ref: 1d86383512f8cf1551fc81048164885f1ef6f089
parent: 18335f193d4931beb590985881bbf0e02f241f23
author: James Zern <[email protected]>
date: Thu Jun 15 21:33:57 EDT 2017
vp8: remove vp8_fast_quantize_b_mmx and vp8_fast_quantize_b_impl_mmx; this was never enabled in rtcd an sse2 version exists so there isn't much reason to keep a mmx implementation around. Change-Id: I8b3ee7f46ba194ffa0d0a6225a0f299f2a4dea90
--- a/vp8/encoder/x86/quantize_mmx.asm
+++ /dev/null
@@ -1,286 +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.
-;
-
-
-%include "vpx_ports/x86_abi_support.asm"
-
-;int vp8_fast_quantize_b_impl_mmx(short *coeff_ptr, short *zbin_ptr,
-; short *qcoeff_ptr,short *dequant_ptr,
-; short *scan_mask, short *round_ptr,
-; short *quant_ptr, short *dqcoeff_ptr);
-global sym(vp8_fast_quantize_b_impl_mmx) PRIVATE
-sym(vp8_fast_quantize_b_impl_mmx):
- push rbp
- mov rbp, rsp
- SHADOW_ARGS_TO_STACK 8
- push rsi
- push rdi
- ; end prolog
-
-
- mov rsi, arg(0) ;coeff_ptr
- movq mm0, [rsi]
-
- mov rax, arg(1) ;zbin_ptr
- movq mm1, [rax]
-
- movq mm3, mm0
- psraw mm0, 15
-
- pxor mm3, mm0
- psubw mm3, mm0 ; abs
-
- movq mm2, mm3
- pcmpgtw mm1, mm2
-
- pandn mm1, mm2
- movq mm3, mm1
-
- mov rdx, arg(6) ;quant_ptr
- movq mm1, [rdx]
-
- mov rcx, arg(5) ;round_ptr
- movq mm2, [rcx]
-
- paddw mm3, mm2
- pmulhuw mm3, mm1
-
- pxor mm3, mm0
- psubw mm3, mm0 ;gain the sign back
-
- mov rdi, arg(2) ;qcoeff_ptr
- movq mm0, mm3
-
- movq [rdi], mm3
-
- mov rax, arg(3) ;dequant_ptr
- movq mm2, [rax]
-
- pmullw mm3, mm2
- mov rax, arg(7) ;dqcoeff_ptr
-
- movq [rax], mm3
-
- ; next 8
- movq mm4, [rsi+8]
-
- mov rax, arg(1) ;zbin_ptr
- movq mm5, [rax+8]
-
- movq mm7, mm4
- psraw mm4, 15
-
- pxor mm7, mm4
- psubw mm7, mm4 ; abs
-
- movq mm6, mm7
- pcmpgtw mm5, mm6
-
- pandn mm5, mm6
- movq mm7, mm5
-
- movq mm5, [rdx+8]
- movq mm6, [rcx+8]
-
- paddw mm7, mm6
- pmulhuw mm7, mm5
-
- pxor mm7, mm4
- psubw mm7, mm4;gain the sign back
-
- mov rdi, arg(2) ;qcoeff_ptr
-
- movq mm1, mm7
- movq [rdi+8], mm7
-
- mov rax, arg(3) ;dequant_ptr
- movq mm6, [rax+8]
-
- pmullw mm7, mm6
- mov rax, arg(7) ;dqcoeff_ptr
-
- movq [rax+8], mm7
-
-
- ; next 8
- movq mm4, [rsi+16]
-
- mov rax, arg(1) ;zbin_ptr
- movq mm5, [rax+16]
-
- movq mm7, mm4
- psraw mm4, 15
-
- pxor mm7, mm4
- psubw mm7, mm4 ; abs
-
- movq mm6, mm7
- pcmpgtw mm5, mm6
-
- pandn mm5, mm6
- movq mm7, mm5
-
- movq mm5, [rdx+16]
- movq mm6, [rcx+16]
-
- paddw mm7, mm6
- pmulhuw mm7, mm5
-
- pxor mm7, mm4
- psubw mm7, mm4;gain the sign back
-
- mov rdi, arg(2) ;qcoeff_ptr
-
- movq mm1, mm7
- movq [rdi+16], mm7
-
- mov rax, arg(3) ;dequant_ptr
- movq mm6, [rax+16]
-
- pmullw mm7, mm6
- mov rax, arg(7) ;dqcoeff_ptr
-
- movq [rax+16], mm7
-
-
- ; next 8
- movq mm4, [rsi+24]
-
- mov rax, arg(1) ;zbin_ptr
- movq mm5, [rax+24]
-
- movq mm7, mm4
- psraw mm4, 15
-
- pxor mm7, mm4
- psubw mm7, mm4 ; abs
-
- movq mm6, mm7
- pcmpgtw mm5, mm6
-
- pandn mm5, mm6
- movq mm7, mm5
-
- movq mm5, [rdx+24]
- movq mm6, [rcx+24]
-
- paddw mm7, mm6
- pmulhuw mm7, mm5
-
- pxor mm7, mm4
