/*
* Copyright (C) 2024 Linux Studio Plugins Project
* (C) 2024 Vladimir Sadovnikov
*
* This file is part of lsp-dsp-lib
* Created on: 31 мар. 2020 г.
*
* lsp-dsp-lib 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 3 of the License, or
* any later version.
*
* lsp-dsp-lib 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 lsp-dsp-lib. If not, see .
*/
#ifndef PRIVATE_DSP_ARCH_X86_AVX_MSMATRIX_H_
#define PRIVATE_DSP_ARCH_X86_AVX_MSMATRIX_H_
#ifndef PRIVATE_DSP_ARCH_X86_AVX_IMPL
#error "This header should not be included directly"
#endif /* PRIVATE_DSP_ARCH_X86_AVX_IMPL */
namespace lsp
{
namespace avx
{
IF_ARCH_X86(
static const float msmatrix_const[] __lsp_aligned32 =
{
LSP_DSP_VEC8(0.5f)
};
)
void lr_to_ms(float *m, float *s, const float *l, const float *r, size_t count)
{
IF_ARCH_X86(size_t off);
ARCH_X86_ASM(
__ASM_EMIT("xor %[off], %[off]")
__ASM_EMIT("vmovaps %[X_HALF], %%ymm7")
// 16x blocks
__ASM_EMIT32("subl $16, %[count]")
__ASM_EMIT64("sub $16, %[count]")
__ASM_EMIT("vmovaps %%ymm7, %%ymm6")
__ASM_EMIT("jb 2f")
__ASM_EMIT("1:")
__ASM_EMIT("vmulps 0x00(%[left], %[off]), %%ymm6, %%ymm0") // ymm0 = l*0.5
__ASM_EMIT("vmulps 0x20(%[left], %[off]), %%ymm7, %%ymm1")
__ASM_EMIT("vmulps 0x00(%[right], %[off]), %%ymm6, %%ymm4") // ymm4 = r*0.5
__ASM_EMIT("vmulps 0x20(%[right], %[off]), %%ymm7, %%ymm5")
__ASM_EMIT("vsubps %%ymm4, %%ymm0, %%ymm2") // ymm2 = s = (l-r)*0.5
__ASM_EMIT("vsubps %%ymm5, %%ymm1, %%ymm3")
__ASM_EMIT("vaddps %%ymm4, %%ymm0, %%ymm0") // ymm0 = m = (l+r)*0.5
__ASM_EMIT("vaddps %%ymm5, %%ymm1, %%ymm1")
__ASM_EMIT("vmovups %%ymm0, 0x00(%[mid], %[off])")
__ASM_EMIT("vmovups %%ymm1, 0x20(%[mid], %[off])")
__ASM_EMIT("vmovups %%ymm2, 0x00(%[side], %[off])")
__ASM_EMIT("vmovups %%ymm3, 0x20(%[side], %[off])")
__ASM_EMIT("add $0x40, %[off]")
__ASM_EMIT32("subl $16, %[count]")
__ASM_EMIT64("sub $16, %[count]")
__ASM_EMIT("jae 1b")
__ASM_EMIT("2:")
// 8x block
__ASM_EMIT32("addl $8, %[count]")
__ASM_EMIT64("add $8, %[count]")
__ASM_EMIT("jl 4f")
__ASM_EMIT("vmulps 0x00(%[left], %[off]), %%ymm6, %%ymm0") // ymm0 = l*0.5
__ASM_EMIT("vmulps 0x00(%[right], %[off]), %%ymm7, %%ymm4") // ymm4 = r*0.5
__ASM_EMIT("vsubps %%ymm4, %%ymm0, %%ymm2") // ymm2 = s = (l-r)*0.5
__ASM_EMIT("vaddps %%ymm4, %%ymm0, %%ymm0") // ymm0 = m = (l+r)*0.5
__ASM_EMIT("vmovups %%ymm0, 0x00(%[mid], %[off])")
__ASM_EMIT("vmovups %%ymm2, 0x00(%[side], %[off])")
__ASM_EMIT("add $0x20, %[off]")
__ASM_EMIT32("subl $8, %[count]")
__ASM_EMIT64("sub $8, %[count]")
__ASM_EMIT("4:")
// 4x block
