/*
* Copyright (C) 2020 Linux Studio Plugins Project
* (C) 2020 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_AARCH64_ASIMD_FFT_PSCRAMBLE_H_
#define PRIVATE_DSP_ARCH_AARCH64_ASIMD_FFT_PSCRAMBLE_H_
#ifndef PRIVATE_DSP_ARCH_AARCH64_ASIMD_IMPL
#error "This header should not be included directly"
#endif /* PRIVATE_DSP_ARCH_AARCH64_ASIMD_IMPL */
namespace lsp
{
namespace asimd
{
void packed_scramble_self_direct(float *dst, size_t rank)
{
IF_ARCH_AARCH64(
size_t count = 1 << rank;
size_t i, j, rrank = 64 - rank;
float *dp, *src = dst;
);
// Self algorithm
ARCH_AARCH64_ASM(
// Do bit-reverse shuffle
__ASM_EMIT("mov %[i], #1") // i = 1
__ASM_EMIT("1:")
__ASM_EMIT("rbit %[j], %[i]") // j = reverse_bits(i)
__ASM_EMIT("add %[src], %[src], #8")
__ASM_EMIT("lsr %[j], %[j], %[rrank]") // j = reverse_bits(i) >> rank
__ASM_EMIT("cmp %[i], %[j]") // i <=> j
__ASM_EMIT("b.hs 2f") // if (i >= j) continue
__ASM_EMIT("add %[dp], %[dst], %[j], LSL #3") // d_re = &dst[j]
__ASM_EMIT("ldr d0, [%[src]]") // d0 = *src
__ASM_EMIT("ldr d1, [%[dp]]") // d1 = *dp
__ASM_EMIT("str d0, [%[dp]]") // *dp = d0
__ASM_EMIT("str d1, [%[src]]") // *src = d1
__ASM_EMIT("2:")
__ASM_EMIT("add %[i], %[i], #1") // i++
__ASM_EMIT("cmp %[i], %[count]") // i <=> count
__ASM_EMIT("b.lo 1b")
// Perform x8 butterflies
__ASM_EMIT("3:")
__ASM_EMIT("ld4 {v0.4s-v3.4s}, [%[dst]]")
// v0 = r0 r2 r4 r6
// v1 = i0 i2 i4 i6
// v2 = r1 r3 r5 r7
// v3 = i1 i3 i5 i7
__ASM_EMIT("fadd v6.4s, v0.4s, v2.4s") // v6 = r0+r1 r2+r3 r4+r5 r6+r7 = r0' r2' r4' r6'
__ASM_EMIT("fadd v7.4s, v1.4s, v3.4s") // v7 = i0+i1 i2+i3 i4+i5 i6+i7 = i0' i2' i4' i6'
__ASM_EMIT("fsub v0.4s, v0.4s, v2.4s") // v0 = r0-r1 r2-r3 r4-r5 r6-r7 = r1' r3' r5' r7'
__ASM_EMIT("fsub v1.4s, v1.4s, v3.4s") // v1 = i0-i1 i2-i3 i4-i5 i6-i7 = i1' i3' i5' i7'
// v6 = r0' r2' r4' r6'
// v0 = r1' r3' r5' r7'
// v7 = i0' i2' i4' i6'
// v1 = i1' i3' i5' i7'
