vp8: remove mmx functions

When they have sse2 equivalents.

Change-Id: I158f631a3bcecba57b36093ac10114b1904767a7

3 files changed, 0 insertions, 461 deletions

diff --git a/vp8/encoder/x86/dct_mmx.asm b/vp8/encoder/x86/dct_mmx.asm
deleted file mode 100644
index 6f188cb94..000000000
--- a/vp8/encoder/x86/dct_mmx.asm
+++ /dev/null
@@ -1,241 +0,0 @@
-;
-;  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"
-
-;void vp8_short_fdct4x4_mmx(short *input, short *output, int pitch)
-global sym(vp8_short_fdct4x4_mmx) PRIVATE
-sym(vp8_short_fdct4x4_mmx):
-    push        rbp
-    mov         rbp,        rsp
-    SHADOW_ARGS_TO_STACK 3
-    GET_GOT     rbx
-    push        rsi
-    push        rdi
-    ; end prolog
-
-        mov         rsi,        arg(0)      ; input
-        mov         rdi,        arg(1)      ; output
-
-        movsxd      rax,        dword ptr arg(2) ;pitch
-
-        lea         rcx,        [rsi + rax*2]
-        ; read the input data
-        movq        mm0,        [rsi]
-        movq        mm1,        [rsi + rax]
-
-        movq        mm2,        [rcx]
-        movq        mm4,        [rcx + rax]
-
-        ; transpose for the first stage
-        movq        mm3,        mm0         ; 00 01 02 03
-        movq        mm5,        mm2         ; 20 21 22 23
-
-        punpcklwd   mm0,        mm1         ; 00 10 01 11
-        punpckhwd   mm3,        mm1         ; 02 12 03 13
-
-        punpcklwd   mm2,        mm4         ; 20 30 21 31
-        punpckhwd   mm5,        mm4         ; 22 32 23 33
-
-        movq        mm1,        mm0         ; 00 10 01 11
-        punpckldq   mm0,        mm2         ; 00 10 20 30
-
-        punpckhdq   mm1,        mm2         ; 01 11 21 31
-
-        movq        mm2,        mm3         ; 02 12 03 13
-        punpckldq   mm2,        mm5         ; 02 12 22 32
-
-        punpckhdq   mm3,        mm5         ; 03 13 23 33
-
-        ; mm0 0
-        ; mm1 1
-        ; mm2 2
-        ; mm3 3
-
-        ; first stage
-        movq        mm5,        mm0
-        movq        mm4,        mm1
-
-        paddw       mm0,        mm3         ; a1 = 0 + 3
-        paddw       mm1,        mm2         ; b1 = 1 + 2
-
-        psubw       mm4,        mm2         ; c1 = 1 - 2
-        psubw       mm5,        mm3         ; d1 = 0 - 3
-
-        psllw       mm5,        3
-        psllw       mm4,        3
-
-        psllw       mm0,        3
-        psllw       mm1,        3
-
-        ; output 0 and 2
-        movq        mm2,        mm0         ; a1
-
-        paddw       mm0,        mm1         ; op[0] = a1 + b1
-        psubw       mm2,        mm1         ; op[2] = a1 - b1
-
-        ; output 1 and 3
-        ; interleave c1, d1
-        movq        mm1,        mm5         ; d1
-        punpcklwd   mm1,        mm4         ; c1 d1
-        punpckhwd   mm5,        mm4         ; c1 d1
-
-        movq        mm3,        mm1
-        movq        mm4,        mm5
-
-        pmaddwd     mm1,        MMWORD PTR[GLOBAL (_5352_2217)]    ; c1*2217 + d1*5352
-        pmaddwd     mm4,        MMWORD PTR[GLOBAL (_5352_2217)]    ; c1*2217 + d1*5352
-
-        pmaddwd     mm3,        MMWORD PTR[GLOBAL(_2217_neg5352)]  ; d1*2217 - c1*5352
-        pmaddwd     mm5,        MMWORD PTR[GLOBAL(_2217_neg5352)]  ; d1*2217 - c1*5352
-
-        paddd       mm1,        MMWORD PTR[GLOBAL(_14500)]
-        paddd       mm4,        MMWORD PTR[GLOBAL(_14500)]
-        paddd       mm3,        MMWORD PTR[GLOBAL(_7500)]
-        paddd       mm5,        MMWORD PTR[GLOBAL(_7500)]
-
-        psrad       mm1,        12          ; (c1 * 2217 + d1 * 5352 +  14500)>>12
-        psrad       mm4,        12          ; (c1 * 2217 + d1 * 5352 +  14500)>>12
-        psrad       mm3,        12          ; (d1 * 2217 - c1 * 5352 +   7500)>>12
-        psrad       mm5,        12          ; (d1 * 2217 - c1 * 5352 +   7500)>>12
-
-        packssdw    mm1,        mm4         ; op[1]
-        packssdw    mm3,        mm5         ; op[3]
-
-        ; done with vertical
-        ; transpose for the second stage
-        movq        mm4,        mm0         ; 00 10 20 30
-        movq        mm5,        mm2         ; 02 12 22 32