__ASM_EMIT32("addl $4, %[count]")
__ASM_EMIT64("add $4, %[count]")
__ASM_EMIT("jl 6f")
__ASM_EMIT("vmulps 0x00(%[left], %[off]), %%xmm6, %%xmm0") // xmm0 = l*0.5
__ASM_EMIT("vmulps 0x00(%[right], %[off]), %%xmm7, %%xmm4") // xmm4 = r*0.5
__ASM_EMIT("vsubps %%xmm4, %%xmm0, %%xmm2") // xmm2 = s = (l-r)*0.5
__ASM_EMIT("vaddps %%xmm4, %%xmm0, %%xmm0") // xmm0 = m = (l+r)*0.5
__ASM_EMIT("vmovups %%xmm0, 0x00(%[mid], %[off])")
__ASM_EMIT("vmovups %%xmm2, 0x00(%[side], %[off])")
__ASM_EMIT("add $0x10, %[off]")
__ASM_EMIT32("subl $4, %[count]")
__ASM_EMIT64("sub $4, %[count]")
__ASM_EMIT("6:")
// 1x blocks
__ASM_EMIT32("addl $3, %[count]")
__ASM_EMIT64("add $3, %[count]")
__ASM_EMIT("jl 8f")
__ASM_EMIT("7:")
__ASM_EMIT("vmulss 0x00(%[left], %[off]), %%xmm6, %%xmm0") // xmm0 = l*0.5
__ASM_EMIT("vmulss 0x00(%[right], %[off]), %%xmm7, %%xmm4") // xmm4 = r*0.5
__ASM_EMIT("vsubss %%xmm4, %%xmm0, %%xmm2") // xmm2 = s = (l-r)*0.5
__ASM_EMIT("vaddss %%xmm4, %%xmm0, %%xmm0") // xmm0 = m = (l+r)*0.5
__ASM_EMIT("vmovss %%xmm0, 0x00(%[mid], %[off])")
__ASM_EMIT("vmovss %%xmm2, 0x00(%[side], %[off])")
__ASM_EMIT("add $0x04, %[off]")
__ASM_EMIT32("decl %[count]")
__ASM_EMIT64("dec %[count]")
__ASM_EMIT("jge 7b")
__ASM_EMIT("8:")
: [off] "=&r" (off),
[count] X86_PGREG (count)
: [left] "r"(l), [right] "r" (r),
[mid] "r" (m), [side] "r" (s),
[X_HALF] "m" (msmatrix_const)
: "cc", "memory",
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
"%xmm4", "%xmm5", "%xmm6", "%xmm7"
);
}
#define LR_TO_PART(P, L, R, OP) \
__ASM_EMIT("xor %[off], %[off]") \
__ASM_EMIT("vmovaps %[X_HALF], %%ymm7") \
/* 32x blocks */ \
__ASM_EMIT("sub $32, %[count]") \
__ASM_EMIT("vmovaps %%ymm7, %%ymm6") \
__ASM_EMIT("jb 2f") \
__ASM_EMIT("1:") \
__ASM_EMIT("vmovups 0x00(%[" L "], %[off]), %%ymm0") /* ymm0 = l */ \
__ASM_EMIT("vmovups 0x20(%[" L "], %[off]), %%ymm1") \
__ASM_EMIT("vmovups 0x40(%[" L "], %[off]), %%ymm2") \
__ASM_EMIT("vmovups 0x60(%[" L "], %[off]), %%ymm3") \
__ASM_EMIT(OP "ps 0x00(%[" R "], %[off]), %%ymm0, %%ymm0") /* ymm0 = l op r */ \
__ASM_EMIT(OP "ps 0x20(%[" R "], %[off]), %%ymm1, %%ymm1") \
__ASM_EMIT(OP "ps 0x40(%[" R "], %[off]), %%ymm2, %%ymm2") \
__ASM_EMIT(OP "ps 0x60(%[" R "], %[off]), %%ymm3, %%ymm3") \
__ASM_EMIT("vmulps %%ymm6, %%ymm0, %%ymm0") /* ymm0 = (l op r) * 0.5 */ \
__ASM_EMIT("vmulps %%ymm7, %%ymm1, %%ymm1") \
__ASM_EMIT("vmulps %%ymm6, %%ymm2, %%ymm2") \
__ASM_EMIT("vmulps %%ymm7, %%ymm3, %%ymm3") \
__ASM_EMIT("vmovups %%ymm0, 0x00(%[" P "], %[off])") \
__ASM_EMIT("vmovups %%ymm1, 0x20(%[" P "], %[off])") \
__ASM_EMIT("vmovups %%ymm2, 0x40(%[" P "], %[off])") \
__ASM_EMIT("vmovups %%ymm3, 0x60(%[" P "], %[off])") \
__ASM_EMIT("add $0x80, %[off]") \
__ASM_EMIT("sub $32, %[count]") \
__ASM_EMIT("jae 1b") \
__ASM_EMIT("2:") \
/* 16x block */ \
