From: James Yu <[email protected]>

Using NEON intrinsics to support ARMv7 and ARMv8 simultaneously.
Rewrite ARMv7 assembly to intinsics version.

Toolchain:
armv7: gcc-linaro-arm-linux-gnueabihf-4.8-2014.01_linux
armv8: gcc-linaro-aarch64-linux-gnu-4.8-2013.10_linux

[RFC] Currently, libavcodec had a aarch64 dir includes all related assembly
code. However, I put this file under libavcodec/arm dir first, since this
support ARMv7 and ARMv8 simultaneously. And RFC how could we port and
optimize libav to aarch64.

Add libavcodec/arm/mpegvideo_neon.c.

Signed-off-by: James Yu <[email protected]>
---
 libavcodec/aarch64/Makefile     |    1 +
 libavcodec/arm/mpegvideo_neon.c |  127 +++++++++++++++++++++++++++++++++++++++
 2 files changed, 128 insertions(+)
 create mode 100644 libavcodec/arm/mpegvideo_neon.c

diff --git a/libavcodec/aarch64/Makefile b/libavcodec/aarch64/Makefile
index b0ebeb3..97dfb79 100644
--- a/libavcodec/aarch64/Makefile
+++ b/libavcodec/aarch64/Makefile
@@ -24,3 +24,4 @@ NEON-OBJS-$(CONFIG_MPEGAUDIODSP)        += 
aarch64/mpegaudiodsp_neon.o
 NEON-OBJS-$(CONFIG_MDCT)                += aarch64/mdct_neon.o
 
 NEON-OBJS-$(CONFIG_VORBIS_DECODER)      += aarch64/vorbisdsp_neon.o
+NEON-OBJS-$(CONFIG_MPEGVIDEO)          += arm/mpegvideo_neon.o
diff --git a/libavcodec/arm/mpegvideo_neon.c b/libavcodec/arm/mpegvideo_neon.c
new file mode 100644
index 0000000..dcc1204
--- /dev/null
+++ b/libavcodec/arm/mpegvideo_neon.c
@@ -0,0 +1,127 @@
+/*
+ * Copyright (c) 2010 Mans Rullgard
+ * Copyright (c) 2014 James Yu <[email protected]>
+ *
+ * This file is part of Libav.
+ *
+ * Libav 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.
+ *
+ * Libav 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 Libav; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include <arm_neon.h>
+#include "libavutil/avassert.h"
+#include "libavcodec/mpegvideo.h"
+
+void ff_dct_unquantize_h263_inter_neon(MpegEncContext *s, int16_t *block,
+                                       int n, int qscale);
+void ff_dct_unquantize_h263_intra_neon(MpegEncContext *s, int16_t *block,
+                                       int n, int qscale);
+
+static void inline ff_dct_unquantize_h263_neon(int qmul, int qadd,
+                                       int nCoeffs, int16_t *block)
+{
+    int16x8_t q0s16, q2s16, q3s16, q8s16, q10s16, q11s16, q13s16;
+    int16x8_t q14s16, q15s16, qzs16;
+    int16x4_t d0s16, d2s16, d3s16, dzs16;
+    uint16x8_t q1u16, q9u16;
+    uint16x4_t d1u16;
+
+    dzs16 = vdup_n_s16(0);
+    qzs16 = vdupq_n_s16(0);
+
+    q15s16 = vdupq_n_s16(qmul);
+    q14s16 = vdupq_n_s16(qadd);
+    q13s16 = vnegq_s16(q14s16);
+
+    if (nCoeffs > 4) {
+        for (; nCoeffs > 8; nCoeffs -= 16, block += 16) {
+            q0s16 = vld1q_s16(block);
+            q3s16 = vreinterpretq_s16_u16(vcltq_s16(q0s16, qzs16));
+            q8s16 = vld1q_s16(block + 8);
+            q1u16 = vceqq_s16(q0s16, qzs16);
+            q2s16 = vmulq_s16(q0s16, q15s16);
+            q11s16 = vreinterpretq_s16_u16(vcltq_s16(q8s16, qzs16));
+            q10s16 = vmulq_s16(q8s16, q15s16);
+            q3s16 = vbslq_s16(vreinterpretq_u16_s16(q3s16), q13s16, q14s16);
+            q11s16 = vbslq_s16(vreinterpretq_u16_s16(q11s16), q13s16, q14s16);
+            q2s16 = vaddq_s16(q2s16, q3s16);
+            q9u16 = vceqq_s16(q8s16, qzs16);
+            q10s16 = vaddq_s16(q10s16, q11s16);
+            q0s16 = vbslq_s16(q1u16, q0s16, q2s16);
+            q8s16 = vbslq_s16(q9u16, q8s16, q10s16);
+            vst1q_s16(block, q0s16);
+            vst1q_s16(block + 8, q8s16);
+        }
+    }
+    if (nCoeffs <= 0)
+        return;
+
+    d0s16 = vld1_s16(block);
+    d3s16 = vreinterpret_s16_u16(vclt_s16(d0s16, dzs16));
+    d1u16 = vceq_s16(d0s16, dzs16);
+    d2s16 = vmul_s16(d0s16, vget_high_s16(q15s16));
+    d3s16 = vbsl_s16(vreinterpret_u16_s16(d3s16),
+                     vget_high_s16(q13s16), vget_high_s16(q14s16));
+    d2s16 = vadd_s16(d2s16, d3s16);
+    d0s16 = vbsl_s16(d1u16, d0s16, d2s16);
+    vst1_s16(block, d0s16);
+}
+
+void ff_dct_unquantize_h263_inter_neon(MpegEncContext *s, int16_t *block,
+                                       int n, int qscale)
+{
+    int qmul, qadd;
+    int nCoeffs;
+
+    qadd = (qscale - 1) | 1;
+    qmul = qscale << 1;
+
+    nCoeffs = s->inter_scantable.raster_end[s->block_last_index[n]];
+
+    ff_dct_unquantize_h263_neon(qmul, qadd, nCoeffs, block);
+}
+
+void ff_dct_unquantize_h263_intra_neon(MpegEncContext *s, int16_t *block,
+                                       int n, int qscale)
+{
+    int qmul, qadd;
+    int nCoeffs, blk0;
+
+    av_assert0(s->block_last_index[n] >= 0);
+
+    qmul = qscale << 1;
+
+    if (!s->h263_aic) {
+        if (n < 4) {
+            block[0] = block[0] * s->y_dc_scale;
+        } else
+            block[0] = block[0] * s->c_dc_scale;
+        qadd = (qscale - 1) | 1;
+    } else
+        qadd = 0;
+
+    if (s->ac_pred) {
+        nCoeffs = 63;
+    } else {
+        nCoeffs = s->inter_scantable.raster_end[s->block_last_index[n]];
+        if (nCoeffs <= 0)
+            return;
+    }
+
+    blk0 = block[0];
+
+    ff_dct_unquantize_h263_neon(qmul, qadd, nCoeffs + 1, block);
+
+    block[0] = blk0;
+}
-- 
1.7.9.5

_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to