-
-        punpcklwd   mm0,        mm1         ; 00 01 10 11
-        punpckhwd   mm4,        mm1         ; 20 21 30 31
-
-        punpcklwd   mm2,        mm3         ; 02 03 12 13
-        punpckhwd   mm5,        mm3         ; 22 23 32 33
-
-        movq        mm1,        mm0         ; 00 01 10 11
-        punpckldq   mm0,        mm2         ; 00 01 02 03
-
-        punpckhdq   mm1,        mm2         ; 01 22 12 13
-
-        movq        mm2,        mm4         ; 20 31 30 31
-        punpckldq   mm2,        mm5         ; 20 21 22 23
-
-        punpckhdq   mm4,        mm5         ; 30 31 32 33
-
-        ; mm0 0
-        ; mm1 1
-        ; mm2 2
-        ; mm3 4
-
-        movq        mm5,        mm0
-        movq        mm3,        mm1
-
-        paddw       mm0,        mm4         ; a1 = 0 + 3
-        paddw       mm1,        mm2         ; b1 = 1 + 2
-
-        psubw       mm3,        mm2         ; c1 = 1 - 2
-        psubw       mm5,        mm4         ; d1 = 0 - 3
-
-        pxor        mm6,        mm6         ; zero out for compare
-
-        pcmpeqw     mm6,        mm5         ; d1 != 0
-
-        pandn       mm6,        MMWORD PTR[GLOBAL(_cmp_mask)]   ; clear upper,
-                                                                ; and keep bit 0 of lower
-
-        ; output 0 and 2
-        movq        mm2,        mm0         ; a1
-
-        paddw       mm0,        mm1         ; a1 + b1
-        psubw       mm2,        mm1         ; a1 - b1
-
-        paddw       mm0,        MMWORD PTR[GLOBAL(_7w)]
-        paddw       mm2,        MMWORD PTR[GLOBAL(_7w)]
-
-        psraw       mm0,        4           ; op[0] = (a1 + b1 + 7)>>4
-        psraw       mm2,        4           ; op[8] = (a1 - b1 + 7)>>4
-
-        movq        MMWORD PTR[rdi + 0 ],  mm0
-        movq        MMWORD PTR[rdi + 16],  mm2
-
-        ; output 1 and 3
-        ; interleave c1, d1
-        movq        mm1,        mm5         ; d1
-        punpcklwd   mm1,        mm3         ; c1 d1
-        punpckhwd   mm5,        mm3         ; c1 d1
-
-        movq        mm3,        mm1
-        movq        mm4,        mm5
-
-        pmaddwd     mm1,        MMWORD PTR[GLOBAL (_5352_2217)]    ; c1*2217 + d1*5352
-        pmaddwd     mm4,        MMWORD PTR[GLOBAL (_5352_2217)]    ; c1*2217 + d1*5352
-
-        pmaddwd     mm3,        MMWORD PTR[GLOBAL(_2217_neg5352)]  ; d1*2217 - c1*5352
-        pmaddwd     mm5,        MMWORD PTR[GLOBAL(_2217_neg5352)]  ; d1*2217 - c1*5352
-
-        paddd       mm1,        MMWORD PTR[GLOBAL(_12000)]
-        paddd       mm4,        MMWORD PTR[GLOBAL(_12000)]
-        paddd       mm3,        MMWORD PTR[GLOBAL(_51000)]
-        paddd       mm5,        MMWORD PTR[GLOBAL(_51000)]
-
-        psrad       mm1,        16          ; (c1 * 2217 + d1 * 5352 +  14500)>>16
-        psrad       mm4,        16          ; (c1 * 2217 + d1 * 5352 +  14500)>>16
-        psrad       mm3,        16          ; (d1 * 2217 - c1 * 5352 +   7500)>>16
-        psrad       mm5,        16          ; (d1 * 2217 - c1 * 5352 +   7500)>>16
-
-        packssdw    mm1,        mm4         ; op[4]
-        packssdw    mm3,        mm5         ; op[12]
-
-        paddw       mm1,        mm6         ; op[4] += (d1!=0)
-
-        movq        MMWORD PTR[rdi + 8 ],  mm1
-        movq        MMWORD PTR[rdi + 24],  mm3
-
-     ; begin epilog
-    pop         rdi
-    pop         rsi
-    RESTORE_GOT
-    UNSHADOW_ARGS
-    pop         rbp
-    ret
-
-SECTION_RODATA
-align 8
-_5352_2217:
-    dw 5352
-    dw 2217
-    dw 5352
-    dw 2217
-align 8
-_2217_neg5352:
-    dw 2217
-    dw -5352
-    dw 2217
-    dw -5352
-align 8
-_cmp_mask:
-    times 4 dw 1
-align 8
-_7w:
-    times 4 dw 7
-align 8
-_14500:
-    times 2 dd 14500
-align 8
-_7500:
-    times 2 dd 7500
-align 8
-_12000:
-    times 2 dd 12000
-align 8
-_51000:
-    times 2 dd 51000
diff --git a/vp8/encoder/x86/encodeopt.asm b/vp8/encoder/x86/encodeopt.asm
index 9c8780125..0297220ee 100644
--- a/vp8/encoder/x86/encodeopt.asm
+++ b/vp8/encoder/x86/encodeopt.asm
@@ -59,149 +59,6 @@ sym(vp8_block_error_sse2):
     pop         rbp
     ret
 