__ASM_EMIT("uzp1 v4.4s, v6.4s, v0.4s") // v4 = r0' r4' r1' r5'
__ASM_EMIT("uzp2 v0.4s, v6.4s, v0.4s") // v0 = r2' r6' r3' r7'
__ASM_EMIT("uzp1 v5.4s, v7.4s, v1.4s") // v5 = i0' i4' i1' i5'
__ASM_EMIT("uzp2 v1.4s, v7.4s, v1.4s") // v1 = i2' i6' i3' i7'
__ASM_EMIT("mov v2.16b, v0.16b") // v2 = r2' r6' r3' r7'
__ASM_EMIT("ins v0.d[1], v1.d[1]") // v0 = r2' r6' i3' i7'
__ASM_EMIT("ins v1.d[1], v2.d[1]") // v1 = i2' i6' r3' r7'
__ASM_EMIT("fadd v2.4s, v4.4s, v0.4s") // v2 = r0'+r2' r4'+r6' r1'+i3' r5'+i7' = r0" r4" r1" r5"
__ASM_EMIT("fsub v3.4s, v4.4s, v0.4s") // v3 = r0'-r2' r4'-r6' r1'-i3' r5'-i7' = r2" r6" r3" r7"
__ASM_EMIT("fadd v0.4s, v5.4s, v1.4s") // v0 = i0'+i2' i4'+i6' i1'+r3' i5'+r7' = i0" i4" i3" i7"
__ASM_EMIT("fsub v1.4s, v5.4s, v1.4s") // v1 = i0'-i2' i4'-i6' i1'-r3' i5'-r7' = i2" i6" i1" i5"
// v0 = i0" i4" i3" i7"
// v1 = i2" i6" i1" i5"
// v2 = r0" r4" r1" r5"
// v3 = r2" r6" r3" r7"
__ASM_EMIT("mov v4.16b, v0.16b") // v4 = i0" i4" i3" i7"
__ASM_EMIT("ins v0.d[1], v1.d[1]") // v0 = i0" i4" i1" i5"
__ASM_EMIT("ins v1.d[1], v4.d[1]") // v1 = i2" i6" i3" i7"
__ASM_EMIT("uzp1 v4.4s, v2.4s, v3.4s") // v4 = r0" r1" r2" r3"
__ASM_EMIT("uzp2 v5.4s, v2.4s, v3.4s") // v5 = r4" r5" r6" r7"
__ASM_EMIT("uzp1 v6.4s, v0.4s, v1.4s") // v6 = i0" i1" i2" i3"
__ASM_EMIT("uzp2 v7.4s, v0.4s, v1.4s") // v7 = i4" i5" i6" i7"
// Store
__ASM_EMIT("stp q4, q6, [%[dst], #0x00]")
__ASM_EMIT("stp q5, q7, [%[dst], #0x20]")
__ASM_EMIT("subs %[count], %[count], #8") // i <=> count
__ASM_EMIT("add %[dst], %[dst], #0x40")
__ASM_EMIT("b.ne 3b")
: [src] "+r" (src), [dst] "+r" (dst), [dp] "=&r" (dp),
[rrank] "+r" (rrank), [i] "=&r" (i), [j] "=&r" (j),
[count] "+r" (count)
:
: "cc", "memory",
"v0", "v1", "v2", "v3",
"v4", "v5", "v6", "v7"
);
}
void packed_scramble_copy_direct(float *dst, const float *src, size_t rank)
{
IF_ARCH_AARCH64(
size_t i, j;
float *sr1, *sr2;
size_t regs = 1 << (rank - 3), rrank = 64 - (rank - 3);
);
