/* * 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_X86_AVX_IMPL #error "This header should not be included directly" #endif /* PRIVATE_DSP_ARCH_X86_AVX_IMPL */ namespace lsp { namespace avx { static inline void FFT_SCRAMBLE_SELF_DIRECT_NAME(float *dst_re, float *dst_im, size_t rank) { // Calculate number of items size_t items = (1 << rank) - 1; for (size_t i = 1; i < items; ++i) { size_t j = reverse_bits(FFT_TYPE(i), rank); /* Reverse the order of the bits */ if (i >= j) continue; /* Copy the values from the reversed position */ ARCH_X86_ASM ( __ASM_EMIT("vmovss (%[dst_re], %[i], 4), %%xmm0") __ASM_EMIT("vmovss (%[dst_im], %[i], 4), %%xmm1") __ASM_EMIT("vmovss (%[dst_re], %[j], 4), %%xmm2") __ASM_EMIT("vmovss (%[dst_im], %[j], 4), %%xmm3") __ASM_EMIT("vmovss %%xmm2, (%[dst_re], %[i], 4)") __ASM_EMIT("vmovss %%xmm3, (%[dst_im], %[i], 4)") __ASM_EMIT("vmovss %%xmm0, (%[dst_re], %[j], 4)") __ASM_EMIT("vmovss %%xmm1, (%[dst_im], %[j], 4)") : : [dst_re] "r" (dst_re), [dst_im] "r" (dst_im), [i] "r"(i), [j] "r"(j) : "memory", "%xmm0", "%xmm1", "%xmm2", "%xmm3" ); } // Perform butterfly 8x size_t off = 0; items = 1 << (rank - 3); // Perform 4-element butterflies ARCH_X86_ASM ( /* Loop 2x 4-element butterflies */ __ASM_EMIT("sub $2, %[items]") __ASM_EMIT("jb 2f") __ASM_EMIT("1:") /* Load data to registers */ __ASM_EMIT("vmovups 0x00(%[dst_re], %[off]), %%xmm0") /* xmm0 = r0 r1 r2 r3 */ __ASM_EMIT("vmovups 0x10(%[dst_re], %[off]), %%xmm4") /* xmm4 = r4 r5 r6 r7 */ __ASM_EMIT("vinsertf128 $1, 0x20(%[dst_re], %[off]), %%ymm0, %%ymm0") /* ymm0 = r0 r1 r2 r3 */ __ASM_EMIT("vinsertf128 $1, 0x30(%[dst_re], %[off]), %%ymm4, %%ymm4") /* ymm4 = r4 r5 r6 r7 */ __ASM_EMIT("vmovups 0x00(%[dst_im], %[off]), %%xmm2") /* xmm2 = i0 i1 i2 i3 */ __ASM_EMIT("vmovups 0x10(%[dst_im], %[off]), %%xmm6") /* xmm6 = i4 i5 i6 i7 */ __ASM_EMIT("vinsertf128 $1, 0x20(%[dst_im], %[off]), %%ymm2, %%ymm2") /* ymm2 = i0 i1 i2 i3 */ __ASM_EMIT("vinsertf128 $1, 0x30(%[dst_im], %[off]), %%ymm6, %%ymm6") /* ymm6 = i4 i5 i6 i7 */ /* 1st-order 4x butterfly */ __ASM_EMIT("vhsubps %%ymm4, %%ymm0, %%ymm1") /* ymm1 = r0-r1 