[FFmpeg-cvslog] avcodec/hevcdsp: ARM NEON optimized qpel functions

Seppo Tomperi git at videolan.org
Wed Feb 25 18:40:57 CET 2015


ffmpeg | branch: master | Seppo Tomperi <seppo.tomperi at vtt.fi> | Wed Feb 25 09:44:55 2015 +0000| [63ca0fe8288dbd300c9bb814cb671e5d889f691c] | committer: Michael Niedermayer

avcodec/hevcdsp: ARM NEON optimized qpel functions

uses comma as macro parameter separator

Signed-off-by: Michael Niedermayer <michaelni at gmx.at>

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=63ca0fe8288dbd300c9bb814cb671e5d889f691c
---

 libavcodec/arm/Makefile            |    3 +-
 libavcodec/arm/hevcdsp_init_neon.c |  159 ++++++
 libavcodec/arm/hevcdsp_qpel_neon.S |  999 ++++++++++++++++++++++++++++++++++++
 3 files changed, 1160 insertions(+), 1 deletion(-)

diff --git a/libavcodec/arm/Makefile b/libavcodec/arm/Makefile
index 240ee3f..d1b4826 100644
--- a/libavcodec/arm/Makefile
+++ b/libavcodec/arm/Makefile
@@ -138,7 +138,8 @@ NEON-OBJS-$(CONFIG_DCA_DECODER)        += arm/dcadsp_neon.o             \
                                           arm/synth_filter_neon.o
 NEON-OBJS-$(CONFIG_HEVC_DECODER)       += arm/hevcdsp_init_neon.o       \
                                           arm/hevcdsp_deblock_neon.o    \
-                                          arm/hevcdsp_idct_neon.o
+                                          arm/hevcdsp_idct_neon.o       \
+                                          arm/hevcdsp_qpel_neon.o
 NEON-OBJS-$(CONFIG_RV30_DECODER)       += arm/rv34dsp_neon.o
 NEON-OBJS-$(CONFIG_RV40_DECODER)       += arm/rv34dsp_neon.o            \
                                           arm/rv40dsp_neon.o
diff --git a/libavcodec/arm/hevcdsp_init_neon.c b/libavcodec/arm/hevcdsp_init_neon.c
index 1f132db..61e6462 100644
--- a/libavcodec/arm/hevcdsp_init_neon.c
+++ b/libavcodec/arm/hevcdsp_init_neon.c
@@ -42,10 +42,110 @@ void ff_hevc_transform_add_16x16_neon_8(uint8_t *_dst, int16_t *coeffs,
 void ff_hevc_transform_add_32x32_neon_8(uint8_t *_dst, int16_t *coeffs,
                                       ptrdiff_t stride);
 
+#define PUT_PIXELS(name) \
+    void name(int16_t *dst, uint8_t *src, \
+                                ptrdiff_t srcstride, int height, \
+                                intptr_t mx, intptr_t my, int width)
+PUT_PIXELS(ff_hevc_put_pixels_w2_neon_8);
+PUT_PIXELS(ff_hevc_put_pixels_w4_neon_8);
+PUT_PIXELS(ff_hevc_put_pixels_w6_neon_8);
+PUT_PIXELS(ff_hevc_put_pixels_w8_neon_8);
+PUT_PIXELS(ff_hevc_put_pixels_w12_neon_8);
+PUT_PIXELS(ff_hevc_put_pixels_w16_neon_8);
+PUT_PIXELS(ff_hevc_put_pixels_w24_neon_8);
+PUT_PIXELS(ff_hevc_put_pixels_w32_neon_8);
+PUT_PIXELS(ff_hevc_put_pixels_w48_neon_8);
+PUT_PIXELS(ff_hevc_put_pixels_w64_neon_8);
+#undef PUT_PIXELS
+
+static void (*put_hevc_qpel_neon[4][4])(int16_t *dst, ptrdiff_t dststride, uint8_t *src, ptrdiff_t srcstride,
+                                   int height, int width);
+static void (*put_hevc_qpel_uw_neon[4][4])(uint8_t *dst, ptrdiff_t dststride, uint8_t *_src, ptrdiff_t _srcstride,
+                                   int width, int height, int16_t* src2, ptrdiff_t src2stride);
+void ff_hevc_put_qpel_neon_wrapper(int16_t *dst, uint8_t *src, ptrdiff_t srcstride,
+                                   int height, intptr_t mx, intptr_t my, int width);
+void ff_hevc_put_qpel_uni_neon_wrapper(uint8_t *dst, ptrdiff_t dststride, uint8_t *src, ptrdiff_t srcstride,
+                                   int height, intptr_t mx, intptr_t my, int width);
+void ff_hevc_put_qpel_bi_neon_wrapper(uint8_t *dst, ptrdiff_t dststride, uint8_t *src, ptrdiff_t srcstride,
+                                       int16_t *src2,
+                                       int height, intptr_t mx, intptr_t my, int width);
+#define QPEL_FUNC(name) \
+    void name(int16_t *dst, ptrdiff_t dststride, uint8_t *src, ptrdiff_t srcstride, \
+                                   int height, int width)
+
+QPEL_FUNC(ff_hevc_put_qpel_v1_neon_8);
+QPEL_FUNC(ff_hevc_put_qpel_v2_neon_8);
+QPEL_FUNC(ff_hevc_put_qpel_v3_neon_8);
+QPEL_FUNC(ff_hevc_put_qpel_h1_neon_8);
+QPEL_FUNC(ff_hevc_put_qpel_h2_neon_8);
+QPEL_FUNC(ff_hevc_put_qpel_h3_neon_8);
+QPEL_FUNC(ff_hevc_put_qpel_h1v1_neon_8);
+QPEL_FUNC(ff_hevc_put_qpel_h1v2_neon_8);
+QPEL_FUNC(ff_hevc_put_qpel_h1v3_neon_8);
+QPEL_FUNC(ff_hevc_put_qpel_h2v1_neon_8);
+QPEL_FUNC(ff_hevc_put_qpel_h2v2_neon_8);
+QPEL_FUNC(ff_hevc_put_qpel_h2v3_neon_8);
+QPEL_FUNC(ff_hevc_put_qpel_h3v1_neon_8);
+QPEL_FUNC(ff_hevc_put_qpel_h3v2_neon_8);
+QPEL_FUNC(ff_hevc_put_qpel_h3v3_neon_8);
+#undef QPEL_FUNC
+
+#define QPEL_FUNC_UW_PIX(name) \
+    void name(uint8_t *dst, ptrdiff_t dststride, uint8_t *_src, ptrdiff_t _srcstride, \
+                                   int height, intptr_t mx, intptr_t my, int width);
+QPEL_FUNC_UW_PIX(ff_hevc_put_qpel_uw_pixels_w4_neon_8);
+QPEL_FUNC_UW_PIX(ff_hevc_put_qpel_uw_pixels_w8_neon_8);
+QPEL_FUNC_UW_PIX(ff_hevc_put_qpel_uw_pixels_w16_neon_8);
+QPEL_FUNC_UW_PIX(ff_hevc_put_qpel_uw_pixels_w24_neon_8);
+QPEL_FUNC_UW_PIX(ff_hevc_put_qpel_uw_pixels_w32_neon_8);