ARCH_AARCH64_ASM(
__ASM_EMIT("eor %[i], %[i], %[i]") // i = 0
__ASM_EMIT("1:")
__ASM_EMIT("rbit %[j], %[i]") // j = reverse_bits(i)
__ASM_EMIT("lsr %[j], %[j], %[rrank]") // j = reverse_bits(i) >> rank
__ASM_EMIT("add %[sr1], %[src], %[j], LSL #3") // sr1 = &src[j]
__ASM_EMIT("add %[sr2], %[sr1], %[regs], LSL #3") // sr2 = &src[j + regs]
__ASM_EMIT("ld2 {v0.s, v1.s}[0], [%[sr1]]") // v0 = r0 ? ? ?, v1 = i0 ? ? ?
__ASM_EMIT("ld2 {v0.s, v1.s}[2], [%[sr2]]") // v0 = r0 ? r4 ?, v1 = i0 ? i4 ?
__ASM_EMIT("add %[sr1], %[sr1], %[regs], LSL #4") // sr1 = &src[j + regs*2]
__ASM_EMIT("add %[sr2], %[sr2], %[regs], LSL #4") // sr2 = &src[j + regs*3]
__ASM_EMIT("ld2 {v0.s, v1.s}[1], [%[sr1]]") // v0 = r0 r2 r4 ?, v1 = i0 i2 i4 ?
__ASM_EMIT("ld2 {v0.s, v1.s}[3], [%[sr2]]") // v0 = r0 r2 r4 r6, v1 = i0 i2 i4 i6
__ASM_EMIT("add %[sr1], %[sr1], %[regs], LSL #4") // sr1 = &src[j + regs*4]
__ASM_EMIT("add %[sr2], %[sr2], %[regs], LSL #4") // sr2 = &src[j + regs*5]
__ASM_EMIT("ld2 {v2.s, v3.s}[0], [%[sr1]]") // v2 = r1 ? ? ?, v3 = i1 ? ? ?
__ASM_EMIT("ld2 {v2.s, v3.s}[2], [%[sr2]]") // v2 = r1 ? r5 ?, v3 = i1 ? i5 ?
__ASM_EMIT("add %[sr1], %[sr1], %[regs], LSL #4") // sr1 = &src[j + regs*6]
__ASM_EMIT("add %[sr2], %[sr2], %[regs], LSL #4") // sr2 = &src[j + regs*7]
__ASM_EMIT("ld2 {v2.s, v3.s}[1], [%[sr1]]") // v2 = r1 r3 r5 ?, v3 = i1 i3 i5 ?
__ASM_EMIT("ld2 {v2.s, v3.s}[3], [%[sr2]]") // v2 = r1 r3 r5 r7, v3 = i1 i3 i5 i7
// v0 = r0 r2 r4 r6
// v1 = i0 i2 i4 i6
// v2 = r1 r3 r5 r7
// v3 = i1 i3 i5 i7
__ASM_EMIT("fadd v6.4s, v0.4s, v2.4s") // v6 = r0+r1 r2+r3 r4+r5 r6+r7 = r0' r2' r4' r6'
__ASM_EMIT("fadd v7.4s, v1.4s, v3.4s") // v7 = i0+i1 i2+i3 i4+i5 i6+i7 = i0' i2' i4' i6'
__ASM_EMIT("fsub v0.4s, v0.4s, v2.4s") // v0 = r0-r1 r2-r3 r4-r5 r6-r7 = r1' r3' r5' r7'
__ASM_EMIT("fsub v1.4s, v1.4s, v3.4s") // v1 = i0-i1 i2-i3 i4-i5 i6-i7 = i1' i3' i5' i7'
// v6 = r0' r2' r4' r6'
// v0 = r1' r3' r5' r7'
// v7 = i0' i2' i4' i6'
// v1 = i1' i3' i5' i7'
__ASM_EMIT("uzp1 v4.4s, v6.4s, v0.4s") // v4 = r0' r4' r1' r5'
__ASM_EMIT("uzp2 v0.4s, v6.4s, v0.4s") // v0 = r2' r6' r3' r7'
__ASM_EMIT("uzp1 v5.4s, v7.4s, v1.4s") // v5 = i0' i4' i1' i5'
__ASM_EMIT("uzp2 v1.4s, v7.4s, v1.4s") // v1 = i2' i6' i3' i7'
__ASM_EMIT("mov v2.16b, v0.16b") // v2 = r2' r6' r3' r7'
__ASM_EMIT("ins v0.d[1], v1.d[1]") // v0 = r2' r6' i3' i7'
__ASM_EMIT("ins v1.d[1], v2.d[1]") // v1 = i2' i6' r3' r7'