- psubw mm7, mm4;gain the sign back
-
- mov rdi, arg(2) ;qcoeff_ptr
-
- movq mm1, mm7
- movq [rdi+24], mm7
-
- mov rax, arg(3) ;dequant_ptr
- movq mm6, [rax+24]
-
- pmullw mm7, mm6
- mov rax, arg(7) ;dqcoeff_ptr
-
- movq [rax+24], mm7
-
-
-
- mov rdi, arg(4) ;scan_mask
- mov rsi, arg(2) ;qcoeff_ptr
-
- pxor mm5, mm5
- pxor mm7, mm7
-
- movq mm0, [rsi]
- movq mm1, [rsi+8]
-
- movq mm2, [rdi]
- movq mm3, [rdi+8];
-
- pcmpeqw mm0, mm7
- pcmpeqw mm1, mm7
-
- pcmpeqw mm6, mm6
- pxor mm0, mm6
-
- pxor mm1, mm6
- psrlw mm0, 15
-
- psrlw mm1, 15
- pmaddwd mm0, mm2
-
- pmaddwd mm1, mm3
- movq mm5, mm0
-
- paddd mm5, mm1
-
- movq mm0, [rsi+16]
- movq mm1, [rsi+24]
-
- movq mm2, [rdi+16]
- movq mm3, [rdi+24];
-
- pcmpeqw mm0, mm7
- pcmpeqw mm1, mm7
-
- pcmpeqw mm6, mm6
- pxor mm0, mm6
-
- pxor mm1, mm6
- psrlw mm0, 15
-
- psrlw mm1, 15
- pmaddwd mm0, mm2
-
- pmaddwd mm1, mm3
- paddd mm5, mm0
-
- paddd mm5, mm1
- movq mm0, mm5
-
- psrlq mm5, 32
- paddd mm0, mm5
-
- ; eob adjustment begins here
- movq rcx, mm0
- and rcx, 0xffff
-
- xor rdx, rdx
- sub rdx, rcx ; rdx=-rcx
-
- bsr rax, rcx
- inc rax
-
- sar rdx, 31
- and rax, rdx
- ; Substitute the sse assembly for the old mmx mixed assembly/C. The
- ; following is kept as reference
- ; movq rcx, mm0
- ; bsr rax, rcx
- ;
- ; mov eob, rax
- ; mov eee, rcx
- ;
- ;if(eee==0)
- ;{- ; eob=-1;
- ;}
- ;else if(eee<0)
- ;{- ; eob=15;
- ;}
- ;d->eob = eob+1;
-
- ; begin epilog
- pop rdi
- pop rsi
- UNSHADOW_ARGS
- pop rbp
- ret
--- a/vp8/encoder/x86/vp8_enc_stubs_mmx.c
+++ /dev/null
@@ -1,34 +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.
- */
-
-#include "vpx_config.h"
-#include "vp8_rtcd.h"
-#include "vpx_ports/x86.h"
-#include "vp8/encoder/block.h"
-
-int vp8_fast_quantize_b_impl_mmx(short *coeff_ptr, short *zbin_ptr,
- short *qcoeff_ptr, short *dequant_ptr,
- const short *scan_mask, short *round_ptr,
- short *quant_ptr, short *dqcoeff_ptr);
-void vp8_fast_quantize_b_mmx(BLOCK *b, BLOCKD *d) {- const short *scan_mask = vp8_default_zig_zag_mask;
- short *coeff_ptr = b->coeff;
- short *zbin_ptr = b->zbin;
- short *round_ptr = b->round;
- short *quant_ptr = b->quant_fast;
- short *qcoeff_ptr = d->qcoeff;
- short *dqcoeff_ptr = d->dqcoeff;
- short *dequant_ptr = d->dequant;
-
- *d->eob = (char)vp8_fast_quantize_b_impl_mmx(
- coeff_ptr, zbin_ptr, qcoeff_ptr, dequant_ptr, scan_mask,
-
- round_ptr, quant_ptr, dqcoeff_ptr);
-}
--- a/vp8/vp8cx.mk
+++ b/vp8/vp8cx.mk
@@ -78,7 +78,6 @@
VP8_CX_SRCS_REMOVE-yes += encoder/temporal_filter.c
endif
-VP8_CX_SRCS-$(HAVE_MMX) += encoder/x86/vp8_enc_stubs_mmx.c
VP8_CX_SRCS-$(HAVE_SSE2) += encoder/x86/dct_sse2.asm
VP8_CX_SRCS-$(HAVE_SSE2) += encoder/x86/fwalsh_sse2.asm
VP8_CX_SRCS-$(HAVE_SSE2) += encoder/x86/vp8_quantize_sse2.c
@@ -91,7 +90,6 @@
VP8_CX_SRCS-$(HAVE_SSE2) += encoder/x86/temporal_filter_apply_sse2.asm
VP8_CX_SRCS-$(HAVE_SSE2) += encoder/x86/vp8_enc_stubs_sse2.c
-VP8_CX_SRCS-$(ARCH_X86)$(ARCH_X86_64) += encoder/x86/quantize_mmx.asm
VP8_CX_SRCS-$(ARCH_X86)$(ARCH_X86_64) += encoder/x86/encodeopt.asm
ifeq ($(CONFIG_REALTIME_ONLY),yes)