [r1], r0
+ vld1.64 {d5}, [r1], r2
+ vld1.64 {d6}, [r1], r0
+ vld1.64 {d7}, [r1], r2
+ vld1.64 {d16}, [r1], r0
+ vld1.64 {d17}, [r1], r2
+
+ vaddw.u8 q9, q0, d2 ; dest[x] + a1
+ vaddw.u8 q10, q0, d3 ; dest[x] + a1
+ vaddw.u8 q11, q0, d4 ; dest[x] + a1
+ vaddw.u8 q12, q0, d5 ; dest[x] + a1
+ vqmovun.s16 d2, q9 ; clip_pixel
+ vqmovun.s16 d3, q10 ; clip_pixel
+ vqmovun.s16 d30, q11 ; clip_pixel
+ vqmovun.s16 d31, q12 ; clip_pixel
+ vst1.64 {d2}, [r12], r0
+ vst1.64 {d3}, [r12], r2
+ vst1.64 {d30}, [r12], r0
+ vst1.64 {d31}, [r12], r2
+
+ vaddw.u8 q9, q0, d6 ; dest[x] + a1
+ vaddw.u8 q10, q0, d7 ; dest[x] + a1
+ vaddw.u8 q11, q0, d16 ; dest[x] + a1
+ vaddw.u8 q12, q0, d17 ; dest[x] + a1
+ vqmovun.s16 d2, q9 ; clip_pixel
+ vqmovun.s16 d3, q10 ; clip_pixel
+ vqmovun.s16 d30, q11 ; clip_pixel
+ vqmovun.s16 d31, q12 ; clip_pixel
+ vst1.64 {d2}, [r12], r0
+ vst1.64 {d3}, [r12], r2
+ vst1.64 {d30}, [r12], r0
+ vst1.64 {d31}, [r12], r2
+
+ bx lr
+ ENDP ; |vp9_idct16x16_1_add_neon|
+
+ END
« no previous file with comments | « source/libvpx/vp8_scalable_patterns.c ('k') | source/libvpx/vp9/common/arm/neon/vp9_idct16x16_add_neon.asm » ('j') | no next file with comments »

Powered by Google App Engine
This is Rietveld 408576698