__ASM_EMIT("fadd v2.4s, v4.4s, v0.4s") // v2 = r0'+r2' r4'+r6' r1'+i3' r5'+i7' = r0" r4" r1" r5"
__ASM_EMIT("fsub v3.4s, v4.4s, v0.4s") // v3 = r0'-r2' r4'-r6' r1'-i3' r5'-i7' = r2" r6" r3" r7"
__ASM_EMIT("fadd v0.4s, v5.4s, v1.4s") // v0 = i0'+i2' i4'+i6' i1'+r3' i5'+r7' = i0" i4" i3" i7"
__ASM_EMIT("fsub v1.4s, v5.4s, v1.4s") // v1 = i0'-i2' i4'-i6' i1'-r3' i5'-r7' = i2" i6" i1" i5"
// v0 = i0" i4" i3" i7"
// v1 = i2" i6" i1" i5"
// v2 = r0" r4" r1" r5"
// v3 = r2" r6" r3" r7"
__ASM_EMIT("mov v4.16b, v0.16b") // v4 = i0" i4" i3" i7"
__ASM_EMIT("ins v0.d[1], v1.d[1]") // v0 = i0" i4" i1" i5"
__ASM_EMIT("ins v1.d[1], v4.d[1]") // v1 = i2" i6" i3" i7"
__ASM_EMIT("uzp1 v4.4s, v2.4s, v3.4s") // v4 = r0" r1" r2" r3"
__ASM_EMIT("uzp2 v5.4s, v2.4s, v3.4s") // v5 = r4" r5" r6" r7"
__ASM_EMIT("uzp1 v6.4s, v0.4s, v1.4s") // v6 = i0" i1" i2" i3"
__ASM_EMIT("uzp2 v7.4s, v0.4s, v1.4s") // v7 = i4" i5" i6" i7"
// Store
__ASM_EMIT("add %[i], %[i], #1")
__ASM_EMIT("stp q4, q6, [%[dst], #0x00]")
__ASM_EMIT("stp q5, q7, [%[dst], #0x20]")
__ASM_EMIT("cmp %[i], %[regs]")
__ASM_EMIT("add %[dst], %[dst], #0x40")
__ASM_EMIT("b.lo 1b")
: [dst] "+r" (dst), [sr1] "=&r" (sr1), [sr2] "=&r" (sr2),
[rrank] "+r" (rrank), [i] "=&r" (i), [j] "=&r" (j)
: [src] "r" (src), [regs] "r" (regs)
: "cc", "memory",
"v0", "v1", "v2", "v3",
"v4", "v5", "v6", "v7"
);
}
void packed_scramble_self_reverse(float *dst, size_t rank)
{
IF_ARCH_AARCH64(
size_t count = 1 << rank;
size_t i, j, rrank = 64 - rank;
float *dp, *src = dst;
);
// Self algorithm
ARCH_AARCH64_ASM(
// Do bit-reverse shuffle
__ASM_EMIT("mov %[i], #1") // i = 1
__ASM_EMIT("1:")
__ASM_EMIT("rbit %[j], %[i]") // j = reverse_bits(i)
__ASM_EMIT("add %[src], %[src], #8")
__ASM_EMIT("lsr %[j], %[j], %[rrank]") // j = reverse_bits(i) >> rank
__ASM_EMIT("cmp %[i], %[j]") // i <=> j
__ASM_EMIT("b.hs 2f") // if (i >= j) continue
__ASM_EMIT("add %[dp], %[dst], %[j], LSL #3") // d_re = &dst[j]
__ASM_EMIT("ldr d0, [%[src]]") // d0 = *src
__ASM_EMIT("ldr d1, [%[dp]]") // d1 = *dp
__ASM_EMIT("str d0, [%[dp]]") // *dp = d0
__ASM_EMIT("str d1, [%[src]]") // *src = d1
__ASM_EMIT("2:")
__ASM_EMIT("add %[i], %[i], #1") // i++
__ASM_EMIT("cmp %[i], %[count]") // i <=> count
__ASM_EMIT("b.lo 1b")
// Perform x8 butterflies
__ASM_EMIT("3:")
__ASM_EMIT("ld4 {v0.4s-v3.4s}, [%[dst]]")
// v0 = r0 r2 r4 r6
// v1 = i0 i2 i4 i6
// v2 = r1 r3 r5 r7
// v3 = i1 i3 i5 i7