__ASM_EMIT("add $16, %[count]") \
__ASM_EMIT("jl 4f") \
__ASM_EMIT("vmovups 0x00(%[" L "], %[off]), %%ymm0") /* ymm0 = l */ \
__ASM_EMIT("vmovups 0x20(%[" L "], %[off]), %%ymm1") \
__ASM_EMIT(OP "ps 0x00(%[" R "], %[off]), %%ymm0, %%ymm0") /* ymm0 = l op r */ \
__ASM_EMIT(OP "ps 0x20(%[" R "], %[off]), %%ymm1, %%ymm1") \
__ASM_EMIT("vmulps %%ymm6, %%ymm0, %%ymm0") /* ymm0 = (l op r) * 0.5 */ \
__ASM_EMIT("vmulps %%ymm7, %%ymm1, %%ymm1") \
__ASM_EMIT("vmovups %%ymm0, 0x00(%[" P "], %[off])") \
__ASM_EMIT("vmovups %%ymm1, 0x20(%[" P "], %[off])") \
__ASM_EMIT("sub $16, %[count]") \
__ASM_EMIT("add $0x40, %[off]") \
__ASM_EMIT("4:") \
/* 8x block */ \
__ASM_EMIT("add $8, %[count]") \
__ASM_EMIT("jl 6f") \
__ASM_EMIT("vmovups 0x00(%[" L "], %[off]), %%ymm0") /* ymm0 = l */ \
__ASM_EMIT(OP "ps 0x00(%[" R "], %[off]), %%ymm0, %%ymm0") /* ymm0 = l op r */ \
__ASM_EMIT("vmulps %%ymm6, %%ymm0, %%ymm0") /* ymm0 = (l op r) * 0.5 */ \
__ASM_EMIT("vmovups %%ymm0, 0x00(%[" P "], %[off])") \
__ASM_EMIT("sub $8, %[count]") \
__ASM_EMIT("add $0x20, %[off]") \
__ASM_EMIT("6:") \
/* 4x block */ \
__ASM_EMIT("add $4, %[count]") \
__ASM_EMIT("jl 8f") \
__ASM_EMIT("vmovups 0x00(%[" L "], %[off]), %%xmm0") /* xmm0 = l */ \
__ASM_EMIT(OP "ps 0x00(%[" R "], %[off]), %%xmm0, %%xmm0") /* xmm0 = l op r */ \
__ASM_EMIT("vmulps %%xmm6, %%xmm0, %%xmm0") /* xmm0 = (l op r) * 0.5 */ \
__ASM_EMIT("vmovups %%xmm0, 0x00(%[" P "], %[off])") \
__ASM_EMIT("sub $4, %[count]") \
__ASM_EMIT("add $0x10, %[off]") \
__ASM_EMIT("8:") \
/* 1x blocks */ \
__ASM_EMIT("add $3, %[count]") \
__ASM_EMIT("jl 10f") \
__ASM_EMIT("9:") \
__ASM_EMIT("vmovss 0x00(%[" L "], %[off]), %%xmm0") /* xmm0 = l */ \
__ASM_EMIT(OP "ss 0x00(%[" R "], %[off]), %%xmm0, %%xmm0") /* xmm0 = l op r */ \
__ASM_EMIT("vmulss %%xmm6, %%xmm0, %%xmm0") /* xmm0 = (l op r) * 0.5 */ \
__ASM_EMIT("vmovss %%xmm0, 0x00(%[" P "], %[off])") \
__ASM_EMIT("add $0x04, %[off]") \
__ASM_EMIT("dec %[count]") \
__ASM_EMIT("jge 9b") \
__ASM_EMIT("10:")
void lr_to_mid(float *m, const float *l, const float *r, size_t count)
{
IF_ARCH_X86(size_t off);
ARCH_X86_ASM (
LR_TO_PART("mid", "left", "right", "vadd")
: [off] "=&r" (off), [count] "+r" (count)
: [left] "r" (l), [right] "r" (r),
[mid] "r" (m),
[X_HALF] "m" (msmatrix_const)
: "cc", "memory",
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
"%xmm6", "%xmm7"
);
}
void lr_to_side(float *s, const float *l, const float *r, size_t count)
{
IF_ARCH_X86(size_t off);
ARCH_X86_ASM (
LR_TO_PART("side", "left", "right", "vsub")