-;int vp8_block_error_mmx(short *coeff_ptr,  short *dcoef_ptr)
-global sym(vp8_block_error_mmx) PRIVATE
-sym(vp8_block_error_mmx):
-    push        rbp
-    mov         rbp, rsp
-    SHADOW_ARGS_TO_STACK 2
-    push rsi
-    push rdi
-    ; end prolog
-
-
-        mov         rsi,        arg(0) ;coeff_ptr
-        pxor        mm7,        mm7
-
-        mov         rdi,        arg(1) ;dcoef_ptr
-        movq        mm3,        [rsi]
-
-        movq        mm4,        [rdi]
-        movq        mm5,        [rsi+8]
-
-        movq        mm6,        [rdi+8]
-        pxor        mm1,        mm1 ; from movd mm1, dc ; dc =0
-
-        movq        mm2,        mm7
-        psubw       mm5,        mm6
-
-        por         mm1,        mm2
-        pmaddwd     mm5,        mm5
-
-        pcmpeqw     mm1,        mm7
-        psubw       mm3,        mm4
-
-        pand        mm1,        mm3
-        pmaddwd     mm1,        mm1
-
-        paddd       mm1,        mm5
-        movq        mm3,        [rsi+16]
-
-        movq        mm4,        [rdi+16]
-        movq        mm5,        [rsi+24]
-
-        movq        mm6,        [rdi+24]
-        psubw       mm5,        mm6
-
-        pmaddwd     mm5,        mm5
-        psubw       mm3,        mm4
-
-        pmaddwd     mm3,        mm3
-        paddd       mm3,        mm5
-
-        paddd       mm1,        mm3
-        movq        mm0,        mm1
-
-        psrlq       mm1,        32
-        paddd       mm0,        mm1
-
-        movq        rax,        mm0
-
-    pop rdi
-    pop rsi
-    ; begin epilog
-    UNSHADOW_ARGS
-    pop         rbp
-    ret
-
-
-;int vp8_mbblock_error_mmx_impl(short *coeff_ptr, short *dcoef_ptr, int dc);
-global sym(vp8_mbblock_error_mmx_impl) PRIVATE
-sym(vp8_mbblock_error_mmx_impl):
-    push        rbp
-    mov         rbp, rsp
-    SHADOW_ARGS_TO_STACK 3
-    push rsi
-    push rdi
-    ; end prolog
-
-
-        mov         rsi,        arg(0) ;coeff_ptr
-        pxor        mm7,        mm7
-
-        mov         rdi,        arg(1) ;dcoef_ptr
-        pxor        mm2,        mm2
-
-        movd        mm1,        dword ptr arg(2) ;dc
-        por         mm1,        mm2
-
-        pcmpeqw     mm1,        mm7
-        mov         rcx,        16
-
-.mberror_loop_mmx:
-        movq        mm3,       [rsi]
-        movq        mm4,       [rdi]
-
-        movq        mm5,       [rsi+8]
-        movq        mm6,       [rdi+8]
-
-
-        psubw       mm5,        mm6
-        pmaddwd     mm5,        mm5
-
-        psubw       mm3,        mm4
-        pand        mm3,        mm1
-
-        pmaddwd     mm3,        mm3
-        paddd       mm2,        mm5
-
-        paddd       mm2,        mm3
-        movq        mm3,       [rsi+16]
-
-        movq        mm4,       [rdi+16]
-        movq        mm5,       [rsi+24]
-
-        movq        mm6,       [rdi+24]
-        psubw       mm5,        mm6
-
-        pmaddwd     mm5,        mm5
-        psubw       mm3,        mm4
-
-        pmaddwd     mm3,        mm3
-        paddd       mm2,        mm5
-
-        paddd       mm2,        mm3
-        add         rsi,        32
-
-        add         rdi,        32
-        sub         rcx,        1
-
-        jnz         .mberror_loop_mmx
-
-        movq        mm0,        mm2
-        psrlq       mm2,        32
-
-        paddd       mm0,        mm2
-        movq        rax,        mm0
-
-    pop rdi
-    pop rsi
-    ; begin epilog
-    UNSHADOW_ARGS
-    pop         rbp
-    ret
-
-
 ;int vp8_mbblock_error_sse2_impl(short *coeff_ptr, short *dcoef_ptr, int dc);
 global sym(vp8_mbblock_error_sse2_impl) PRIVATE
 sym(vp8_mbblock_error_sse2_impl):
@@ -272,63 +129,6 @@ sym(vp8_mbblock_error_sse2_impl):
     ret
 