__ASM_EMIT("fadd v6.4s, v0.4s, v2.4s") // v6 = r0+r1 r2+r3 r4+r5 r6+r7 = r0' r2' r4' r6'
__ASM_EMIT("fadd v7.4s, v1.4s, v3.4s") // v7 = i0+i1 i2+i3 i4+i5 i6+i7 = i0' i2' i4' i6'
__ASM_EMIT("fsub v0.4s, v0.4s, v2.4s") // v0 = r0-r1 r2-r3 r4-r5 r6-r7 = r1' r3' r5' r7'
__ASM_EMIT("fsub v1.4s, v1.4s, v3.4s") // v1 = i0-i1 i2-i3 i4-i5 i6-i7 = i1' i3' i5' i7'
// v6 = r0' r2' r4' r6'
// v0 = r1' r3' r5' r7'
// v7 = i0' i2' i4' i6'
// v1 = i1' i3' i5' i7'
__ASM_EMIT("uzp1 v4.4s, v6.4s, v0.4s") // v4 = r0' r4' r1' r5'
__ASM_EMIT("uzp2 v0.4s, v6.4s, v0.4s") // v0 = r2' r6' r3' r7'
__ASM_EMIT("uzp1 v5.4s, v7.4s, v1.4s") // v5 = i0' i4' i1' i5'
__ASM_EMIT("uzp2 v1.4s, v7.4s, v1.4s") // v1 = i2' i6' i3' i7'
__ASM_EMIT("mov v2.16b, v0.16b") // v2 = r2' r6' r3' r7'
__ASM_EMIT("ins v0.d[1], v1.d[1]") // v0 = r2' r6' i3' i7'
__ASM_EMIT("ins v1.d[1], v2.d[1]") // v1 = i2' i6' r3' r7'
__ASM_EMIT("fadd v2.4s, v4.4s, v0.4s") // v2 = r0'+r2' r4'+r6' r1'+i3' r5'+i7' = r0" r4" r3" r7"
__ASM_EMIT("fsub v3.4s, v4.4s, v0.4s") // v3 = r0'-r2' r4'-r6' r1'-i3' r5'-i7' = r2" r6" r1" r5"
__ASM_EMIT("fadd v0.4s, v5.4s, v1.4s") // v0 = i0'+i2' i4'+i6' i1'+r3' i5'+r7' = i0" i4" i1" i5"
__ASM_EMIT("fsub v1.4s, v5.4s, v1.4s") // v1 = i0'-i2' i4'-i6' i1'-r3' i5'-r7' = i2" i6" i3" i7"
// v0 = i0" i4" i1" i5"
// v1 = i2" i6" i3" i7"
// v2 = r0" r4" r3" r7"
// v3 = r2" r6" r1" r5"
__ASM_EMIT("mov v4.16b, v2.16b") // v4 = r0" r4" r3" r7"
__ASM_EMIT("ins v2.d[1], v3.d[1]") // v2 = r0" r4" r1" r5"
__ASM_EMIT("ins v3.d[1], v4.d[1]") // v3 = r2" r6" r3" r7"
__ASM_EMIT("uzp1 v4.4s, v2.4s, v3.4s") // v4 = r0" r1" r2" r3"
__ASM_EMIT("uzp2 v5.4s, v2.4s, v3.4s") // v5 = r4" r5" r6" r7"
__ASM_EMIT("uzp1 v6.4s, v0.4s, v1.4s") // v6 = i0" i1" i2" i3"
__ASM_EMIT("uzp2 v7.4s, v0.4s, v1.4s") // v7 = i4" i5" i6" i7"
// Store
__ASM_EMIT("stp q4, q6, [%[dst], #0x00]")
__ASM_EMIT("stp q5, q7, [%[dst], #0x20]")
__ASM_EMIT("subs %[count], %[count], #8") // i <=> count
__ASM_EMIT("add %[dst], %[dst], #0x40")
__ASM_EMIT("b.ne 3b")
: [src] "+r" (src), [dst] "+r" (dst), [dp] "=&r" (dp),
[rrank] "+r" (rrank), [i] "=&r" (i), [j] "=&r" (j),
[count] "+r" (count)
:
: "cc", "memory",
"v0", "v1", "v2", "v3",
"v4", "v5", "v6", "v7"
);
}
void packed_scramble_copy_reverse(float *dst, const float *src, size_t rank)
{
IF_ARCH_AARCH64(
size_t i, j;
float *sr1, *sr2;
size_t regs = 1 << (rank - 3), rrank = 64 - (rank - 3);
);
