ref: 510e0ab4679ef76f1d7c5ed1b792f8012444f0cc
parent: 83a91e789ccd9fc0a9d05cb4357131ad18ba3b04
author: John Koleszar <jkoleszar@google.com>
date: Thu Jan 12 11:55:44 EST 2012
RTCD: add FDCT functions This commit continues the process of converting to the new RTCD system. Change-Id: I3f9c07db65eb206f6363d21bdb80e871570da767
--- a/vp8/common/rtcd_defs.sh
+++ b/vp8/common/rtcd_defs.sh
@@ -392,5 +392,20 @@
specialize vp8_ssim_parms_16x16 $sse2_on_x86_64
fi
+#
+# Forward DCT
+#
+prototype void vp8_short_fdct4x4 "short *input, short *output, int pitch"
+specialize vp8_short_fdct4x4 mmx sse2 media neon
+vp8_short_fdct4x4_media=vp8_short_fdct4x4_armv6
+
+prototype void vp8_short_fdct8x4 "short *input, short *output, int pitch"
+specialize vp8_short_fdct8x4 mmx sse2 media neon
+vp8_short_fdct8x4_media=vp8_short_fdct8x4_armv6
+
+prototype void vp8_short_walsh4x4 "short *input, short *output, int pitch"
+specialize vp8_short_walsh4x4 mmx sse2 media neon
+vp8_short_walsh4x4_media=vp8_short_walsh4x4_armv6
+
# End of encoder only functions
fi
--- a/vp8/encoder/arm/arm_csystemdependent.c
+++ b/vp8/encoder/arm/arm_csystemdependent.c
@@ -32,12 +32,6 @@
#if HAVE_MEDIA
if (flags & HAS_MEDIA)
{- cpi->rtcd.fdct.short4x4 = vp8_short_fdct4x4_armv6;
- cpi->rtcd.fdct.short8x4 = vp8_short_fdct8x4_armv6;
- cpi->rtcd.fdct.fast4x4 = vp8_short_fdct4x4_armv6;
- cpi->rtcd.fdct.fast8x4 = vp8_short_fdct8x4_armv6;
- cpi->rtcd.fdct.walsh_short4x4 = vp8_short_walsh4x4_armv6;
-
/*cpi->rtcd.encodemb.berr = vp8_block_error_c;
cpi->rtcd.encodemb.mberr = vp8_mbblock_error_c;
cpi->rtcd.encodemb.mbuverr = vp8_mbuverror_c;*/
@@ -53,12 +47,6 @@
#if HAVE_NEON
if (flags & HAS_NEON)
{- cpi->rtcd.fdct.short4x4 = vp8_short_fdct4x4_neon;
- cpi->rtcd.fdct.short8x4 = vp8_short_fdct8x4_neon;
- cpi->rtcd.fdct.fast4x4 = vp8_short_fdct4x4_neon;
- cpi->rtcd.fdct.fast8x4 = vp8_short_fdct8x4_neon;
- cpi->rtcd.fdct.walsh_short4x4 = vp8_short_walsh4x4_neon;
-
/*cpi->rtcd.encodemb.berr = vp8_block_error_c;
cpi->rtcd.encodemb.mberr = vp8_mbblock_error_c;
cpi->rtcd.encodemb.mbuverr = vp8_mbuverror_c;*/
--- a/vp8/encoder/arm/dct_arm.c
+++ b/vp8/encoder/arm/dct_arm.c
@@ -9,7 +9,7 @@
*/
#include "vpx_config.h"
-#include "vp8/encoder/dct.h"
+#include "vpx_rtcd.h"
#if HAVE_MEDIA
--- a/vp8/encoder/arm/dct_arm.h
+++ /dev/null
@@ -1,65 +1,0 @@
-/*
- * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- * Use of this source code is governed by a BSD-style license
- * that can be found in the LICENSE file in the root of the source
- * tree. An additional intellectual property rights grant can be found