: [off] "=&r" (off), [count] "+r" (count)
: [left] "r" (l), [right] "r" (r),
[side] "r" (s),
[X_HALF] "m" (msmatrix_const)
: "cc", "memory",
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
"%xmm6", "%xmm7"
);
}
#undef LR_TO_PART
void ms_to_lr(float *l, float *r, const float *m, const float *s, size_t count)
{
IF_ARCH_X86(size_t off);
ARCH_X86_ASM(
__ASM_EMIT("xor %[off], %[off]")
// 32x blocks
__ASM_EMIT32("subl $32, %[count]")
__ASM_EMIT64("sub $32, %[count]")
__ASM_EMIT("jb 2f")
__ASM_EMIT("1:")
__ASM_EMIT("vmovups 0x00(%[mid], %[off]), %%ymm0") // ymm0 = m
__ASM_EMIT("vmovups 0x20(%[mid], %[off]), %%ymm1")
__ASM_EMIT("vmovups 0x40(%[mid], %[off]), %%ymm2")
__ASM_EMIT("vmovups 0x60(%[mid], %[off]), %%ymm3")
__ASM_EMIT("vsubps 0x00(%[side], %[off]), %%ymm0, %%ymm4") // ymm4 = m - s
__ASM_EMIT("vsubps 0x20(%[side], %[off]), %%ymm1, %%ymm5")
__ASM_EMIT("vsubps 0x40(%[side], %[off]), %%ymm2, %%ymm6")
__ASM_EMIT("vsubps 0x60(%[side], %[off]), %%ymm3, %%ymm7")
__ASM_EMIT("vaddps 0x00(%[side], %[off]), %%ymm0, %%ymm0") // ymm0 = m + s
__ASM_EMIT("vaddps 0x20(%[side], %[off]), %%ymm1, %%ymm1")
__ASM_EMIT("vaddps 0x40(%[side], %[off]), %%ymm2, %%ymm2")
__ASM_EMIT("vaddps 0x60(%[side], %[off]), %%ymm3, %%ymm3")
__ASM_EMIT("vmovups %%ymm0, 0x00(%[left], %[off])")
__ASM_EMIT("vmovups %%ymm1, 0x20(%[left], %[off])")
__ASM_EMIT("vmovups %%ymm2, 0x40(%[left], %[off])")
__ASM_EMIT("vmovups %%ymm3, 0x60(%[left], %[off])")
__ASM_EMIT("vmovups %%ymm4, 0x00(%[right], %[off])")
__ASM_EMIT("vmovups %%ymm5, 0x20(%[right], %[off])")
__ASM_EMIT("vmovups %%ymm6, 0x40(%[right], %[off])")
__ASM_EMIT("vmovups %%ymm7, 0x60(%[right], %[off])")
__ASM_EMIT("add $0x80, %[off]")
__ASM_EMIT32("subl $32, %[count]")
__ASM_EMIT64("sub $32, %[count]")
__ASM_EMIT("jae 1b")
__ASM_EMIT("2:")
// 16x block
__ASM_EMIT32("addl $16, %[count]")
__ASM_EMIT64("add $16, %[count]")
__ASM_EMIT("jl 4f")
__ASM_EMIT("vmovups 0x00(%[mid], %[off]), %%ymm0") // ymm0 = m
__ASM_EMIT("vmovups 0x20(%[mid], %[off]), %%ymm1")
__ASM_EMIT("vsubps 0x00(%[side], %[off]), %%ymm0, %%ymm4") // ymm4 = m - s
__ASM_EMIT("vsubps 0x20(%[side], %[off]), %%ymm1, %%ymm5")
__ASM_EMIT("vaddps 0x00(%[side], %[off]), %%ymm0, %%ymm0") // ymm0 = m + s
__ASM_EMIT("vaddps 0x20(%[side], %[off]), %%ymm1, %%ymm1")
__ASM_EMIT("vmovups %%ymm0, 0x00(%[left], %[off])")
__ASM_EMIT("vmovups %%ymm1, 0x20(%[left], %[off])")
__ASM_EMIT("vmovups %%ymm4, 0x00(%[right], %[off])")
__ASM_EMIT("vmovups %%ymm5, 0x20(%[right], %[off])")
__ASM_EMIT32("subl $16, %[count]")
__ASM_EMIT64("sub $16, %[count]")
__ASM_EMIT("add $0x40, %[off]")
__ASM_EMIT("4:")
// 8x block
