From 8992cb1d0d07edc33d274f6d7924ecdf6f83d994 Mon Sep 17 00:00:00 2001
From: Tim Redfern <tim@eclectronics.org>
Date: Thu, 5 Sep 2013 17:57:22 +0100
Subject: making act segmenter

---
 ffmpeg/libavcodec/x86/fmtconvert_init.c | 148 ++++++++++++++++++++++++++++++++
 1 file changed, 148 insertions(+)
 create mode 100644 ffmpeg/libavcodec/x86/fmtconvert_init.c

(limited to 'ffmpeg/libavcodec/x86/fmtconvert_init.c')

diff --git a/ffmpeg/libavcodec/x86/fmtconvert_init.c b/ffmpeg/libavcodec/x86/fmtconvert_init.c
new file mode 100644
index 0000000..4a4c017
--- /dev/null
+++ b/ffmpeg/libavcodec/x86/fmtconvert_init.c
@@ -0,0 +1,148 @@
+/*
+ * Format Conversion Utils
+ * Copyright (c) 2000, 2001 Fabrice Bellard
+ * Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * 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
+ *
+ * MMX optimization by Nick Kurshev <nickols_k@mail.ru>
+ */
+
+#include "libavutil/attributes.h"
+#include "libavutil/cpu.h"
+#include "libavutil/x86/asm.h"
+#include "libavutil/x86/cpu.h"
+#include "libavcodec/fmtconvert.h"
+
+#if HAVE_YASM
+
+void ff_int32_to_float_fmul_scalar_sse (float *dst, const int *src, float mul, int len);
+void ff_int32_to_float_fmul_scalar_sse2(float *dst, const int *src, float mul, int len);
+
+void ff_float_to_int16_3dnow(int16_t *dst, const float *src, long len);
+void ff_float_to_int16_sse  (int16_t *dst, const float *src, long len);
+void ff_float_to_int16_sse2 (int16_t *dst, const float *src, long len);
+
+void ff_float_to_int16_step_3dnow(int16_t *dst, const float *src, long len, long step);
+void ff_float_to_int16_step_sse  (int16_t *dst, const float *src, long len, long step);
+void ff_float_to_int16_step_sse2 (int16_t *dst, const float *src, long len, long step);
+
+void ff_float_to_int16_interleave2_3dnow(int16_t *dst, const float **src, long len);
+void ff_float_to_int16_interleave2_sse  (int16_t *dst, const float **src, long len);
+void ff_float_to_int16_interleave2_sse2 (int16_t *dst, const float **src, long len);
+
+void ff_float_to_int16_interleave6_sse(int16_t *dst, const float **src, int len);
+void ff_float_to_int16_interleave6_3dnow(int16_t *dst, const float **src, int len);
+void ff_float_to_int16_interleave6_3dnowext(int16_t *dst, const float **src, int len);
+
+#define ff_float_to_int16_interleave6_sse2 ff_float_to_int16_interleave6_sse
+
+#define FLOAT_TO_INT16_INTERLEAVE(cpu) \
+/* gcc pessimizes register allocation if this is in the same function as float_to_int16_interleave_sse2*/\
+static av_noinline void float_to_int16_interleave_misc_##cpu(int16_t *dst, const float **src, long len, int channels){\
+    int c;\
+    for(c=0; c<channels; c++){\
+        ff_float_to_int16_step_##cpu(dst+c, src[c], len, channels);\
+    }\
+}\
+\
+static void float_to_int16_interleave_##cpu(int16_t *dst, const float **src, long len, int channels){\
+    if(channels==1)\
+        ff_float_to_int16_##cpu(dst, src[0], len);\
+    else if(channels==2){\
+        ff_float_to_int16_interleave2_##cpu(dst, src, len);\
+    }else if(channels==6){\
+        ff_float_to_int16_interleave6_##cpu(dst, src, len);\
+    }else\
+        float_to_int16_interleave_misc_##cpu(dst, src, len, channels);\
+}
+
+FLOAT_TO_INT16_INTERLEAVE(3dnow)
+FLOAT_TO_INT16_INTERLEAVE(sse)
+FLOAT_TO_INT16_INTERLEAVE(sse2)
+
+static void float_to_int16_interleave_3dnowext(int16_t *dst, const float **src,
+                                               long len, int channels)
+{
+    if(channels==6)
+        ff_float_to_int16_interleave6_3dnowext(dst, src, len);
+    else
+        float_to_int16_interleave_3dnow(dst, src, len, channels);
+}
+
+void ff_float_interleave2_mmx(float *dst, const float **src, unsigned int len);
+void ff_float_interleave2_sse(float *dst, const float **src, unsigned int len);
+
+void ff_float_interleave6_mmx(float *dst, const float **src, unsigned int len);
+void ff_float_interleave6_sse(float *dst, const float **src, unsigned int len);
+
+static void float_interleave_mmx(float *dst, const float **src,
+                                 unsigned int len, int channels)
+{
+    if (channels == 2) {
+        ff_float_interleave2_mmx(dst, src, len);
+    } else if (channels == 6)
+        ff_float_interleave6_mmx(dst, src, len);
+    else
+        ff_float_interleave_c(dst, src, len, channels);
+}
+
+static void float_interleave_sse(float *dst, const float **src,
+                                 unsigned int len, int channels)
+{
+    if (channels == 2) {
+        ff_float_interleave2_sse(dst, src, len);
+    } else if (channels == 6)
+        ff_float_interleave6_sse(dst, src, len);
+    else
+        ff_float_interleave_c(dst, src, len, channels);
+}
+#endif /* HAVE_YASM */
+
+av_cold void ff_fmt_convert_init_x86(FmtConvertContext *c, AVCodecContext *avctx)
+{
+#if HAVE_YASM
+    int mm_flags = av_get_cpu_flags();
+
+    if (EXTERNAL_MMX(mm_flags)) {
+        c->float_interleave = float_interleave_mmx;
+
+        if (EXTERNAL_AMD3DNOW(mm_flags)) {
+            if(!(avctx->flags & CODEC_FLAG_BITEXACT)){
+                c->float_to_int16 = ff_float_to_int16_3dnow;
+                c->float_to_int16_interleave = float_to_int16_interleave_3dnow;
+            }
+        }
+        if (EXTERNAL_AMD3DNOWEXT(mm_flags)) {
+            if(!(avctx->flags & CODEC_FLAG_BITEXACT)){
+                c->float_to_int16_interleave = float_to_int16_interleave_3dnowext;
+            }
+        }
+        if (EXTERNAL_SSE(mm_flags)) {
+            c->int32_to_float_fmul_scalar = ff_int32_to_float_fmul_scalar_sse;
+            c->float_to_int16 = ff_float_to_int16_sse;
+            c->float_to_int16_interleave = float_to_int16_interleave_sse;
+            c->float_interleave = float_interleave_sse;
+        }
+        if (EXTERNAL_SSE2(mm_flags)) {
+            c->int32_to_float_fmul_scalar = ff_int32_to_float_fmul_scalar_sse2;
+            c->float_to_int16 = ff_float_to_int16_sse2;
+            c->float_to_int16_interleave = float_to_int16_interleave_sse2;
+        }
+    }
+#endif /* HAVE_YASM */
+}
-- 
cgit v1.2.3