- * in the file PATENTS. All contributing project authors may
- * be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef DCT_ARM_H
-#define DCT_ARM_H
-
-#if HAVE_MEDIA
-extern prototype_fdct(vp8_short_walsh4x4_armv6);
-extern prototype_fdct(vp8_short_fdct4x4_armv6);
-extern prototype_fdct(vp8_short_fdct8x4_armv6);
-
-#if !CONFIG_RUNTIME_CPU_DETECT
-#undef vp8_fdct_walsh_short4x4
-#define vp8_fdct_walsh_short4x4 vp8_short_walsh4x4_armv6
-
-#undef vp8_fdct_short4x4
-#define vp8_fdct_short4x4 vp8_short_fdct4x4_armv6
-
-#undef vp8_fdct_short8x4
-#define vp8_fdct_short8x4 vp8_short_fdct8x4_armv6
-
-#undef vp8_fdct_fast4x4
-#define vp8_fdct_fast4x4 vp8_short_fdct4x4_armv6
-
-#undef vp8_fdct_fast8x4
-#define vp8_fdct_fast8x4 vp8_short_fdct8x4_armv6
-#endif
-
-#endif /* HAVE_MEDIA */
-
-#if HAVE_NEON
-extern prototype_fdct(vp8_short_fdct4x4_neon);
-extern prototype_fdct(vp8_short_fdct8x4_neon);
-extern prototype_fdct(vp8_fast_fdct4x4_neon);
-extern prototype_fdct(vp8_fast_fdct8x4_neon);
-extern prototype_fdct(vp8_short_walsh4x4_neon);
-
-#if !CONFIG_RUNTIME_CPU_DETECT
-#undef vp8_fdct_short4x4
-#define vp8_fdct_short4x4 vp8_short_fdct4x4_neon
-
-#undef vp8_fdct_short8x4
-#define vp8_fdct_short8x4 vp8_short_fdct8x4_neon
-
-#undef vp8_fdct_fast4x4
-#define vp8_fdct_fast4x4 vp8_short_fdct4x4_neon
-
-#undef vp8_fdct_fast8x4
-#define vp8_fdct_fast8x4 vp8_short_fdct8x4_neon
-
-#undef vp8_fdct_walsh_short4x4
-#define vp8_fdct_walsh_short4x4 vp8_short_walsh4x4_neon
-#endif
-
-#endif /* HAVE_NEON */
-
-#endif
--- a/vp8/encoder/block.h
+++ b/vp8/encoder/block.h
@@ -117,8 +117,8 @@
int optimize;
int q_index;
- void (*vp8_short_fdct4x4)(short *input, short *output, int pitch);
- void (*vp8_short_fdct8x4)(short *input, short *output, int pitch);
+ void (*short_fdct4x4)(short *input, short *output, int pitch);
+ void (*short_fdct8x4)(short *input, short *output, int pitch);
void (*short_walsh4x4)(short *input, short *output, int pitch);
void (*quantize_b)(BLOCK *b, BLOCKD *d);
void (*quantize_b_pair)(BLOCK *b1, BLOCK *b2, BLOCKD *d0, BLOCKD *d1);
--- a/vp8/encoder/dct.h
+++ /dev/null
@@ -1,65 +1,0 @@
-/*
- * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- * Use of this source code is governed by a BSD-style license
- * that can be found in the LICENSE file in the root of the source
- * tree. An additional intellectual property rights grant can be found
- * in the file PATENTS. All contributing project authors may
- * be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef __INC_DCT_H