__ASM_EMIT32("addl $8, %[count]")
__ASM_EMIT64("add $8, %[count]")
__ASM_EMIT("jl 6f")
__ASM_EMIT("vmovups 0x00(%[mid], %[off]), %%ymm0") // ymm0 = m
__ASM_EMIT("vsubps 0x00(%[side], %[off]), %%ymm0, %%ymm4") // ymm4 = m - s
__ASM_EMIT("vaddps 0x00(%[side], %[off]), %%ymm0, %%ymm0") // ymm0 = m + s
__ASM_EMIT("vmovups %%ymm0, 0x00(%[left], %[off])")
__ASM_EMIT("vmovups %%ymm4, 0x00(%[right], %[off])")
__ASM_EMIT32("subl $8, %[count]")
__ASM_EMIT64("sub $8, %[count]")
__ASM_EMIT("add $0x20, %[off]")
__ASM_EMIT("6:")
// 4x block
__ASM_EMIT32("addl $4, %[count]")
__ASM_EMIT64("add $4, %[count]")
__ASM_EMIT("jl 8f")
__ASM_EMIT("vmovups 0x00(%[mid], %[off]), %%xmm0") // xmm0 = m
__ASM_EMIT("vsubps 0x00(%[side], %[off]), %%xmm0, %%xmm4") // xmm4 = m - s
__ASM_EMIT("vaddps 0x00(%[side], %[off]), %%xmm0, %%xmm0") // xmm0 = m + s
__ASM_EMIT("vmovups %%xmm0, 0x00(%[left], %[off])")
__ASM_EMIT("vmovups %%xmm4, 0x00(%[right], %[off])")
__ASM_EMIT32("subl $4, %[count]")
__ASM_EMIT64("sub $4, %[count]")
__ASM_EMIT("add $0x10, %[off]")
__ASM_EMIT("8:")
// 1x blocks
__ASM_EMIT32("addl $3, %[count]")
__ASM_EMIT64("add $3, %[count]")
__ASM_EMIT("jl 10f")
__ASM_EMIT("9:")
__ASM_EMIT("vmovss 0x00(%[mid], %[off]), %%xmm0") // xmm0 = m
__ASM_EMIT("vsubss 0x00(%[side], %[off]), %%xmm0, %%xmm4") // xmm4 = m - s
__ASM_EMIT("vaddss 0x00(%[side], %[off]), %%xmm0, %%xmm0") // xmm0 = m + s
__ASM_EMIT("vmovss %%xmm0, 0x00(%[left], %[off])")
__ASM_EMIT("vmovss %%xmm4, 0x00(%[right], %[off])")
__ASM_EMIT("add $0x04, %[off]")
__ASM_EMIT32("decl %[count]")
__ASM_EMIT64("dec %[count]")
__ASM_EMIT("jge 9b")
__ASM_EMIT("10:")
: [off] "=&r" (off),
[count] X86_PGREG (count)
: [left] "r"(l), [right] "r" (r),
[mid] "r" (m), [side] "r" (s)
: "cc", "memory",
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
"%xmm4", "%xmm5", "%xmm6", "%xmm7"
);
}
#define MS_TO_PART(P, M, S, OP) \
__ASM_EMIT("xor %[off], %[off]") \
/* 32x blocks */ \
__ASM_EMIT("sub $32, %[count]") \
__ASM_EMIT("jb 2f") \
__ASM_EMIT("1:") \
__ASM_EMIT("vmovups 0x00(%[" M "], %[off]), %%ymm0") /* ymm0 = m */ \
__ASM_EMIT("vmovups 0x20(%[" M "], %[off]), %%ymm1") \
__ASM_EMIT("vmovups 0x40(%[" M "], %[off]), %%ymm2") \
__ASM_EMIT("vmovups 0x60(%[" M "], %[off]), %%ymm3") \
__ASM_EMIT(OP "ps 0x00(%[" S "], %[off]), %%ymm0, %%ymm0") /* ymm4 = m - s */ \
__ASM_EMIT(OP "ps 0x20(%[" S "], %[off]), %%ymm1, %%ymm1") \
__ASM_EMIT(OP "ps 0x40(%[" S "], %[off]), %%ymm2, %%ymm2") \
__ASM_EMIT(OP "ps 0x60(%[" S "], %[off]), %%ymm3, %%ymm3") \
__ASM_EMIT("vmovups %%ymm0, 0x00(%[" P "], %[off])") \
__ASM_EMIT("vmovups %%ymm1, 0x20(%[" P "], %[off])") \
__ASM_EMIT("vmovups %%ymm2, 0x40(%[" P "], %[off])") \
__ASM_EMIT("vmovups %%ymm3, 0x60(%[" P "], %[off])") \
__ASM_EMIT("add $0x80, %[off]") \
__ASM_EMIT("sub $32, %[count]") \
__ASM_EMIT("jae 1b") \
__ASM_EMIT("2:") \
/* 16x block */ \