ARCH_AARCH64_ASM(
__ASM_EMIT("eor %[i], %[i], %[i]") // i = 0
__ASM_EMIT("1:")
__ASM_EMIT("rbit %[j], %[i]") // j = reverse_bits(i)
__ASM_EMIT("lsr %[j], %[j], %[rrank]") // j = reverse_bits(i) >> rank
__ASM_EMIT("add %[sr1], %[src], %[j], LSL #3") // sr1 = &src[j]
__ASM_EMIT("add %[sr2], %[sr1], %[regs], LSL #3") // sr2 = &src[j + regs]
__ASM_EMIT("ld2 {v0.s, v1.s}[0], [%[sr1]]") // v0 = r0 ? ? ?, v1 = i0 ? ? ?
__ASM_EMIT("ld2 {v0.s, v1.s}[2], [%[sr2]]") // v0 = r0 ? r4 ?, v1 = i0 ? i4 ?
__ASM_EMIT("add %[sr1], %[sr1], %[regs], LSL #4") // sr1 = &src[j + regs*2]
__ASM_EMIT("add %[sr2], %[sr2], %[regs], LSL #4") // sr2 = &src[j + regs*3]
__ASM_EMIT("ld2 {v0.s, v1.s}[1], [%[sr1]]") // v0 = r0 r2 r4 ?, v1 = i0 i2 i4 ?
__ASM_EMIT("ld2 {v0.s, v1.s}[3], [%[sr2]]") // v0 = r0 r2 r4 r6, v1 = i0 i2 i4 i6
__ASM_EMIT("add %[sr1], %[sr1], %[regs], LSL #4") // sr1 = &src[j + regs*4]
__ASM_EMIT("add %[sr2], %[sr2], %[regs], LSL #4") // sr2 = &src[j + regs*5]
__ASM_EMIT("ld2 {v2.s, v3.s}[0], [%[sr1]]") // v2 = r1 ? ? ?, v3 = i1 ? ? ?
__ASM_EMIT("ld2 {v2.s, v3.s}[2], [%[sr2]]") // v2 = r1 ? r5 ?, v3 = i1 ? i5 ?
__ASM_EMIT("add %[sr1], %[sr1], %[regs], LSL #4") // sr1 = &src[j + regs*6]
__ASM_EMIT("add %[sr2], %[sr2], %[regs], LSL #4") // sr2 = &src[j + regs*7]
__ASM_EMIT("ld2 {v2.s, v3.s}[1], [%[sr1]]") // v2 = r1 r3 r5 ?, v3 = i1 i3 i5 ?
__ASM_EMIT("ld2 {v2.s, v3.s}[3], [%[sr2]]") // v2 = r1 r3 r5 r7, v3 = i1 i3 i5 i7
// v0 = r0 r2 r4 r6
// v1 = i0 i2 i4 i6
// v2 = r1 r3 r5 r7
// v3 = i1 i3 i5 i7
__ASM_EMIT("fadd v6.4s, v0.4s, v2.4s") // v6 = r0+r1 r2+r3 r4+r5 r6+r7 = r0' r2' r4' r6'
__ASM_EMIT("fadd v7.4s, v1.4s, v3.4s") // v7 = i0+i1 i2+i3 i4+i5 i6+i7 = i0' i2' i4' i6'
__ASM_EMIT("fsub v0.4s, v0.4s, v2.4s") // v0 = r0-r1 r2-r3 r4-r5 r6-r7 = r1' r3' r5' r7'
__ASM_EMIT("fsub v1.4s, v1.4s, v3.4s") // v1 = i0-i1 i2-i3 i4-i5 i6-i7 = i1' i3' i5' i7'
// v6 = r0' r2' r4' r6'
// v0 = r1' r3' r5' r7'
// v7 = i0' i2' i4' i6'
// v1 = i1' i3' i5' i7'
__ASM_EMIT("uzp1 v4.4s, v6.4s, v0.4s") // v4 = r0' r4' r1' r5'
__ASM_EMIT("uzp2 v0.4s, v6.4s, v0.4s") // v0 = r2' r6' r3' r7'
__ASM_EMIT("uzp1 v5.4s, v7.4s, v1.4s") // v5 = i0' i4' i1' i5'
__ASM_EMIT("uzp2 v1.4s, v7.4s, v1.4s") // v1 = i2' i6' i3' i7'
__ASM_EMIT("mov v2.16b, v0.16b") // v2 = r2' r6' r3' r7'
__ASM_EMIT("ins v0.d[1], v1.d[1]") // v0 = r2' r6' i3' i7'
__ASM_EMIT("ins v1.d[1], v2.d[1]") // v1 = i2' i6' r3' r7'
__ASM_EMIT("fadd v2.4s, v4.4s, v0.4s") // v2 = r0'+r2' r4'+r6' r1'+i3' r5'+i7' = r0" r4" r3" r7"
__ASM_EMIT("fsub v3.4s, v4.4s, v0.4s") // v3 = r0'-r2' r4'-r6' r1'-i3' r5'-i7' = r2" r6" r1" r5"
__ASM_EMIT("fadd v0.4s, v5.4s, v1.4s") // v0 = i0'+i2' i4'+i6' i1'+r3' i5'+r7' = i0" i4" i1" i5"
__ASM_EMIT("fsub v1.4s, v5.4s, v1.4s") // v1 = i0'-i2' i4'-i6' i1'-r3' i5'-r7' = i2" i6" i3" i7"
// v0 = i0" i4" i1" i5"