-#define __INC_DCT_H
-
-#define prototype_fdct(sym) void (sym)(short *input, short *output, int pitch)
-
-#if ARCH_X86 || ARCH_X86_64
-#include "x86/dct_x86.h"
-#endif
-
-#if ARCH_ARM
-#include "arm/dct_arm.h"
-#endif
-
-#ifndef vp8_fdct_short4x4
-#define vp8_fdct_short4x4 vp8_short_fdct4x4_c
-#endif
-extern prototype_fdct(vp8_fdct_short4x4);
-
-#ifndef vp8_fdct_short8x4
-#define vp8_fdct_short8x4 vp8_short_fdct8x4_c
-#endif
-extern prototype_fdct(vp8_fdct_short8x4);
-
-// There is no fast4x4 (for now)
-#ifndef vp8_fdct_fast4x4
-#define vp8_fdct_fast4x4 vp8_short_fdct4x4_c
-#endif
-
-#ifndef vp8_fdct_fast8x4
-#define vp8_fdct_fast8x4 vp8_short_fdct8x4_c
-#endif
-
-#ifndef vp8_fdct_walsh_short4x4
-#define vp8_fdct_walsh_short4x4 vp8_short_walsh4x4_c
-#endif
-extern prototype_fdct(vp8_fdct_walsh_short4x4);
-
-typedef prototype_fdct(*vp8_fdct_fn_t);
-typedef struct
-{- vp8_fdct_fn_t short4x4;
- vp8_fdct_fn_t short8x4;
- vp8_fdct_fn_t fast4x4;
- vp8_fdct_fn_t fast8x4;
- vp8_fdct_fn_t walsh_short4x4;
-} vp8_fdct_rtcd_vtable_t;
-
-#if CONFIG_RUNTIME_CPU_DETECT
-#define FDCT_INVOKE(ctx,fn) (ctx)->fn
-#else
-#define FDCT_INVOKE(ctx,fn) vp8_fdct_##fn
-#endif
-
-#endif
--- a/vp8/encoder/encodeintra.c
+++ b/vp8/encoder/encodeintra.c
@@ -15,7 +15,6 @@
#include "vp8/common/reconintra4x4.h"
#include "encodemb.h"
#include "vp8/common/invtrans.h"
-#include "dct.h"
#include "encodeintra.h"
@@ -70,7 +69,7 @@
ENCODEMB_INVOKE(&rtcd->encodemb, subb)(be, b, 16);
- x->vp8_short_fdct4x4(be->src_diff, be->coeff, 32);
+ x->short_fdct4x4(be->src_diff, be->coeff, 32);
x->quantize_b(be, b);
--- a/vp8/encoder/encodemb.c
+++ b/vp8/encoder/encodemb.c
@@ -16,7 +16,6 @@
#include "quantize.h"
#include "tokenize.h"
#include "vp8/common/invtrans.h"
-#include "dct.h"
#include "vpx_mem/vpx_mem.h"
#include "rdopt.h"
@@ -127,7 +126,7 @@
for (i = 16; i < 24; i += 2)
{- x->vp8_short_fdct8x4(&x->block[i].src_diff[0],
+ x->short_fdct8x4(&x->block[i].src_diff[0],
&x->block[i].coeff[0], 16);
}
}
@@ -139,7 +138,7 @@
for (i = 0; i < 16; i += 2)
{- x->vp8_short_fdct8x4(&x->block[i].src_diff[0],
+ x->short_fdct8x4(&x->block[i].src_diff[0],
&x->block[i].coeff[0], 32);
}
@@ -159,7 +158,7 @@
for (i = 0; i < 16; i += 2)
{- x->vp8_short_fdct8x4(&x->block[i].src_diff[0],
+ x->short_fdct8x4(&x->block[i].src_diff[0],
&x->block[i].coeff[0], 32);
}
@@ -169,7 +168,7 @@
for (i = 16; i < 24; i += 2)
{- x->vp8_short_fdct8x4(&x->block[i].src_diff[0],
+ x->short_fdct8x4(&x->block[i].src_diff[0],
&x->block[i].coeff[0], 16);
}
@@ -187,7 +186,7 @@
for (i = 0; i < 16; i += 2)