+QPEL_FUNC_UW_PIX(ff_hevc_put_qpel_uw_pixels_w48_neon_8);
+QPEL_FUNC_UW_PIX(ff_hevc_put_qpel_uw_pixels_w64_neon_8);
+#undef QPEL_FUNC_UW_PIX
+
+#define QPEL_FUNC_UW(name) \
+    void name(uint8_t *dst, ptrdiff_t dststride, uint8_t *_src, ptrdiff_t _srcstride, \
+                                   int width, int height, int16_t* src2, ptrdiff_t src2stride);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_pixels_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_v1_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_v2_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_v3_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_h1_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_h2_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_h3_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_h1v1_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_h1v2_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_h1v3_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_h2v1_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_h2v2_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_h2v3_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_h3v1_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_h3v2_neon_8);
+QPEL_FUNC_UW(ff_hevc_put_qpel_uw_h3v3_neon_8);
+#undef QPEL_FUNC_UW
+
+void ff_hevc_put_qpel_neon_wrapper(int16_t *dst, uint8_t *src, ptrdiff_t srcstride,
+                                   int height, intptr_t mx, intptr_t my, int width) {
+
+    put_hevc_qpel_neon[my][mx](dst, MAX_PB_SIZE, src, srcstride, height, width);
+}
+
+void ff_hevc_put_qpel_uni_neon_wrapper(uint8_t *dst, ptrdiff_t dststride, uint8_t *src, ptrdiff_t srcstride,
+                                   int height, intptr_t mx, intptr_t my, int width) {
+
+    put_hevc_qpel_uw_neon[my][mx](dst, dststride, src, srcstride, width, height, NULL, 0);
+}
+
+void ff_hevc_put_qpel_bi_neon_wrapper(uint8_t *dst, ptrdiff_t dststride, uint8_t *src, ptrdiff_t srcstride,
+                                       int16_t *src2,
+                                       int height, intptr_t mx, intptr_t my, int width) {
+    put_hevc_qpel_uw_neon[my][mx](dst, dststride, src, srcstride, width, height, src2, MAX_PB_SIZE);
+}
+
 static av_cold void hevcdsp_init_neon(HEVCDSPContext *c, const int bit_depth)
 {
 #if HAVE_NEON
     if (bit_depth == 8) {
+        int x;
         c->hevc_v_loop_filter_luma     = ff_hevc_v_loop_filter_luma_neon;
         c->hevc_h_loop_filter_luma     = ff_hevc_h_loop_filter_luma_neon;
         c->hevc_v_loop_filter_chroma   = ff_hevc_v_loop_filter_chroma_neon;
@@ -61,6 +161,65 @@ static av_cold void hevcdsp_init_neon(HEVCDSPContext *c, const int bit_depth)
         c->transform_add[2]            = ff_hevc_transform_add_16x16_neon_8;
         c->transform_add[3]            = ff_hevc_transform_add_32x32_neon_8;
         c->idct_4x4_luma               = ff_hevc_transform_luma_4x4_neon_8;
+        put_hevc_qpel_neon[1][0]       = ff_hevc_put_qpel_v1_neon_8;
+        put_hevc_qpel_neon[2][0]       = ff_hevc_put_qpel_v2_neon_8;
+        put_hevc_qpel_neon[3][0]       = ff_hevc_put_qpel_v3_neon_8;
+        put_hevc_qpel_neon[0][1]       = ff_hevc_put_qpel_h1_neon_8;
+        put_hevc_qpel_neon[0][2]       = ff_hevc_put_qpel_h2_neon_8;
+        put_hevc_qpel_neon[0][3]       = ff_hevc_put_qpel_h3_neon_8;
+        put_hevc_qpel_neon[1][1]       = ff_hevc_put_qpel_h1v1_neon_8;
+        put_hevc_qpel_neon[1][2]       = ff_hevc_put_qpel_h2v1_neon_8;
+        put_hevc_qpel_neon[1][3]       = ff_hevc_put_qpel_h3v1_neon_8;
+        put_hevc_qpel_neon[2][1]       = ff_hevc_put_qpel_h1v2_neon_8;
+        put_hevc_qpel_neon[2][2]       = ff_hevc_put_qpel_h2v2_neon_8;
+        put_hevc_qpel_neon[2][3]       = ff_hevc_put_qpel_h3v2_neon_8;
+        put_hevc_qpel_neon[3][1]       = ff_hevc_put_qpel_h1v3_neon_8;
+        put_hevc_qpel_neon[3][2]       = ff_hevc_put_qpel_h2v3_neon_8;
+        put_hevc_qpel_neon[3][3]       = ff_hevc_put_qpel_h3v3_neon_8;
+        put_hevc_qpel_uw_neon[1][0]      = ff_hevc_put_qpel_uw_v1_neon_8;
+        put_hevc_qpel_uw_neon[2][0]      = ff_hevc_put_qpel_uw_v2_neon_8;
+        put_hevc_qpel_uw_neon[3][0]      = ff_hevc_put_qpel_uw_v3_neon_8;
+        put_hevc_qpel_uw_neon[0][1]      = ff_hevc_put_qpel_uw_h1_neon_8;
+        put_hevc_qpel_uw_neon[0][2]      = ff_hevc_put_qpel_uw_h2_neon_8;
+        put_hevc_qpel_uw_neon[0][3]      = ff_hevc_put_qpel_uw_h3_neon_8;
+        put_hevc_qpel_uw_neon[1][1]      = ff_hevc_put_qpel_uw_h1v1_neon_8;
+        put_hevc_qpel_uw_neon[1][2]      = ff_hevc_put_qpel_uw_h2v1_neon_8;
+        put_hevc_qpel_uw_neon[1][3]      = ff_hevc_put_qpel_uw_h3v1_neon_8;
+        put_hevc_qpel_uw_neon[2][1]      = ff_hevc_put_qpel_uw_h1v2_neon_8;
+        put_hevc_qpel_uw_neon[2][2]      = ff_hevc_put_qpel_uw_h2v2_neon_8;
+        put_hevc_qpel_uw_neon[2][3]      = ff_hevc_put_qpel_uw_h3v2_neon_8;
+        put_hevc_qpel_uw_neon[3][1]      = ff_hevc_put_qpel_uw_h1v3_neon_8;
+        put_hevc_qpel_uw_neon[3][2]      = ff_hevc_put_qpel_uw_h2v3_neon_8;