__ASM_EMIT("add $16, %[count]") \
__ASM_EMIT("jl 4f") \
__ASM_EMIT("vmovups 0x00(%[" M "], %[off]), %%ymm0") /* ymm0 = m */ \
__ASM_EMIT("vmovups 0x20(%[" M "], %[off]), %%ymm1") \
__ASM_EMIT(OP "ps 0x00(%[" S "], %[off]), %%ymm0, %%ymm0") /* ymm4 = m - s */ \
__ASM_EMIT(OP "ps 0x20(%[" S "], %[off]), %%ymm1, %%ymm1") \
__ASM_EMIT("vmovups %%ymm0, 0x00(%[" P "], %[off])") \
__ASM_EMIT("vmovups %%ymm1, 0x20(%[" P "], %[off])") \
__ASM_EMIT("sub $16, %[count]") \
__ASM_EMIT("add $0x40, %[off]") \
__ASM_EMIT("4:") \
/* 8x block */ \
__ASM_EMIT("add $8, %[count]") \
__ASM_EMIT("jl 6f") \
__ASM_EMIT("vmovups 0x00(%[" M "], %[off]), %%ymm0") /* ymm0 = m */ \
__ASM_EMIT(OP "ps 0x00(%[" S "], %[off]), %%ymm0, %%ymm0") /* ymm4 = m - s */ \
__ASM_EMIT("vmovups %%ymm0, 0x00(%[" P "], %[off])") \
__ASM_EMIT("sub $8, %[count]") \
__ASM_EMIT("add $0x20, %[off]") \
__ASM_EMIT("6:") \
/* 4x block */ \
__ASM_EMIT("add $4, %[count]") \
__ASM_EMIT("jl 8f") \
__ASM_EMIT("vmovups 0x00(%[" M "], %[off]), %%xmm0") /* xmm0 = m */ \
__ASM_EMIT(OP "ps 0x00(%[" S "], %[off]), %%xmm0, %%xmm0") /* xmm4 = m - s */ \
__ASM_EMIT("vmovups %%xmm0, 0x00(%[" P "], %[off])") \
__ASM_EMIT("sub $4, %[count]") \
__ASM_EMIT("add $0x10, %[off]") \
__ASM_EMIT("8:") \
/* 1x blocks */ \
__ASM_EMIT("add $3, %[count]") \
__ASM_EMIT("jl 10f") \
__ASM_EMIT("9:") \
__ASM_EMIT("vmovss 0x00(%[" M "], %[off]), %%xmm0") /* xmm0 = m */ \
__ASM_EMIT(OP "ss 0x00(%[" S "], %[off]), %%xmm0, %%xmm0") /* xmm4 = m - s */ \
__ASM_EMIT("vmovss %%xmm0, 0x00(%[" P "], %[off])") \
__ASM_EMIT("add $0x04, %[off]") \
__ASM_EMIT("dec %[count]") \
__ASM_EMIT("jge 9b") \
__ASM_EMIT("10:")
void ms_to_left(float *l, const float *m, const float *s, size_t count)
{
size_t off;
ARCH_X86_ASM (
MS_TO_PART("left", "mid", "side", "vadd")
: [off] "=&r" (off), [count] "+r" (count)
: [left] "r" (l),
[mid] "r" (m), [side] "r" (s)
: "cc", "memory",
"%xmm0", "%xmm1", "%xmm2", "%xmm3"
);
}
void ms_to_right(float *r, const float *m, const float *s, size_t count)
{
size_t off;
ARCH_X86_ASM (
MS_TO_PART("right", "mid", "side", "vsub")
: [off] "=&r" (off), [count] "+r" (count)
: [right] "r" (r),
[mid] "r" (m), [side] "r" (s)
: "cc", "memory",
"%xmm0", "%xmm1", "%xmm2", "%xmm3"
);
}
#undef MS_TO_PART
} /* namespace avx */
} /* namespace lsp */
#endif /* PRIVATE_DSP_ARCH_X86_AVX_MSMATRIX_H_ */