{- x->vp8_short_fdct8x4(&x->block[i].src_diff[0],
+ x->short_fdct8x4(&x->block[i].src_diff[0],
&x->block[i].coeff[0], 32);
}
--- a/vp8/encoder/ethreading.c
+++ b/vp8/encoder/ethreading.c
@@ -304,8 +304,8 @@
z->mv_row_max = x->mv_row_max;
*/
- z->vp8_short_fdct4x4 = x->vp8_short_fdct4x4;
- z->vp8_short_fdct8x4 = x->vp8_short_fdct8x4;
+ z->short_fdct4x4 = x->short_fdct4x4;
+ z->short_fdct8x4 = x->short_fdct8x4;
z->short_walsh4x4 = x->short_walsh4x4;
z->quantize_b = x->quantize_b;
z->quantize_b_pair = x->quantize_b_pair;
--- a/vp8/encoder/generic/csystemdependent.c
+++ b/vp8/encoder/generic/csystemdependent.c
@@ -25,12 +25,6 @@
void vp8_cmachine_specific_config(VP8_COMP *cpi)
{#if CONFIG_RUNTIME_CPU_DETECT
- cpi->rtcd.fdct.short4x4 = vp8_short_fdct4x4_c;
- cpi->rtcd.fdct.short8x4 = vp8_short_fdct8x4_c;
- cpi->rtcd.fdct.fast4x4 = vp8_short_fdct4x4_c;
- cpi->rtcd.fdct.fast8x4 = vp8_short_fdct8x4_c;
- cpi->rtcd.fdct.walsh_short4x4 = vp8_short_walsh4x4_c;
-
cpi->rtcd.encodemb.berr = vp8_block_error_c;
cpi->rtcd.encodemb.mberr = vp8_mbblock_error_c;
cpi->rtcd.encodemb.mbuverr = vp8_mbuverror_c;
--- a/vp8/encoder/onyx_if.c
+++ b/vp8/encoder/onyx_if.c
@@ -141,9 +141,6 @@
extern unsigned int inter_b_modes[15];
#endif
-extern void (*vp8_short_fdct4x4)(short *input, short *output, int pitch);
-extern void (*vp8_short_fdct8x4)(short *input, short *output, int pitch);
-
extern const int vp8_bits_per_mb[2][QINDEX_RANGE];
extern const int qrounding_factors[129];
@@ -979,16 +976,17 @@
if (cpi->sf.improved_dct)
{- cpi->mb.vp8_short_fdct8x4 = FDCT_INVOKE(&cpi->rtcd.fdct, short8x4);
- cpi->mb.vp8_short_fdct4x4 = FDCT_INVOKE(&cpi->rtcd.fdct, short4x4);
+ cpi->mb.short_fdct8x4 = vp8_short_fdct8x4;
+ cpi->mb.short_fdct4x4 = vp8_short_fdct4x4;
}
else
{- cpi->mb.vp8_short_fdct8x4 = FDCT_INVOKE(&cpi->rtcd.fdct, fast8x4);
- cpi->mb.vp8_short_fdct4x4 = FDCT_INVOKE(&cpi->rtcd.fdct, fast4x4);
+ /* No fast FDCT defined for any platform at this time. */
+ cpi->mb.short_fdct8x4 = vp8_short_fdct8x4;
+ cpi->mb.short_fdct4x4 = vp8_short_fdct4x4;
}
- cpi->mb.short_walsh4x4 = FDCT_INVOKE(&cpi->rtcd.fdct, walsh_short4x4);
+ cpi->mb.short_walsh4x4 = vp8_short_walsh4x4;
if (cpi->sf.improved_quant)
{--- a/vp8/encoder/onyx_int.h
+++ b/vp8/encoder/onyx_int.h
@@ -19,7 +19,6 @@
#include "tokenize.h"
#include "vp8/common/onyxc_int.h"
#include "variance.h"
-#include "dct.h"
#include "encodemb.h"
#include "quantize.h"
#include "vp8/common/entropy.h"
@@ -224,7 +223,6 @@
typedef struct VP8_ENCODER_RTCD
{- vp8_fdct_rtcd_vtable_t fdct;
vp8_encodemb_rtcd_vtable_t encodemb;