+        put_hevc_qpel_uw_neon[3][3]      = ff_hevc_put_qpel_uw_h3v3_neon_8;
+        for (x = 0; x < 10; x++) {
+            c->put_hevc_qpel[x][1][0]         = ff_hevc_put_qpel_neon_wrapper;
+            c->put_hevc_qpel[x][0][1]         = ff_hevc_put_qpel_neon_wrapper;
+            c->put_hevc_qpel[x][1][1]         = ff_hevc_put_qpel_neon_wrapper;
+            c->put_hevc_qpel_uni[x][1][0]     = ff_hevc_put_qpel_uni_neon_wrapper;
+            c->put_hevc_qpel_uni[x][0][1]     = ff_hevc_put_qpel_uni_neon_wrapper;
+            c->put_hevc_qpel_uni[x][1][1]     = ff_hevc_put_qpel_uni_neon_wrapper;
+            c->put_hevc_qpel_bi[x][1][0]      = ff_hevc_put_qpel_bi_neon_wrapper;
+            c->put_hevc_qpel_bi[x][0][1]      = ff_hevc_put_qpel_bi_neon_wrapper;
+            c->put_hevc_qpel_bi[x][1][1]      = ff_hevc_put_qpel_bi_neon_wrapper;
+        }
+        c->put_hevc_qpel[0][0][0]  = ff_hevc_put_pixels_w2_neon_8;
+        c->put_hevc_qpel[1][0][0]  = ff_hevc_put_pixels_w4_neon_8;
+        c->put_hevc_qpel[2][0][0]  = ff_hevc_put_pixels_w6_neon_8;
+        c->put_hevc_qpel[3][0][0]  = ff_hevc_put_pixels_w8_neon_8;
+        c->put_hevc_qpel[4][0][0]  = ff_hevc_put_pixels_w12_neon_8;
+        c->put_hevc_qpel[5][0][0]  = ff_hevc_put_pixels_w16_neon_8;
+        c->put_hevc_qpel[6][0][0]  = ff_hevc_put_pixels_w24_neon_8;
+        c->put_hevc_qpel[7][0][0]  = ff_hevc_put_pixels_w32_neon_8;
+        c->put_hevc_qpel[8][0][0]  = ff_hevc_put_pixels_w48_neon_8;
+        c->put_hevc_qpel[9][0][0]  = ff_hevc_put_pixels_w64_neon_8;
+
+        c->put_hevc_qpel_uni[1][0][0]  = ff_hevc_put_qpel_uw_pixels_w4_neon_8;
+        c->put_hevc_qpel_uni[3][0][0]  = ff_hevc_put_qpel_uw_pixels_w8_neon_8;
+        c->put_hevc_qpel_uni[5][0][0]  = ff_hevc_put_qpel_uw_pixels_w16_neon_8;
+        c->put_hevc_qpel_uni[6][0][0]  = ff_hevc_put_qpel_uw_pixels_w24_neon_8;
+        c->put_hevc_qpel_uni[7][0][0]  = ff_hevc_put_qpel_uw_pixels_w32_neon_8;
+        c->put_hevc_qpel_uni[8][0][0]  = ff_hevc_put_qpel_uw_pixels_w48_neon_8;
+        c->put_hevc_qpel_uni[9][0][0]  = ff_hevc_put_qpel_uw_pixels_w64_neon_8;
     }
 #endif // HAVE_NEON
 }
diff --git a/libavcodec/arm/hevcdsp_qpel_neon.S b/libavcodec/arm/hevcdsp_qpel_neon.S
new file mode 100644
index 0000000..86f92cf
--- /dev/null
+++ b/libavcodec/arm/hevcdsp_qpel_neon.S
@@ -0,0 +1,999 @@
+/*
+ * Copyright (c) 2014 - 2015 Seppo Tomperi <seppo.tomperi at vtt.fi>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavutil/arm/asm.S"
+#include "neon.S"
+
+#define MAX_PB_SIZE #64
+
+.macro regshuffle_d8
+    vmov d16, d17
+    vmov d17, d18
+    vmov d18, d19
+    vmov d19, d20
+    vmov d20, d21
+    vmov d21, d22
+    vmov d22, d23
+.endm
+
+.macro regshuffle_q8
+    vmov q0, q1
+    vmov q1, q2
+    vmov q2, q3
+    vmov q3, q4
+    vmov q4, q5
+    vmov q5, q6
+    vmov q6, q7
+.endm
+
+.macro vextin8
+        pld       [r2]
+        vld1.8    {q11}, [r2], r3
+        vext.8    d16, d22, d23, #1
+        vext.8    d17, d22, d23, #2
+        vext.8    d18, d22, d23, #3
+        vext.8    d19, d22, d23, #4
+        vext.8    d20, d22, d23, #5
+        vext.8    d21, d22, d23, #6
+        vext.8    d22, d22, d23, #7
+.endm
+
+.macro loadin8
+        pld       [r2]
+        vld1.8    {d16}, [r2], r3
+        pld       [r2]
+        vld1.8    {d17}, [r2], r3
+        pld       [r2]
+        vld1.8    {d18}, [r2], r3
+        pld       [r2]
+        vld1.8    {d19}, [r2], r3
+        pld       [r2]
+        vld1.8    {d20}, [r2], r3
+        pld       [r2]
+        vld1.8    {d21}, [r2], r3
+        pld       [r2]
+        vld1.8    {d22}, [r2], r3
+        pld       [r2]
+        vld1.8    {d23}, [r2], r3
+.endm
+
+.macro qpel_filter_1_32b
+        vmov.i16   d16, #58
+        vmov.i16   d17, #10
+        vmull.s16   q9, d6, d16   // 58 * d0
+        vmull.s16  q10, d7, d16   // 58 * d1
+        vmov.i16   d16, #17
+        vmull.s16  q11, d4, d17   // 10 * c0
+        vmull.s16  q12, d5, d17   // 10 * c1
+        vmov.i16   d17, #5
+        vmull.s16  q13, d8, d16   // 17 * e0
+        vmull.s16  q14, d9, d16   // 17 * e1
+        vmull.s16  q15, d10, d17  //  5 * f0
+        vmull.s16   q8, d11, d17  //  5 * f1
+        vsub.s32    q9, q11       // 58 * d0 - 10 * c0
+        vsub.s32   q10, q12       // 58 * d1 - 10 * c1
+        vshll.s16  q11, d2, #2    // 4 * b0
+        vshll.s16  q12, d3, #2    // 4 * b1
+        vadd.s32    q9, q13       // 58 * d0 - 10 * c0 + 17 * e0
+        vadd.s32   q10, q14       // 58 * d1 - 10 * c1 + 17 * e1
+        vsubl.s16  q13, d12, d0   // g0 - a0
+        vsubl.s16  q14, d13, d1   // g1 - a1
+        vadd.s32    q9, q11       // 58 * d0 - 10 * c0 + 17 * e0 + 4 * b0
+        vadd.s32   q10, q12       // 58 * d1 - 10 * c1 + 17 * e1 + 4 * b1
+        vsub.s32   q13, q15       // g0 - a0 - 5 * f0
+        vsub.s32   q14, q8        // g1 - a1 - 5 * f1
+        vadd.s32    q9, q13       // 58 * d0 - 10 * c0 + 17 * e0 + 4 * b0 + g0 - a0 - 5 * f0
+        vadd.s32   q10, q14       // 58 * d1 - 10 * c1 + 17 * e1 + 4 * b1 + g1 - a1 - 5 * f1