r2-r3 r4-r5 r6-r7 = r1' r3' r5' r7' */ __ASM_EMIT("vhsubps %%ymm6, %%ymm2, %%ymm3") /* ymm3 = i0-i1 i2-i3 i4-i5 i6-i7 = i1' i3' i5' i7' */ __ASM_EMIT("vhaddps %%ymm4, %%ymm0, %%ymm0") /* ymm0 = r0+r1 r2+r3 r4+r5 r6+r7 = r0' r2' r4' r6' */ __ASM_EMIT("vhaddps %%ymm6, %%ymm2, %%ymm2") /* ymm2 = i0+i1 i2+i3 i4+i5 i6+i7 = i0' i2' i4' i6' */ /* 2nd-order 4x butterfly */ __ASM_EMIT("vblendps $0xaa, %%ymm3, %%ymm1, %%ymm4") /* ymm4 = r1' i3' r5' i7' */ __ASM_EMIT("vblendps $0xaa, %%ymm1, %%ymm3, %%ymm5") /* ymm5 = i1' r3' i5' r7' */ __ASM_EMIT("vhsubps %%ymm4, %%ymm0, %%ymm1") /* ymm1 = r0'-r2' r4'-r6' r1'-i3' r5'-i7' = r2" r6" r3" r7" */ __ASM_EMIT("vhsubps %%ymm5, %%ymm2, %%ymm3") /* ymm3 = i0'-i2' i4'-i6' i1'-r3' i5'-r7' = i2" i6" i1" i5" */ __ASM_EMIT("vhaddps %%ymm4, %%ymm0, %%ymm0") /* ymm0 = r0'+r2' r4'+r6' r1'+i3' r5'+i7' = r0" r4" r1" r5" */ __ASM_EMIT("vhaddps %%ymm5, %%ymm2, %%ymm2") /* ymm2 = i0'+i2' i4'+i6' i1'+r3' i5'+r7' = i0" i4" i3" i7" */ __ASM_EMIT("vblendps $0xcc, %%ymm3, %%ymm2, %%ymm4") /* ymm4 = i0" i4" i1" i5" */ __ASM_EMIT("vblendps $0xcc, %%ymm2, %%ymm3, %%ymm5") /* ymm5 = i2" i6" i3" i7" */ __ASM_EMIT("vshufps $0x88, %%ymm1, %%ymm0, %%ymm2") /* ymm2 = r0" r1" r2" r3" */ __ASM_EMIT("vshufps $0xdd, %%ymm1, %%ymm0, %%ymm3") /* ymm3 = r4" r5" r6" r7" */ __ASM_EMIT("vshufps $0x88, %%ymm5, %%ymm4, %%ymm6") /* ymm6 = i0" i1" i2" i3" */ __ASM_EMIT("vshufps $0xdd, %%ymm5, %%ymm4, %%ymm7") /* ymm7 = i4" i5" i6" i7" */ /* 3rd-order 8x butterfly */ __ASM_EMIT("vmulps 0x20 + %[FFT_A], %%ymm3, %%ymm4") /* ymm4 = x_im * b_re */ \ __ASM_EMIT("vmulps 0x20 + %[FFT_A], %%ymm7, %%ymm5") /* ymm5 = x_im * b_im */ \ __ASM_EMIT(FFT_FMA("vmulps 0x00 + %[FFT_A], %%ymm3, %%ymm3", "")) /* ymm3 = x_re * b_re */ \ __ASM_EMIT(FFT_FMA("vmulps 0x00 + %[FFT_A], %%ymm7, %%ymm7", "")) /* ymm7 = x_re * b_im */ \ __ASM_EMIT(FFT_FMA("vaddps %%ymm5, %%ymm3, %%ymm5", "vfmadd231ps 0x00 + %[FFT_A], %%ymm3, %%ymm5")) /* ymm5 = c_re = x_re * b_re + x_im * b_im */ \ __ASM_EMIT(FFT_FMA("vsubps %%ymm4, %%ymm7, %%ymm4", "vfmsub231ps 0x00 + %[FFT_A], %%ymm7, %%ymm4")) /* ymm4 = c_im = x_re * b_im - x_im * b_re */ \ __ASM_EMIT("vsubps %%ymm5, %%ymm2, %%ymm0") /* ymm0 = a_re - c_re */ \ __ASM_EMIT("vsubps %%ymm4, %%ymm6, %%ymm1") /* ymm1 = a_im - c_im */ \ __ASM_EMIT("vaddps %%ymm5, %%ymm2, %%ymm2") /* ymm2 = a_re + c_re */ \ __ASM_EMIT("vaddps %%ymm4, %%ymm6, %%ymm3") /* ymm3 = a_im + c_im */ \ /* Store */ __ASM_EMIT("vmovups %%xmm2, 0x00(%[dst_re], %[off])") __ASM_EMIT("vmovups %%xmm0, 0x10(%[dst_re], %[off])") __ASM_EMIT("vextractf128 $1, %%ymm2, 0x20(%[dst_re], %[off])") __ASM_EMIT("vextractf128 $1, %%ymm0, 0x30(%[dst_re], %[off])") __ASM_EMIT("vmovups %%xmm3, 0x00(%[dst_im], %[off])") __ASM_EMIT("vmovups %%xmm1, 0x10(%[dst_im], %[off])") __ASM_EMIT("vextractf128 $1, %%ymm3, 0x20(%[dst_im], %[off])") __ASM_EMIT("vextractf128 $1, %%ymm1, 0x30(%[dst_im], %[off])") /* Move pointers and repeat*/ __ASM_EMIT("add $0x40, %[off]") __ASM_EMIT("sub $2, %[items]") __ASM_EMIT("jae 1b") __ASM_EMIT("2:") /* x4 scramble block */ __ASM_EMIT("add $1, %[items]") __ASM_EMIT("jl 4f") __ASM_EMIT("vmovups 0x00(%[dst_re], %[off]), %%xmm0") /* xmm0 = r0 r1 r2 r3 */ __ASM_EMIT("vmovups 0x10(%[dst_re], %[off]), %%xmm4") /* xmm4 = r4 r5 r6 r7 */ __ASM_EMIT("vmovups 0x00(%[dst_im], %[off]), %%xmm2") /* xmm2 = i0 i1 i2 i3 */ __ASM_EMIT("vmovups 0x10(%[dst_im], %[off]), %%xmm6") /* xmm6 = i4 i5 i6 i7 */ /* 1st-order 4x butterfly */ __ASM_EMIT("vhsubps %%xmm4, %%xmm0, %%xmm1") /* xmm1 = r0-r1 r2-r3 r4-r5 r6-r7 = r1' r3' r5' r7' */ __ASM_EMIT("vhsubps %%xmm6, %%xmm2, %%xmm3") /* xmm3 = i0-i1 i2-i3 i4-i5 i6-i7 = i1' i3' i5' i7' */ __ASM_EMIT("vhaddps %%xmm4, %%xmm0, %%xmm0") /* xmm0 = r0+r1 r2+r3 r4+r5 r6+r7 = r0' r2' r4' r6' */ __ASM_EMIT("vhaddps %%xmm6, %%xmm2, %%xmm2") /* xmm2 = i0+i1 i2+i3 i4+i5 i6+i7 = i0' i2' i4' i6' */ /* 2nd-order 4x butterfly */ __ASM_EMIT("vblendps $0xaa, %%xmm3, %%xmm1, %%xmm4") /* xmm4 = r1' i3' r5' i7' */ __ASM_EMIT("vblendps $0xaa, %%xmm1, %%xmm3, %%xmm5") /* xmm5 = i1' r3' i5' r7' */ __ASM_EMIT("vhsubps %%xmm4, %%xmm0, %%xmm1") /* xmm1 = r0'-r2' r4'-r6' r1'-i3' r5'-i7' = r2" r6" r3" r7" */ __ASM_EMIT("vhsubps %%xmm5, %%xmm2, %%xmm3") /* xmm3 = i0'-i2' i4'-i6' i1'-r3' i5'-r7' = i2" i6" i1" i5" */ __ASM_EMIT("vhaddps %%xmm4, %%xmm0, %%xmm0") /* xmm0 = r0'+r2' r4'+r6' r1'+i3' r5'+i7' = r0" r4" r1" r5" */ __ASM_EMIT("vhaddps %%xmm5, %%xmm2, %%xmm2") /* xmm2 = i0'+i2' i4'+i6' i1'+r3' i5'+r7' = i0" i4" i3" i7" */ __ASM_EMIT("vblendps $0xcc, %%xmm3, %%xmm2, %%xmm4") /* xmm4 = i0" i4" i1" i5" */ __ASM_EMIT("vblendps $0xcc, %%xmm2, %%xmm3, %%xmm5") /* xmm5 = i2" i6" i3" i7" */ __ASM_EMIT("vshufps $0x88, %%xmm1, %%xmm0, %%xmm2") /* xmm2 = r0" r1" r2" r3" */ __ASM_EMIT("vshufps $0xdd, %%xmm1, %%xmm0, %%xmm3") /* xmm3 = r4" r5" r6" r7" */ __ASM_EMIT("vshufps $0x88, %%xmm5, %%xmm4, %%xmm6") /* xmm6 = i0" i1" i2" i3" */ __ASM_EMIT("vshufps $0xdd, %%xmm5, %%xmm4, %%xmm7") /* xmm7 = i4" i5" i6" i7" */ /* 3rd-order 8x butterfly */ __ASM_EMIT("vmulps 0x20 + %[FFT_A], %%xmm3, %%xmm4") /* xmm4 = x_im * b_re */ \ __ASM_EMIT("vmulps 0x20 + %[FFT_A], %%xmm7, %%xmm5") /* xmm5 = x_im * b_im */ \ __ASM_EMIT(FFT_FMA("vmulps 0x00 + %[FFT_A], %%xmm3, %%xmm3", "")) /* xmm3 = x_re * b_re */ \ __ASM_EMIT(FFT_FMA("vmulps 0x00 + %[FFT_A], %%xmm7, %%xmm7", "")) /* xmm7 = x_re * b_im */ \ __ASM_EMIT(FFT_FMA("vaddps %%xmm5, %%xmm3, %%xmm5", "vfmadd231ps 0x00 + %[FFT_A], %%xmm3, %%xmm5")) /* xmm5 = c_re = x_re * b_re + x_im * b_im */ \ __ASM_EMIT(FFT_FMA("vsubps %%xmm4, %%xmm7, %%xmm4", "vfmsub231ps 0x00 + %[FFT_A], %%xmm7, %%xmm4")) /* xmm4 = c_im = x_re * b_im - x_im * b_re */ \ __ASM_EMIT("vsubps %%xmm5, %%xmm2, %%xmm0") /* xmm0 = a_re - c_re */ \ __ASM_EMIT("vsubps %%xmm4, %%xmm6, %%xmm1") /* xmm1 = a_im - c_im */ \ __ASM_EMIT("vaddps %%xmm5, %%xmm2, %%xmm2") /* xmm2 = a_re + c_re */ \ __ASM_EMIT("vaddps %%xmm4, %%xmm6, %%xmm3") /* xmm3 = a_im + c_im */ \ /* Store */ __ASM_EMIT("vmovups %%xmm2, 0x00(%[dst_re], %[off])") __ASM_EMIT("vmovups %%xmm0, 0x10(%[dst_re], %[off])") __ASM_EMIT("vmovups %%xmm3, 0x00(%[dst_im], %[off])") __ASM_EMIT("vmovups %%xmm1, 0x10(%[dst_im], %[off])") __ASM_EMIT("4:") : [dst_re] "+r"(dst_re), [dst_im] "+r"(dst_im), [off] "+r" (off), [items] "+r"(items) : [FFT_A] "o" (FFT_A) : "cc", "memory", "%xmm0", "%xmm1", "%xmm2", "%xmm3", "%xmm4", "%xmm5", "%xmm6", "%xmm7" ); } static