vp8_quantize_rtcd_vtable_t quantize;
vp8_search_rtcd_vtable_t search;
--- a/vp8/encoder/rdopt.c
+++ b/vp8/encoder/rdopt.c
@@ -32,7 +32,6 @@
#include "mcomp.h"
#include "rdopt.h"
#include "vpx_mem/vpx_mem.h"
-#include "dct.h"
#include "vp8/common/systemdependent.h"
#if CONFIG_RUNTIME_CPU_DETECT
@@ -580,7 +579,7 @@
// Fdct and building the 2nd order block
for (beptr = mb->block; beptr < mb->block + 16; beptr += 2)
{- mb->vp8_short_fdct8x4(beptr->src_diff, beptr->coeff, 32);
+ mb->short_fdct8x4(beptr->src_diff, beptr->coeff, 32);
*Y2DCPtr++ = beptr->coeff[0];
*Y2DCPtr++ = beptr->coeff[16];
}
@@ -656,7 +655,7 @@
(*(b->base_dst) + b->dst, b->dst_stride,
mode, b->predictor, 16);
ENCODEMB_INVOKE(IF_RTCD(&cpi->rtcd.encodemb), subb)(be, b, 16);
- x->vp8_short_fdct4x4(be->src_diff, be->coeff, 32);
+ x->short_fdct4x4(be->src_diff, be->coeff, 32);
x->quantize_b(be, b);
tempa = ta;
@@ -1028,7 +1027,7 @@
vp8_build_inter_predictors_b(bd, 16, x->e_mbd.subpixel_predict);
ENCODEMB_INVOKE(rtcd, subb)(be, bd, 16);
- x->vp8_short_fdct4x4(be->src_diff, be->coeff, 32);
+ x->short_fdct4x4(be->src_diff, be->coeff, 32);
// set to 0 no way to account for 2nd order DC so discount
//be->coeff[0] = 0;
--- a/vp8/encoder/x86/dct_x86.h
+++ /dev/null
@@ -1,73 +1,0 @@
-/*
- * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- * Use of this source code is governed by a BSD-style license
- * that can be found in the LICENSE file in the root of the source
- * tree. An additional intellectual property rights grant can be found
- * in the file PATENTS. All contributing project authors may
- * be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef DCT_X86_H
-#define DCT_X86_H
-
-
-/* Note:
- *
- * This platform is commonly built for runtime CPU detection. If you modify
- * any of the function mappings present in this file, be sure to also update
- * them in the function pointer initialization code
- */
-#if HAVE_MMX
-extern prototype_fdct(vp8_short_fdct4x4_mmx);
-extern prototype_fdct(vp8_short_fdct8x4_mmx);
-
-#if !CONFIG_RUNTIME_CPU_DETECT
-
-#undef vp8_fdct_short4x4
-#define vp8_fdct_short4x4 vp8_short_fdct4x4_mmx
-
-#undef vp8_fdct_short8x4
-#define vp8_fdct_short8x4 vp8_short_fdct8x4_mmx
-
-#undef vp8_fdct_fast4x4
-#define vp8_fdct_fast4x4 vp8_short_fdct4x4_mmx
-
-#undef vp8_fdct_fast8x4
-#define vp8_fdct_fast8x4 vp8_short_fdct8x4_mmx
-
-#endif
-
-#endif
-
-
-#if HAVE_SSE2
-extern prototype_fdct(vp8_short_fdct8x4_sse2);
-extern prototype_fdct(vp8_short_walsh4x4_sse2);
-
-extern prototype_fdct(vp8_short_fdct4x4_sse2);
-