+        vqshrn.s32  d16, q9, #6
+        vqshrn.s32  d17, q10, #6
+.endm
+
+// input  q0 - q7
+// output q8
+.macro qpel_filter_2_32b
+        vmov.i32   q8, #11
+        vaddl.s16   q9, d6, d8   // d0 + e0
+        vaddl.s16  q10, d7, d9   // d1 + e1
+        vaddl.s16  q11, d4, d10  // c0 + f0
+        vaddl.s16  q12, d5, d11  // c1 + f1
+        vmul.s32   q11, q8       // 11 * (c0 + f0)
+        vmul.s32   q12, q8       // 11 * (c1 + f1)
+        vmov.i32   q8, #40
+        vaddl.s16  q15, d2, d12  // b0 + g0
+        vmul.s32    q9, q8       // 40 * (d0 + e0)
+        vmul.s32   q10, q8       // 40 * (d1 + e1)
+        vaddl.s16   q8, d3, d13  // b1 + g1
+        vaddl.s16  q13, d0, d14  // a0 + h0
+        vaddl.s16  q14, d1, d15  // a1 + h1
+        vshl.s32   q15, #2       // 4*(b0+g0)
+        vshl.s32    q8, #2       // 4*(b1+g1)
+        vadd.s32   q11, q13      // 11 * (c0 + f0) + a0 + h0
+        vadd.s32   q12, q14      // 11 * (c1 + f1) + a1 + h1
+        vadd.s32   q9, q15       // 40 * (d0 + e0) + 4*(b0+g0)
+        vadd.s32   q10, q8       // 40 * (d1 + e1) + 4*(b1+g1)
+        vsub.s32   q9, q11       // 40 * (d0 + e0) + 4*(b0+g0) - (11 * (c0 + f0) + a0 + h0)
+        vsub.s32   q10, q12      // 40 * (d1 + e1) + 4*(b1+g1) - (11 * (c1 + f1) + a1 + h1)
+        vqshrn.s32  d16, q9, #6
+        vqshrn.s32  d17, q10, #6
+.endm
+
+.macro qpel_filter_3_32b
+        vmov.i16   d16, #58
+        vmov.i16   d17, #10
+        vmull.s16   q9, d8, d16   // 58 * d0
+        vmull.s16  q10, d9, d16   // 58 * d1
+        vmov.i16   d16, #17
+        vmull.s16  q11, d10, d17  // 10 * c0
+        vmull.s16  q12, d11, d17  // 10 * c1
+        vmov.i16   d17, #5
+        vmull.s16  q13, d6, d16   // 17 * e0
+        vmull.s16  q14, d7, d16   // 17 * e1
+        vmull.s16  q15, d4, d17   //  5 * f0
+        vmull.s16   q8, d5, d17   //  5 * f1
+        vsub.s32    q9, q11       // 58 * d0 - 10 * c0
+        vsub.s32   q10, q12       // 58 * d1 - 10 * c1
+        vshll.s16  q11, d12, #2   // 4 * b0
+        vshll.s16  q12, d13, #2   // 4 * b1
+        vadd.s32    q9, q13       // 58 * d0 - 10 * c0 + 17 * e0
+        vadd.s32   q10, q14       // 58 * d1 - 10 * c1 + 17 * e1
+        vsubl.s16  q13, d2, d14   // g0 - a0
+        vsubl.s16  q14, d3, d15   // g1 - a1
+        vadd.s32    q9, q11       // 58 * d0 - 10 * c0 + 17 * e0 + 4 * b0
+        vadd.s32   q10, q12       // 58 * d1 - 10 * c1 + 17 * e1 + 4 * b1
+        vsub.s32   q13, q15       // g0 - a0 - 5 * f0
+        vsub.s32   q14, q8        // g1 - a1 - 5 * f1
+        vadd.s32    q9, q13       // 58 * d0 - 10 * c0 + 17 * e0 + 4 * b0 + g0 - a0 - 5 * f0
+        vadd.s32   q10, q14       // 58 * d1 - 10 * c1 + 17 * e1 + 4 * b1 + g1 - a1 - 5 * f1
+        vqshrn.s32  d16, q9, #6
+        vqshrn.s32  d17, q10, #6
+.endm
+
+.macro qpel_filter_1 out=q7
+        vmov.u8    d24, #58
+        vmov.u8    d25, #10
+        vshll.u8   q13, d20, #4   // 16*e
+        vshll.u8   q14, d21, #2   // 4*f
+        vmull.u8  \out, d19, d24  // 58*d
+        vaddw.u8   q13, q13, d20  // 17*e
+        vmull.u8   q15, d18, d25  // 10*c
+        vaddw.u8   q14, q14, d21  // 5*f
+        vsubl.u8   q12, d22, d16  // g - a
+        vadd.u16  \out, q13       // 58d + 17e
+        vshll.u8   q13, d17, #2   // 4*b
+        vadd.u16   q15, q14       // 10*c + 5*f
+        vadd.s16   q13, q12       // - a + 4*b + g
+        vsub.s16  \out, q15       // -10*c + 58*d + 17*e -5*f
+        vadd.s16  \out, q13       // -a + 4*b -10*c + 58*d + 17*e -5*f
+.endm
+
+.macro qpel_filter_2 out=q7
+        vmov.i16   q12, #10
+        vmov.i16   q14, #11
+        vaddl.u8   q13, d19, d20   // d + e
+        vaddl.u8   q15, d18, d21   // c + f
+        vmul.u16   q13, q12        // 10 * (d+e)
+        vmul.u16   q15, q14        // 11 * ( c + f)
+        vaddl.u8  \out, d17, d22   // b + g
+        vaddl.u8   q12, d16, d23   // a + h
+        vadd.u16  \out, q13        // b + 10 * (d + e) + g
+        vadd.s16   q12, q15
+        vshl.u16  \out, #2         // 4 * (b + 10 * (d + e) + g)
+        vsub.s16  \out, q12
+.endm
+
+.macro qpel_filter_3 out=q7
+        vmov.u8    d24, #58
+        vmov.u8    d25, #10
+        vshll.u8   q13, d19, #4     // 16*e
+        vshll.u8   q14, d18, #2     // 4*f
+        vmull.u8  \out, d20, d24    // 58*d
+        vaddw.u8   q13, q13, d19    // 17*e
+        vmull.u8   q15, d21, d25    // 10*c
+        vaddw.u8   q14, q14, d18    // 5*f
+        vsubl.u8   q12, d17, d23    // g - a
+        vadd.u16  \out, q13         // 58d + 17e
+        vshll.u8   q13, d22, #2     // 4*b
+        vadd.u16   q15, q14         // 10*c + 5*f
+        vadd.s16   q13, q12         // - a + 4*b + g
+        vsub.s16  \out, q15         // -10*c + 58*d + 17*e -5*f
+        vadd.s16  \out, q13         // -a + 4*b -10*c + 58*d + 17*e -5*f
+.endm
+
+.macro  hevc_put_qpel_vX_neon_8 filter
+        push   {r4, r5, r6, r7}
+        ldr    r4, [sp, #16] // height
+        ldr    r5, [sp, #20] // width
+        vpush {d8-d15}
+        sub       r2, r2, r3, lsl #1