// v1 = i2" i6" i3" i7"
// v2 = r0" r4" r3" r7"
// v3 = r2" r6" r1" r5"
__ASM_EMIT("mov v4.16b, v2.16b") // v4 = r0" r4" r3" r7"
__ASM_EMIT("ins v2.d[1], v3.d[1]") // v2 = r0" r4" r1" r5"
__ASM_EMIT("ins v3.d[1], v4.d[1]") // v3 = r2" r6" r3" r7"
__ASM_EMIT("uzp1 v4.4s, v2.4s, v3.4s") // v4 = r0" r1" r2" r3"
__ASM_EMIT("uzp2 v5.4s, v2.4s, v3.4s") // v5 = r4" r5" r6" r7"
__ASM_EMIT("uzp1 v6.4s, v0.4s, v1.4s") // v6 = i0" i1" i2" i3"
__ASM_EMIT("uzp2 v7.4s, v0.4s, v1.4s") // v7 = i4" i5" i6" i7"
// Store
__ASM_EMIT("add %[i], %[i], #1")
__ASM_EMIT("stp q4, q6, [%[dst], #0x00]")
__ASM_EMIT("stp q5, q7, [%[dst], #0x20]")
__ASM_EMIT("cmp %[i], %[regs]")
__ASM_EMIT("add %[dst], %[dst], #0x40")
__ASM_EMIT("b.lo 1b")
: [dst] "+r" (dst), [sr1] "=&r" (sr1), [sr2] "=&r" (sr2),
[rrank] "+r" (rrank), [i] "=&r" (i), [j] "=&r" (j)
: [src] "r" (src), [regs] "r" (regs)
: "cc", "memory",
"v0", "v1", "v2", "v3",
"v4", "v5", "v6", "v7"
);
}
void packed_unscramble_direct(float *dst, size_t rank)
{
IF_ARCH_AARCH64(
size_t count = 1 << rank;
);
ARCH_AARCH64_ASM(
// 16x blocks
__ASM_EMIT("subs %[count], %[count], #16")
__ASM_EMIT("b.lo 2f")
__ASM_EMIT("1:")
__ASM_EMIT("ldp q0, q16, [%[dst], #0x00]")
__ASM_EMIT("ldp q2, q17, [%[dst], #0x20]")
__ASM_EMIT("ldp q4, q18, [%[dst], #0x40]")
__ASM_EMIT("ldp q6, q19, [%[dst], #0x60]")
__ASM_EMIT("zip2 v1.4s, v0.4s, v16.4s")
__ASM_EMIT("zip1 v0.4s, v0.4s, v16.4s")
__ASM_EMIT("zip2 v3.4s, v2.4s, v17.4s")
__ASM_EMIT("zip1 v2.4s, v2.4s, v17.4s")
__ASM_EMIT("zip2 v5.4s, v4.4s, v18.4s")
__ASM_EMIT("zip1 v4.4s, v4.4s, v18.4s")
__ASM_EMIT("zip2 v7.4s, v6.4s, v19.4s")
__ASM_EMIT("zip1 v6.4s, v6.4s, v19.4s")
__ASM_EMIT("stp q0, q1, [%[dst], #0x00]")
__ASM_EMIT("stp q2, q3, [%[dst], #0x20]")
__ASM_EMIT("stp q4, q5, [%[dst], #0x40]")
__ASM_EMIT("stp q6, q7, [%[dst], #0x60]")
__ASM_EMIT("subs %[count], %[count], #16")
__ASM_EMIT("add %[dst], %[dst], #0x80")
__ASM_EMIT("b.hs 1b")
__ASM_EMIT("2:")
// 8x block
__ASM_EMIT("adds %[count], %[count], #8")
__ASM_EMIT("b.lt 4f")
__ASM_EMIT("ldp q0, q16, [%[dst], #0x00]")
__ASM_EMIT("ldp q2, q17, [%[dst], #0x20]")
__ASM_EMIT("zip2 v1.4s, v0.4s, v16.4s")
__ASM_EMIT("zip1 v0.4s, v0.4s, v16.4s")
__ASM_EMIT("zip2 v3.4s, v2.4s, v17.4s")
__ASM_EMIT("zip1 v2.4s, v2.4s, v17.4s")
__ASM_EMIT("stp q0, q1, [%[dst], #0x00]")
__ASM_EMIT("stp q2, q3, [%[dst], #0x20]")
__ASM_EMIT("4:")
: [dst] "+r" (dst),