-#if !CONFIG_RUNTIME_CPU_DETECT
-
-#undef vp8_fdct_short4x4
-#define vp8_fdct_short4x4 vp8_short_fdct4x4_sse2
-
-#undef vp8_fdct_short8x4
-#define vp8_fdct_short8x4 vp8_short_fdct8x4_sse2
-
-#undef vp8_fdct_fast4x4
-#define vp8_fdct_fast4x4 vp8_short_fdct4x4_sse2
-
-#undef vp8_fdct_fast8x4
-#define vp8_fdct_fast8x4 vp8_short_fdct8x4_sse2
-
-#undef vp8_fdct_walsh_short4x4
-#define vp8_fdct_walsh_short4x4 vp8_short_walsh4x4_sse2
-
-#endif
-
-
-#endif
-
-#endif
--- a/vp8/encoder/x86/x86_csystemdependent.c
+++ b/vp8/encoder/x86/x86_csystemdependent.c
@@ -127,13 +127,6 @@
#if HAVE_MMX
if (flags & HAS_MMX)
{- cpi->rtcd.fdct.short4x4 = vp8_short_fdct4x4_mmx;
- cpi->rtcd.fdct.short8x4 = vp8_short_fdct8x4_mmx;
- cpi->rtcd.fdct.fast4x4 = vp8_short_fdct4x4_mmx;
- cpi->rtcd.fdct.fast8x4 = vp8_short_fdct8x4_mmx;
-
- cpi->rtcd.fdct.walsh_short4x4 = vp8_short_walsh4x4_c;
-
cpi->rtcd.encodemb.berr = vp8_block_error_mmx;
cpi->rtcd.encodemb.mberr = vp8_mbblock_error_mmx;
cpi->rtcd.encodemb.mbuverr = vp8_mbuverror_mmx;
@@ -148,13 +141,6 @@
#if HAVE_SSE2
if (flags & HAS_SSE2)
{- cpi->rtcd.fdct.short4x4 = vp8_short_fdct4x4_sse2;
- cpi->rtcd.fdct.short8x4 = vp8_short_fdct8x4_sse2;
- cpi->rtcd.fdct.fast4x4 = vp8_short_fdct4x4_sse2;
- cpi->rtcd.fdct.fast8x4 = vp8_short_fdct8x4_sse2;
-
- cpi->rtcd.fdct.walsh_short4x4 = vp8_short_walsh4x4_sse2 ;
-
cpi->rtcd.encodemb.berr = vp8_block_error_xmm;
cpi->rtcd.encodemb.mberr = vp8_mbblock_error_xmm;
cpi->rtcd.encodemb.mbuverr = vp8_mbuverror_xmm;
--- a/vp8/vp8cx.mk
+++ b/vp8/vp8cx.mk
@@ -48,7 +48,6 @@
VP8_CX_SRCS-yes += encoder/block.h
VP8_CX_SRCS-yes += encoder/boolhuff.h
VP8_CX_SRCS-yes += encoder/bitstream.h
-VP8_CX_SRCS-yes += encoder/dct.h
VP8_CX_SRCS-yes += encoder/encodeintra.h
VP8_CX_SRCS-yes += encoder/encodemb.h
VP8_CX_SRCS-yes += encoder/encodemv.h
@@ -95,7 +94,6 @@
endif
VP8_CX_SRCS-$(ARCH_X86)$(ARCH_X86_64) += encoder/x86/encodemb_x86.h
-VP8_CX_SRCS-$(ARCH_X86)$(ARCH_X86_64) += encoder/x86/dct_x86.h
VP8_CX_SRCS-$(ARCH_X86)$(ARCH_X86_64) += encoder/x86/mcomp_x86.h
VP8_CX_SRCS-$(ARCH_X86)$(ARCH_X86_64) += encoder/x86/quantize_x86.h
VP8_CX_SRCS-$(ARCH_X86)$(ARCH_X86_64) += encoder/x86/temporal_filter_x86.h
--- a/vp8/vp8cx_arm.mk
+++ b/vp8/vp8cx_arm.mk
@@ -16,7 +16,6 @@
VP8_CX_SRCS-$(ARCH_ARM) += encoder/arm/arm_csystemdependent.c
VP8_CX_SRCS-$(ARCH_ARM) += encoder/arm/dct_arm.c
-VP8_CX_SRCS-$(ARCH_ARM) += encoder/arm/dct_arm.h
VP8_CX_SRCS-$(ARCH_ARM) += encoder/arm/encodemb_arm.h
VP8_CX_SRCS-$(ARCH_ARM) += encoder/arm/quantize_arm.c
VP8_CX_SRCS-$(ARCH_ARM) += encoder/arm/quantize_arm.h
--
⑨