+        sub       r2, r3
+        mov       r12, r4
+        mov       r6, r0
+        mov       r7, r2
+        lsl       r1, #1
+0:      loadin8
+        cmp       r5, #4
+        beq       4f
+8:      subs r4, #1
+        \filter
+        vst1.16    {q7}, [r0], r1
+        regshuffle_d8
+        vld1.8    {d23}, [r2], r3
+        bne 8b
+        subs  r5, #8
+        beq       99f
+        mov r4, r12
+        add r6, #16
+        mov r0, r6
+        add r7, #8
+        mov r2, r7
+        b     0b
+4:      subs r4, #1
+        \filter
+        vst1.16    d14, [r0], r1
+        regshuffle_d8
+        vld1.32    {d23[0]}, [r2], r3
+        bne 4b
+99:     vpop {d8-d15}
+        pop {r4, r5, r6, r7}
+        bx lr
+.endm
+
+.macro  hevc_put_qpel_uw_vX_neon_8 filter
+        push   {r4-r10}
+        ldr    r5, [sp, #28] // width
+        ldr    r4, [sp, #32] // height
+        ldr    r8, [sp, #36] // src2
+        ldr    r9, [sp, #40] // src2stride
+        vpush {d8-d15}
+        sub       r2, r2, r3, lsl #1
+        sub       r2, r3
+        mov       r12, r4
+        mov       r6, r0
+        mov       r7, r2
+        cmp       r8, #0
+        bne       .Lbi\@
+0:      loadin8
+        cmp       r5, #4
+        beq       4f
+8:      subs r4, #1
+        \filter
+        vqrshrun.s16   d0, q7, #6
+        vst1.8    d0, [r0], r1
+        regshuffle_d8
+        vld1.8    {d23}, [r2], r3
+        bne 8b
+        subs  r5, #8
+        beq       99f
+        mov r4, r12
+        add r6, #8
+        mov r0, r6
+        add r7, #8
+        mov r2, r7
+        b     0b
+4:      subs r4, #1
+        \filter
+        vqrshrun.s16   d0, q7, #6
+        vst1.32    d0[0], [r0], r1
+        regshuffle_d8
+        vld1.32    {d23[0]}, [r2], r3
+        bne 4b
+        b   99f
+.Lbi\@: lsl       r9, #1
+        mov       r10, r8
+0:      loadin8
+        cmp       r5, #4
+        beq       4f
+8:      subs r4, #1
+        \filter
+        vld1.16        {q0}, [r8], r9
+        vqadd.s16      q0, q7
+        vqrshrun.s16   d0, q0, #7
+        vst1.8         d0, [r0], r1
+        regshuffle_d8
+        vld1.8    {d23}, [r2], r3
+        bne 8b
+        subs  r5, #8
+        beq       99f
+        mov r4, r12
+        add r6, #8
+        mov r0, r6
+        add r10, #16
+        mov r8, r10
+        add r7, #8
+        mov r2, r7
+        b     0b
+4:      subs r4, #1
+        \filter
+        vld1.16      d0, [r8], r9
+        vqadd.s16    d0, d14
+        vqrshrun.s16 d0, q0, #7
+        vst1.32      d0[0], [r0], r1
+        regshuffle_d8
+        vld1.32    {d23[0]}, [r2], r3
+        bne 4b
+99:     vpop {d8-d15}
+        pop {r4-r10}
+        bx lr
+.endm
+
+function ff_hevc_put_qpel_v1_neon_8, export=1
+        hevc_put_qpel_vX_neon_8 qpel_filter_1
+endfunc
+
+function ff_hevc_put_qpel_v2_neon_8, export=1
+        hevc_put_qpel_vX_neon_8 qpel_filter_2
+endfunc
+
+function ff_hevc_put_qpel_v3_neon_8, export=1
+        hevc_put_qpel_vX_neon_8 qpel_filter_3
+endfunc
+
+
+function ff_hevc_put_qpel_uw_v1_neon_8, export=1
+        hevc_put_qpel_uw_vX_neon_8 qpel_filter_1
+endfunc
+
+function ff_hevc_put_qpel_uw_v2_neon_8, export=1
+        hevc_put_qpel_uw_vX_neon_8 qpel_filter_2
+endfunc
+
+function ff_hevc_put_qpel_uw_v3_neon_8, export=1
+        hevc_put_qpel_uw_vX_neon_8 qpel_filter_3
+endfunc
+
+.macro hevc_put_qpel_hX_neon_8 filter
+        push     {r4, r5, r6, r7}
+        ldr    r4, [sp, #16] // height
+        ldr    r5, [sp, #20] // width
+
+        vpush    {d8-d15}
+        sub       r2, #4
+        lsl       r1, #1
+        mov      r12, r4
+        mov       r6, r0
+        mov       r7, r2
+        cmp       r5, #4
+        beq       4f
+8:      subs      r4, #1
+        vextin8
+        \filter
+        vst1.16   {q7}, [r0], r1
+        bne       8b
+        subs      r5, #8
+        beq      99f
+        mov       r4, r12
+        add       r6, #16
+        mov       r0, r6
+        add       r7, #8
+        mov       r2, r7
+        cmp       r5, #4
+        bne       8b
+4:      subs      r4, #1
+        vextin8
+        \filter
+        vst1.16  d14, [r0], r1
+        bne       4b
+99:     vpop     {d8-d15}
+        pop      {r4, r5, r6, r7}
+        bx lr
+.endm
+
+.macro hevc_put_qpel_uw_hX_neon_8 filter
+        push     {r4-r10}
+        ldr       r5, [sp, #28] // width
+        ldr       r4, [sp, #32] // height
+        ldr       r8, [sp, #36] // src2
+        ldr       r9, [sp, #40] // src2stride
+        vpush    {d8-d15}
+        sub       r2, #4
+        mov      r12, r4
+        mov       r6, r0
+        mov       r7, r2
+        cmp       r8, #0
+        bne       .Lbi\@
+        cmp       r5, #4
+        beq       4f
+8:      subs      r4, #1
+        vextin8
+        \filter
+        vqrshrun.s16   d0, q7, #6
+        vst1.8    d0, [r0], r1
+        bne       8b
+        subs      r5, #8
+        beq      99f
+        mov       r4, r12
+        add       r6, #8
+        mov       r0, r6
+        add       r7, #8
+        mov       r2, r7
+        cmp       r5, #4