[count] "+r" (count)
:
: "cc", "memory",
"v0", "v1", "v2", "v3",
"v4", "v5", "v6", "v7",
"v16", "v17", "v18", "v19"
);
}
void packed_unscramble_reverse(float *dst, size_t rank)
{
IF_ARCH_AARCH64(
size_t count = 1 << rank;
float k = 1.0f/(1 << rank);
);
ARCH_AARCH64_ASM(
__ASM_EMIT("dup v31.4s, %[k].s[0]") // v31 = k
// 16x blocks
__ASM_EMIT("subs %[count], %[count], #16")
__ASM_EMIT("mov v30.16b, v31.16b")
__ASM_EMIT("b.lo 2f")
__ASM_EMIT("1:")
__ASM_EMIT("ldp q0, q16, [%[dst], #0x00]")
__ASM_EMIT("ldp q2, q17, [%[dst], #0x20]")
__ASM_EMIT("ldp q4, q18, [%[dst], #0x40]")
__ASM_EMIT("ldp q6, q19, [%[dst], #0x60]")
__ASM_EMIT("zip2 v1.4s, v0.4s, v16.4s")
__ASM_EMIT("zip1 v0.4s, v0.4s, v16.4s")
__ASM_EMIT("fmul v1.4s, v1.4s, v30.4s")
__ASM_EMIT("fmul v0.4s, v0.4s, v31.4s")
__ASM_EMIT("zip2 v3.4s, v2.4s, v17.4s")
__ASM_EMIT("zip1 v2.4s, v2.4s, v17.4s")
__ASM_EMIT("fmul v3.4s, v3.4s, v30.4s")
__ASM_EMIT("fmul v2.4s, v2.4s, v31.4s")
__ASM_EMIT("zip2 v5.4s, v4.4s, v18.4s")
__ASM_EMIT("zip1 v4.4s, v4.4s, v18.4s")
__ASM_EMIT("fmul v5.4s, v5.4s, v30.4s")
__ASM_EMIT("fmul v4.4s, v4.4s, v31.4s")
__ASM_EMIT("zip2 v7.4s, v6.4s, v19.4s")
__ASM_EMIT("zip1 v6.4s, v6.4s, v19.4s")
__ASM_EMIT("fmul v7.4s, v7.4s, v30.4s")
__ASM_EMIT("fmul v6.4s, v6.4s, v31.4s")
__ASM_EMIT("stp q0, q1, [%[dst], #0x00]")
__ASM_EMIT("stp q2, q3, [%[dst], #0x20]")
__ASM_EMIT("stp q4, q5, [%[dst], #0x40]")
__ASM_EMIT("stp q6, q7, [%[dst], #0x60]")
__ASM_EMIT("subs %[count], %[count], #16")
__ASM_EMIT("add %[dst], %[dst], #0x80")
__ASM_EMIT("b.hs 1b")
__ASM_EMIT("2:")
// 8x block
__ASM_EMIT("adds %[count], %[count], #8")
__ASM_EMIT("b.lt 4f")
__ASM_EMIT("ldp q0, q16, [%[dst], #0x00]")
__ASM_EMIT("ldp q2, q17, [%[dst], #0x20]")
__ASM_EMIT("zip2 v1.4s, v0.4s, v16.4s")
__ASM_EMIT("zip1 v0.4s, v0.4s, v16.4s")
__ASM_EMIT("fmul v1.4s, v1.4s, v30.4s")
__ASM_EMIT("fmul v0.4s, v0.4s, v31.4s")
__ASM_EMIT("zip2 v3.4s, v2.4s, v17.4s")
__ASM_EMIT("zip1 v2.4s, v2.4s, v17.4s")
__ASM_EMIT("fmul v3.4s, v3.4s, v30.4s")
__ASM_EMIT("fmul v2.4s, v2.4s, v31.4s")
__ASM_EMIT("stp q0, q1, [%[dst], #0x00]")
__ASM_EMIT("stp q2, q3, [%[dst], #0x20]")
__ASM_EMIT("4:")
: [dst] "+r" (dst),
[count] "+r" (count),
[k] "+w" (k)
:
: "cc", "memory",
"v0", "v1", "v2", "v3",
"v4", "v5", "v6", "v7",
"v16", "v17", "v18", "v19",
"v30", "v31"
);
}
}
}
#endif /* PRIVATE_DSP_ARCH_AARCH64_ASIMD_FFT_PSCRAMBLE_H_ */