 
-;int vp8_mbuverror_mmx_impl(short *s_ptr, short *d_ptr);
-global sym(vp8_mbuverror_mmx_impl) PRIVATE
-sym(vp8_mbuverror_mmx_impl):
-    push        rbp
-    mov         rbp, rsp
-    SHADOW_ARGS_TO_STACK 2
-    push rsi
-    push rdi
-    ; end prolog
-
-
-        mov             rsi,        arg(0) ;s_ptr
-        mov             rdi,        arg(1) ;d_ptr
-
-        mov             rcx,        16
-        pxor            mm7,        mm7
-
-.mbuverror_loop_mmx:
-
-        movq            mm1,        [rsi]
-        movq            mm2,        [rdi]
-
-        psubw           mm1,        mm2
-        pmaddwd         mm1,        mm1
-
-
-        movq            mm3,        [rsi+8]
-        movq            mm4,        [rdi+8]
-
-        psubw           mm3,        mm4
-        pmaddwd         mm3,        mm3
-
-
-        paddd           mm7,        mm1
-        paddd           mm7,        mm3
-
-
-        add             rsi,        16
-        add             rdi,        16
-
-        dec             rcx
-        jnz             .mbuverror_loop_mmx
-
-        movq            mm0,        mm7
-        psrlq           mm7,        32
-
-        paddd           mm0,        mm7
-        movq            rax,        mm0
-
-    pop rdi
-    pop rsi
-    ; begin epilog
-    UNSHADOW_ARGS
-    pop         rbp
-    ret
-
-
 ;int vp8_mbuverror_sse2_impl(short *s_ptr, short *d_ptr);
 global sym(vp8_mbuverror_sse2_impl) PRIVATE
 sym(vp8_mbuverror_sse2_impl):
diff --git a/vp8/encoder/x86/vp8_enc_stubs_mmx.c b/vp8/encoder/x86/vp8_enc_stubs_mmx.c
index d00abd4d0..4406dd0cc 100644
--- a/vp8/encoder/x86/vp8_enc_stubs_mmx.c
+++ b/vp8/encoder/x86/vp8_enc_stubs_mmx.c
@@ -13,12 +13,6 @@
 #include "vpx_ports/x86.h"
 #include "vp8/encoder/block.h"
 
-void vp8_short_fdct4x4_mmx(short *input, short *output, int pitch);
-void vp8_short_fdct8x4_mmx(short *input, short *output, int pitch) {
-  vp8_short_fdct4x4_mmx(input, output, pitch);
-  vp8_short_fdct4x4_mmx(input + 4, output + 16, pitch);
-}
-
 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,
@@ -38,17 +32,3 @@ void vp8_fast_quantize_b_mmx(BLOCK *b, BLOCKD *d) {
 
       round_ptr, quant_ptr, dqcoeff_ptr);
 }
-
-int vp8_mbblock_error_mmx_impl(short *coeff_ptr, short *dcoef_ptr, int dc);
-int vp8_mbblock_error_mmx(MACROBLOCK *mb, int dc) {
-  short *coeff_ptr = mb->block[0].coeff;
-  short *dcoef_ptr = mb->e_mbd.block[0].dqcoeff;
-  return vp8_mbblock_error_mmx_impl(coeff_ptr, dcoef_ptr, dc);
-}
-
-int vp8_mbuverror_mmx_impl(short *s_ptr, short *d_ptr);
-int vp8_mbuverror_mmx(MACROBLOCK *mb) {
-  short *s_ptr = &mb->coeff[256];
-  short *d_ptr = &mb->e_mbd.dqcoeff[256];
-  return vp8_mbuverror_mmx_impl(s_ptr, d_ptr);
-}