+        bne       8b
+4:      subs      r4, #1
+        vextin8
+        \filter
+        vqrshrun.s16   d0, q7, #6
+        vst1.32  d0[0], [r0], r1
+        bne       4b
+        b         99f
+.Lbi\@:
+        lsl       r9, #1
+        cmp       r5, #4
+        beq       4f
+        mov       r10, r8
+8:      subs      r4, #1
+        vextin8
+        \filter
+        vld1.16        {q0}, [r8], r9
+        vqadd.s16      q0, q7
+        vqrshrun.s16   d0, q0, #7
+        vst1.8         d0, [r0], r1
+        bne       8b
+        subs      r5, #8
+        beq      99f
+        mov       r4, r12
+        add       r6, #8
+        add       r10, #16
+        mov       r8, r10
+        mov       r0, r6
+        add       r7, #8
+        mov       r2, r7
+        cmp       r5, #4
+        bne       8b
+4:      subs      r4, #1
+        vextin8
+        \filter
+        vld1.16      d0, [r8], r9
+        vqadd.s16    d0, d14
+        vqrshrun.s16 d0, q0, #7
+        vst1.32      d0[0], [r0], r1
+        bne       4b
+99:     vpop     {d8-d15}
+        pop      {r4-r10}
+        bx lr
+.endm
+
+function ff_hevc_put_qpel_h1_neon_8, export=1
+        hevc_put_qpel_hX_neon_8 qpel_filter_1
+endfunc
+
+function ff_hevc_put_qpel_h2_neon_8, export=1
+        hevc_put_qpel_hX_neon_8 qpel_filter_2
+endfunc
+
+function ff_hevc_put_qpel_h3_neon_8, export=1
+        hevc_put_qpel_hX_neon_8 qpel_filter_3
+endfunc
+
+
+function ff_hevc_put_qpel_uw_h1_neon_8, export=1
+        hevc_put_qpel_uw_hX_neon_8 qpel_filter_1
+endfunc
+
+function ff_hevc_put_qpel_uw_h2_neon_8, export=1
+        hevc_put_qpel_uw_hX_neon_8 qpel_filter_2
+endfunc
+
+function ff_hevc_put_qpel_uw_h3_neon_8, export=1
+        hevc_put_qpel_uw_hX_neon_8 qpel_filter_3
+endfunc
+
+.macro hevc_put_qpel_hXvY_neon_8 filterh filterv
+        push   {r4, r5, r6, r7}
+        ldr    r4, [sp, #16] // height
+        ldr    r5, [sp, #20] // width
+
+        vpush {d8-d15}
+        sub       r2, #4
+        sub       r2, r2, r3, lsl #1
+        sub       r2, r3  // extra_before 3
+        lsl       r1, #1
+        mov       r12, r4
+        mov       r6, r0
+        mov       r7, r2
+0:      vextin8
+        \filterh q0
+        vextin8
+        \filterh q1
+        vextin8
+        \filterh q2
+        vextin8
+        \filterh q3
+        vextin8
+        \filterh q4
+        vextin8
+        \filterh q5
+        vextin8
+        \filterh q6
+        vextin8
+        \filterh q7
+        cmp r5, #4
+        beq 4f
+8:      subs  r4, #1
+        \filterv
+        vst1.16    {q8}, [r0], r1
+        regshuffle_q8
+        vextin8
+        \filterh q7
+        bne 8b
+        subs  r5, #8
+        beq 99f
+        mov r4, r12
+        add r6, #16
+        mov r0, r6
+        add r7, #8
+        mov r2, r7
+        b 0b
+4:      subs  r4, #1
+        \filterv
+        vst1.16    d16, [r0], r1
+        regshuffle_q8
+        vextin8
+        \filterh q7
+        bne 4b
+99:     vpop {d8-d15}
+        pop {r4, r5, r6, r7}
+        bx lr
+.endm
+
+.macro hevc_put_qpel_uw_hXvY_neon_8 filterh filterv
+        push     {r4-r10}
+        ldr       r5, [sp, #28] // width
+        ldr       r4, [sp, #32] // height
+        ldr       r8, [sp, #36] // src2
+        ldr       r9, [sp, #40] // src2stride
+        vpush {d8-d15}
+        sub       r2, #4
+        sub       r2, r2, r3, lsl #1
+        sub       r2, r3  // extra_before 3
+        mov       r12, r4
+        mov       r6, r0
+        mov       r7, r2
+        cmp       r8, #0
+        bne       .Lbi\@
+0:      vextin8
+        \filterh q0
+        vextin8
+        \filterh q1
+        vextin8
+        \filterh q2
+        vextin8
+        \filterh q3
+        vextin8
+        \filterh q4
+        vextin8
+        \filterh q5
+        vextin8
+        \filterh q6
+        vextin8
+        \filterh q7
+        cmp r5, #4
+        beq 4f
+8:      subs  r4, #1
+        \filterv
+        vqrshrun.s16   d0, q8, #6
+        vst1.8    d0, [r0], r1
+        regshuffle_q8
+        vextin8
+        \filterh q7
+        bne 8b
+        subs  r5, #8
+        beq 99f
+        mov r4, r12
+        add r6, #8
+        mov r0, r6
+        add r7, #8
+        mov r2, r7
+        b 0b
+4:      subs  r4, #1
+        \filterv
+        vqrshrun.s16   d0, q8, #6
+        vst1.32        d0[0], [r0], r1
+        regshuffle_q8
+        vextin8
+        \filterh q7
+        bne 4b
+        b   99f
+.Lbi\@: lsl      r9, #1
+        mov      r10, r8
+0:      vextin8
+        \filterh q0
+        vextin8
+        \filterh q1
+        vextin8
+        \filterh q2
+        vextin8
+        \filterh q3
+        vextin8
+        \filterh q4
+        vextin8
+        \filterh q5
+        vextin8
+        \filterh q6
+        vextin8
+        \filterh q7
+        cmp r5, #4
+        beq 4f
+8:      subs  r4, #1
+        \filterv
+        vld1.16        {q0}, [r8], r9
+        vqadd.s16      q0, q8
+        vqrshrun.s16   d0, q0, #7
+        vst1.8         d0, [r0], r1
+        regshuffle_q8
+        vextin8
+        \filterh q7