inline void FFT_SCRAMBLE_SELF_REVERSE_NAME(float *dst_re, float *dst_im, size_t rank) { // Calculate number of items size_t items = (1 << rank) - 1; for (size_t i = 1; i < items; ++i) { size_t j = reverse_bits(FFT_TYPE(i), rank); /* Reverse the order of the bits */ if (i >= j) continue; /* Copy the values from the reversed position */ ARCH_X86_ASM ( __ASM_EMIT("vmovss (%[dst_re], %[i], 4), %%xmm0") __ASM_EMIT("vmovss (%[dst_im], %[i], 4), %%xmm1") __ASM_EMIT("vmovss (%[dst_re], %[j], 4), %%xmm2") __ASM_EMIT("vmovss (%[dst_im], %[j], 4), %%xmm3") __ASM_EMIT("vmovss %%xmm2, (%[dst_re], %[i], 4)") __ASM_EMIT("vmovss %%xmm3, (%[dst_im], %[i], 4)") __ASM_EMIT("vmovss %%xmm0, (%[dst_re], %[j], 4)") __ASM_EMIT("vmovss %%xmm1, (%[dst_im], %[j], 4)") : : [dst_re] "r"(dst_re), [dst_im] "r"(dst_im), [i] "r"(i), [j] "r"(j) : "memory", "%xmm0", "%xmm1", "%xmm2", "%xmm3" ); } // Perform butterfly 8x size_t off = 0; items = 1 << (rank - 3); // Perform 4-element butterflies ARCH_X86_ASM ( /* Loop 2x 4-element butterflies */ __ASM_EMIT("sub $2, %[items]") __ASM_EMIT("jb 2f") __ASM_EMIT("1:") /* Load data to registers */ __ASM_EMIT("vmovups 0x00(%[dst_re], %[off]), %%xmm0") /* xmm0 = r0 r1 r2 r3 */ __ASM_EMIT("vmovups 0x10(%[dst_re], %[off]), %%xmm4") /* xmm4 = r4 r5 r6 r7 */ __ASM_EMIT("vinsertf128 $1, 0x20(%[dst_re], %[off]), %%ymm0, %%ymm0") /* ymm0 = r0 r1 r2 r3 */ __ASM_EMIT("vinsertf128 $1, 0x30(%[dst_re], %[off]), %%ymm4, %%ymm4") /* ymm4 = r4 r5 r6 r7 */ __ASM_EMIT("vmovups 0x00(%[dst_im], %[off]), %%xmm2") /* xmm2 = i0 i1 i2 i3 */ __ASM_EMIT("vmovups 0x10(%[dst_im], %[off]), %%xmm6") /* xmm6 = i4 i5 i6 i7 */ __ASM_EMIT("vinsertf128 $1, 0x20(%[dst_im], %[off]), %%ymm2, %%ymm2") /* ymm2 = i0 i1 i2 i3 */ __ASM_EMIT("vinsertf128 $1, 0x30(%[dst_im], %[off]), %%ymm6, %%ymm6") /* ymm6 = i4 i5 i6 i7 */ /* 1st-order 4x butterfly */ __ASM_EMIT("vhsubps %%ymm4, %%ymm0, %%ymm1") /* ymm1 = r0-r1 r2-r3 r4-r5 r6-r7 = r1' r3' r5' r7' */ __ASM_EMIT("vhsubps %%ymm6, %%ymm2, %%ymm3") /* ymm3 = i0-i1 i2-i3 i4-i5 i6-i7 = i1' i3' i5' i7' */ __ASM_EMIT("vhaddps %%ymm4, %%ymm0, %%ymm0") /* ymm0 = r0+r1 