+        bne 8b
+        subs  r5, #8
+        beq 99f
+        mov r4, r12
+        add r6, #8
+        mov r0, r6
+        add r10, #16
+        mov r8, r10
+        add r7, #8
+        mov r2, r7
+        b 0b
+4:      subs  r4, #1
+        \filterv
+        vld1.16      d0, [r8], r9
+        vqadd.s16    d0, d16
+        vqrshrun.s16 d0, q0, #7
+        vst1.32      d0[0], [r0], r1
+        regshuffle_q8
+        vextin8
+        \filterh q7
+        bne 4b
+99:     vpop {d8-d15}
+        pop {r4-r10}
+        bx lr
+.endm
+
+
+function ff_hevc_put_qpel_h1v1_neon_8, export=1
+        hevc_put_qpel_hXvY_neon_8 qpel_filter_1 qpel_filter_1_32b
+endfunc
+
+function ff_hevc_put_qpel_h2v1_neon_8, export=1
+        hevc_put_qpel_hXvY_neon_8 qpel_filter_2 qpel_filter_1_32b
+endfunc
+
+function ff_hevc_put_qpel_h3v1_neon_8, export=1
+        hevc_put_qpel_hXvY_neon_8 qpel_filter_3 qpel_filter_1_32b
+endfunc
+
+function ff_hevc_put_qpel_h1v2_neon_8, export=1
+        hevc_put_qpel_hXvY_neon_8 qpel_filter_1 qpel_filter_2_32b
+endfunc
+
+function ff_hevc_put_qpel_h2v2_neon_8, export=1
+        hevc_put_qpel_hXvY_neon_8 qpel_filter_2 qpel_filter_2_32b
+endfunc
+
+function ff_hevc_put_qpel_h3v2_neon_8, export=1
+        hevc_put_qpel_hXvY_neon_8 qpel_filter_3 qpel_filter_2_32b
+endfunc
+
+function ff_hevc_put_qpel_h1v3_neon_8, export=1
+        hevc_put_qpel_hXvY_neon_8 qpel_filter_1 qpel_filter_3_32b
+endfunc
+
+function ff_hevc_put_qpel_h2v3_neon_8, export=1
+        hevc_put_qpel_hXvY_neon_8 qpel_filter_2 qpel_filter_3_32b
+endfunc
+
+function ff_hevc_put_qpel_h3v3_neon_8, export=1
+        hevc_put_qpel_hXvY_neon_8 qpel_filter_3 qpel_filter_3_32b
+endfunc
+
+
+function ff_hevc_put_qpel_uw_h1v1_neon_8, export=1
+        hevc_put_qpel_uw_hXvY_neon_8 qpel_filter_1 qpel_filter_1_32b
+endfunc
+
+function ff_hevc_put_qpel_uw_h2v1_neon_8, export=1
+        hevc_put_qpel_uw_hXvY_neon_8 qpel_filter_2 qpel_filter_1_32b
+endfunc
+
+function ff_hevc_put_qpel_uw_h3v1_neon_8, export=1
+        hevc_put_qpel_uw_hXvY_neon_8 qpel_filter_3 qpel_filter_1_32b
+endfunc
+
+function ff_hevc_put_qpel_uw_h1v2_neon_8, export=1
+        hevc_put_qpel_uw_hXvY_neon_8 qpel_filter_1 qpel_filter_2_32b
+endfunc
+
+function ff_hevc_put_qpel_uw_h2v2_neon_8, export=1
+        hevc_put_qpel_uw_hXvY_neon_8 qpel_filter_2 qpel_filter_2_32b
+endfunc
+
+function ff_hevc_put_qpel_uw_h3v2_neon_8, export=1
+        hevc_put_qpel_uw_hXvY_neon_8 qpel_filter_3 qpel_filter_2_32b
+endfunc
+
+function ff_hevc_put_qpel_uw_h1v3_neon_8, export=1
+        hevc_put_qpel_uw_hXvY_neon_8 qpel_filter_1 qpel_filter_3_32b
+endfunc
+
+function ff_hevc_put_qpel_uw_h2v3_neon_8, export=1
+        hevc_put_qpel_uw_hXvY_neon_8 qpel_filter_2 qpel_filter_3_32b
+endfunc
+
+function ff_hevc_put_qpel_uw_h3v3_neon_8, export=1
+        hevc_put_qpel_uw_hXvY_neon_8 qpel_filter_3 qpel_filter_3_32b
+endfunc
+
+.macro init_put_pixels
+        pld    [r1]
+        pld    [r1, r2]
+        mov    r12, MAX_PB_SIZE
+        lsl    r12, #1
+.endm
+
+function ff_hevc_put_pixels_w2_neon_8, export=1
+        init_put_pixels
+        vmov.u8      d5, #255
+        vshr.u64     d5, #32
+0:      subs r3, #1
+        vld1.32     {d0[0]}, [r1], r2
+        pld [r1]
+        vld1.32     d6, [r0]
+        vshll.u8    q0, d0, #6
+        vbit        d6, d0, d5
+        vst1.32     d6, [r0], r12
+        bne 0b
+        bx lr
+endfunc
+
+function ff_hevc_put_pixels_w4_neon_8, export=1
+        init_put_pixels
+0:      subs r3, #2
+        vld1.32   {d0[0]}, [r1], r2
+        vld1.32   {d0[1]}, [r1], r2
+        pld       [r1]
+        pld       [r1, r2]
+        vshll.u8   q0, d0, #6
+        vst1.64   {d0}, [r0], r12
+        vst1.64   {d1}, [r0], r12
+        bne 0b
+        bx lr
+endfunc
+
+function ff_hevc_put_pixels_w6_neon_8, export=1
+        init_put_pixels
+        vmov.u8      q10, #255
+        vshr.u64     d21, #32
+0:      subs r3, #1
+        vld1.16     {d0}, [r1], r2
+        pld [r1]
+        vshll.u8    q0, d0, #6
+        vld1.8      {q12}, [r0]
+        vbit        q12, q0, q10
+        vst1.8      {q12}, [r0], r12
+        bne 0b
+        bx lr
+endfunc
+
+function ff_hevc_put_pixels_w8_neon_8, export=1
+        init_put_pixels
+0:      subs r3, #2
+        vld1.8   {d0}, [r1], r2
+        vld1.8   {d2}, [r1], r2
+        pld        [r1]
+        pld        [r1, r2]
+        vshll.u8   q0, d0, #6
+        vshll.u8   q1, d2, #6
+        vst1.16   {q0}, [r0], r12
+        vst1.16   {q1}, [r0], r12
+        bne 0b
+        bx lr
+endfunc
+
+function ff_hevc_put_pixels_w12_neon_8, export=1
+        init_put_pixels
+0:      subs r3, #2
+        vld1.64    {d0}, [r1]
+        add       r1, #8
+        vld1.32   {d1[0]}, [r1], r2
+        sub       r1, #8
+        vld1.64    {d2}, [r1]
+        add       r1, #8
+        vld1.32   {d1[1]}, [r1], r2
+        sub       r1, #8
+        pld       [r1]
+        pld       [r1, r2]
+        vshll.u8  q8, d0, #6
+        vshll.u8  q9, d1, #6
+        vshll.u8  q10, d2, #6
+        vmov      d22, d19
+        vst1.64   {d16, d17, d18}, [r0], r12
+        vst1.64   {d20, d21, d22}, [r0], r12
+        bne 0b
+        bx lr
+endfunc
+
+function ff_hevc_put_pixels_w16_neon_8, export=1
+        init_put_pixels
+0:      subs r3, #2
+        vld1.8   {q0}, [r1], r2
+        vld1.8   {q1}, [r1], r2
+        pld       [r1]
+        pld       [r1, r2]
+        vshll.u8  q8, d0, #6
+        vshll.u8  q9, d1, #6
+        vshll.u8  q10, d2, #6
+        vshll.u8  q11, d3, #6
+        vst1.8    {q8, q9}, [r0], r12
+        vst1.8    {q10, q11}, [r0], r12
+        bne 0b
+        bx lr
+endfunc
+
+function ff_hevc_put_pixels_w24_neon_8, export=1
+        init_put_pixels
+0:      subs r3, #1
+        vld1.8   {d0, d1, d2}, [r1], r2
+        pld       [r1]
+        vshll.u8  q10, d0, #6
+        vshll.u8  q11, d1, #6
+        vshll.u8  q12, d2, #6
+        vstm     r0, {q10, q11, q12}
+        add      r0, r12
+        bne 0b
+        bx lr
+endfunc
+
+function ff_hevc_put_pixels_w32_neon_8, export=1
+        init_put_pixels
+0:      subs r3, #1
+        vld1.8 {q0, q1}, [r1], r2
+        pld       [r1]
+        vshll.u8  q8, d0, #6
+        vshll.u8  q9, d1, #6
+        vshll.u8  q10, d2, #6
+        vshll.u8  q11, d3, #6
+        vstm    r0, {q8, q9, q10, q11}
+        add     r0, r12
+        bne 0b
+        bx lr
+endfunc
+
+function ff_hevc_put_pixels_w48_neon_8, export=1
+        init_put_pixels
+0:      subs r3, #1
+        vld1.8    {q0, q1}, [r1]
+        add r1, #32
+        vld1.8    {q2}, [r1], r2
+        sub r1, #32
+        pld       [r1]
+        vshll.u8  q8, d0, #6
+        vshll.u8  q9, d1, #6
+        vshll.u8  q10, d2, #6
+        vshll.u8  q11, d3, #6
+        vshll.u8  q12, d4, #6
+        vshll.u8  q13, d5, #6
+        vstm r0, {q8, q9, q10, q11, q12, q13}
+        add  r0, r12
+        bne 0b
+        bx lr
+endfunc
+
+function ff_hevc_put_pixels_w64_neon_8, export=1
+        init_put_pixels
+0:      subs r3, #1
+        vld1.8    {q0, q1}, [r1]
+        add      r1, #32
+        vld1.8    {q2, q3}, [r1], r2
+        sub      r1, #32
+        pld       [r1]
+        vshll.u8  q8, d0, #6
+        vshll.u8  q9, d1, #6
+        vshll.u8  q10, d2, #6
+        vshll.u8  q11, d3, #6
+        vshll.u8  q12, d4, #6
+        vshll.u8  q13, d5, #6
+        vshll.u8  q14, d6, #6
+        vshll.u8  q15, d7, #6
+        vstm    r0, {q8, q9, q10, q11, q12, q13, q14, q15}
+        add r0, r12
+        bne 0b
+        bx lr
+endfunc
+
+function ff_hevc_put_qpel_uw_pixels_neon_8, export=1
+        push   {r4-r9}
+        ldr    r5, [sp, #24] // width
+        ldr    r4, [sp, #28] // height
+        ldr    r8, [sp, #32] // src2
+        ldr    r9, [sp, #36] // src2stride
+        vpush {d8-d15}
+        cmp    r8, #0
+        bne    2f
+1:      subs r4, #1
+        vld1.8     {d0}, [r2], r3
+        vst1.8      d0, [r0], r1
+        bne 1b
+        vpop {d8-d15}
+        pop   {r4-r9}
+        bx lr
+2:      subs  r4, #1
+        vld1.8         {d0}, [r2], r3
+        vld1.16        {q1}, [r8], r9
+        vshll.u8       q0, d0, #6
+        vqadd.s16      q0, q1
+        vqrshrun.s16   d0, q0, #7
+        vst1.8      d0, [r0], r1
+        bne 2b
+        vpop {d8-d15}
+        pop   {r4-r9}
+        bx lr
+endfunc
+
+.macro put_qpel_uw_pixels width, regs, regs2, regs3, regs4
+function ff_hevc_put_qpel_uw_pixels_w\width\()_neon_8, export=1
+        ldr    r12, [sp] // height
+1:      subs   r12, #4
+        vld1.32     {\regs}  , [r2], r3
+        vld1.32     {\regs2} , [r2], r3
+        vld1.32     {\regs3} , [r2], r3
+        vld1.32     {\regs4} , [r2], r3
+        vst1.32     {\regs}  , [r0], r1
+        vst1.32     {\regs2} , [r0], r1
+        vst1.32     {\regs3} , [r0], r1
+        vst1.32     {\regs4} , [r0], r1
+        bne 1b
+        bx lr
+endfunc
+.endm
+
+.macro put_qpel_uw_pixels_m width, regs, regs2, regs3, regs4
+function ff_hevc_put_qpel_uw_pixels_w\width\()_neon_8, export=1
+        push   {r4-r5}
+        ldr    r12, [sp, #8] // height
+1:      subs r12, #2
+        mov      r4, r2
+        vld1.32   {\regs} , [r2]!
+        vld1.32   {\regs2} , [r2]
+        add      r2, r4, r3
+        mov      r4, r2
+        vld1.32   {\regs3} , [r2]!
+        vld1.32   {\regs4} , [r2]
+        add      r2, r4, r3
+        mov      r5, r0
+        vst1.32   {\regs} , [r0]!
+        vst1.32   {\regs2} , [r0]
+        add      r0, r5, r1
+        mov      r5, r0
+        vst1.32   {\regs3} , [r0]!
+        vst1.32   {\regs4} , [r0]
+        add      r0, r5, r1
+        bne 1b
+        pop   {r4-r5}
+        bx lr
+endfunc
+.endm
+
+put_qpel_uw_pixels    4, d0[0], d0[1], d1[0], d1[1]
+put_qpel_uw_pixels    8, d0,    d1,    d2,    d3
+put_qpel_uw_pixels_m 12, d0,    d1[0], d2,    d3[0]
+put_qpel_uw_pixels   16, q0,    q1,    q2,    q3
+put_qpel_uw_pixels   24, d0-d2, d3-d5, d16-d18, d19-d21
+put_qpel_uw_pixels   32, q0-q1, q2-q3, q8-q9, q10-q11
+put_qpel_uw_pixels_m 48, q0-q1, q2,    q8-q9, q10
+put_qpel_uw_pixels_m 64, q0-q1, q2-q3, q8-q9, q10-q11




More information about the ffmpeg-cvslog mailing list