r2+r3 r4+r5 r6+r7 = r0' r2' r4' r6' */ __ASM_EMIT("vhaddps %%ymm6, %%ymm2, %%ymm2") /* ymm2 = i0+i1 i2+i3 i4+i5 i6+i7 = i0' i2' i4' i6' */ /* 2nd-order 4x butterfly */ __ASM_EMIT("vblendps $0xaa, %%ymm3, %%ymm1, %%ymm4") /* ymm4 = r1' i3' r5' i7' */ __ASM_EMIT("vblendps $0xaa, %%ymm1, %%ymm3, %%ymm5") /* ymm5 = i1' r3' i5' r7' */ __ASM_EMIT("vhsubps %%ymm4, %%ymm0, %%ymm1") /* ymm1 = r0'-r2' r4'-r6' r1'-i3' r5'-i7' = r2" r6" r1" r5" */ __ASM_EMIT("vhsubps %%ymm5, %%ymm2, %%ymm3") /* ymm3 = i0'-i2' i4'-i6' i1'-r3' i5'-r7' = i2" i6" i3" i7" */ __ASM_EMIT("vhaddps %%ymm4, %%ymm0, %%ymm0") /* ymm0 = r0'+r2' r4'+r6' r1'+i3' r5'+i7' = r0" r4" r3" r7" */ __ASM_EMIT("vhaddps %%ymm5, %%ymm2, %%ymm2") /* ymm2 = i0'+i2' i4'+i6' i1'+r3' i5'+r7' = i0" i4" i1" i5" */ __ASM_EMIT("vblendps $0xcc, %%ymm1, %%ymm0, %%ymm4") /* ymm4 = r0" i4" r1" r5" */ __ASM_EMIT("vblendps $0xcc, %%ymm0, %%ymm1, %%ymm5") /* ymm5 = r2" r6" r3" r7" */ __ASM_EMIT("vshufps $0x88, %%ymm3, %%ymm2, %%ymm6") /* ymm6 = i0" i1" i2" i3" */ __ASM_EMIT("vshufps $0xdd, %%ymm3, %%ymm2, %%ymm7") /* ymm7 = i4" i5" i6" i7" */ __ASM_EMIT("vshufps $0x88, %%ymm5, %%ymm4, %%ymm2") /* ymm2 = r0" r1" r2" r3" */ __ASM_EMIT("vshufps $0xdd, %%ymm5, %%ymm4, %%ymm3") /* ymm3 = r4" r5" r6" r7" */ /* 3rd-order 8x butterfly */ __ASM_EMIT("vmulps 0x20 + %[FFT_A], %%ymm3, %%ymm4") /* ymm4 = x_im * b_re */ \ __ASM_EMIT("vmulps 0x20 + %[FFT_A], %%ymm7, %%ymm5") /* ymm5 = x_im * b_im */ \ __ASM_EMIT(FFT_FMA("vmulps 0x00 + %[FFT_A], %%ymm3, %%ymm3", "")) /* ymm3 = x_re * b_re */ \ __ASM_EMIT(FFT_FMA("vmulps 0x00 + %[FFT_A], %%ymm7, %%ymm7", "")) /* ymm7 = x_re * b_im */ \ __ASM_EMIT(FFT_FMA("vsubps %%ymm5, %%ymm3, %%ymm5", "vfmsub231ps 0x00 + %[FFT_A], %%ymm3, %%ymm5")) /* ymm5 = c_re = x_re * b_re - x_im * b_im */ \ __ASM_EMIT(FFT_FMA("vaddps %%ymm4, %%ymm7, %%ymm4", "vfmadd231ps 0x00 + %[FFT_A], %%ymm7, %%ymm4")) /* ymm4 = c_im = x_re * b_im + x_im * b_re */ \ __ASM_EMIT("vsubps %%ymm5, %%ymm2, %%ymm0") /* ymm0 = a_re - c_re */ \ __ASM_EMIT("vsubps %%ymm4, %%ymm6, %%ymm1") /* ymm1 = a_im - c_im */ \ __ASM_EMIT("vaddps %%ymm5, %%ymm2, %%ymm2") /* ymm2 = a_re + c_re */ \ __ASM_EMIT("vaddps %%ymm4, %%ymm6, %%ymm3") /* ymm3 = a_im + c_im */ \ /* Store */ __ASM_EMIT("vmovups %%xmm2, 0x00(%[dst_re], %[off])") __ASM_EMIT("vmovups %%xmm0, 0x10(%[dst_re], %[off])") __ASM_EMIT("vextractf128 $1, %%ymm2, 0x20(%[dst_re], %[off])") __ASM_EMIT("vextractf128 $1, %%ymm0, 0x30(%[dst_re], %[off])") __ASM_EMIT("vmovups %%xmm3, 0x00(%[dst_im], %[off])") __ASM_EMIT("vmovups %%xmm1, 0x10(%[dst_im], %[off])") __ASM_EMIT("vextractf128 $1, %%ymm3, 0x20(%[dst_im], %[off])") __ASM_EMIT("vextractf128 $1, %%ymm1, 0x30(%[dst_im], %[off])") /* Move pointers and repeat*/ __ASM_EMIT("add $0x40, %[off]") __ASM_EMIT("sub $2, %[items]") __ASM_EMIT("jae 1b") __ASM_EMIT("2:") /* x4 scramble block */ __ASM_EMIT("add $1, %[items]") __ASM_EMIT("jl 4f") __ASM_EMIT("vmovups 0x00(%[dst_re], %[off]), %%xmm0") /* xmm0 = r0 r1 r2 r3 */ __ASM_EMIT("vmovups 0x10(%[dst_re], %[off]), %%xmm4") /* xmm4 = r4 r5 r6 r7 */ __ASM_EMIT("vmovups 0x00(%[dst_im], %[off]), %%xmm2") /* xmm2 = i0 i1 i2 i3 */ __ASM_EMIT("vmovups 0x10(%[dst_im], %[off]), %%xmm6") /* xmm6 = i4 i5 i6 i7 */ /* 1st-order 4x butterfly */ __ASM_EMIT("vhsubps %%xmm4, %%xmm0, %%xmm1") /* xmm1 = r0-r1 r2-r3 r4-r5 r6-r7 = r1' r3' r5' r7' */ __ASM_EMIT("vhsubps %%xmm6, %%xmm2, %%xmm3") /* xmm3 = i0-i1 i2-i3 i4-i5 i6-i7 = i1' i3' i5' i7' */ __ASM_EMIT("vhaddps %%xmm4, %%xmm0, %%xmm0") /* xmm0 = r0+r1 r2+r3 r4+r5 r6+r7 = r0' r2' r4' r6' */ __ASM_EMIT("vhaddps %%xmm6, %%xmm2, %%xmm2") /* xmm2 = i0+i1 i2+i3 i4+i5 i6+i7 = i0' i2' i4' i6' */ /* 2nd-order 4x butterfly */ __ASM_EMIT("vblendps $0xaa, %%xmm3, %%xmm1, %%xmm4") /* xmm4 = r1' i3' r5' i7' */ __ASM_EMIT("vblendps $0xaa, %%xmm1, %%xmm3, %%xmm5") /* xmm5 = i1' r3' i5' r7' */ __ASM_EMIT("vhsubps %%xmm4, %%xmm0, %%xmm1") /* xmm1 = r0'-r2' r4'-r6' r1'-i3' r5'-i7' = r2" r6" r1" r5" */ __ASM_EMIT("vhsubps %%xmm5, %%xmm2, %%xmm3") /* xmm3 = i0'-i2' i4'-i6' i1'-r3' i5'-r7' = i2" i6" i3" i7" */ __ASM_EMIT("vhaddps %%xmm4, %%xmm0, %%xmm0") /* xmm0 = r0'+r2' r4'+r6' r1'+i3' r5'+i7' = r0" r4" r3" r7" */ __ASM_EMIT("vhaddps %%xmm5, %%xmm2, %%xmm2") /* xmm2 = i0'+i2' i4'+i6' i1'+r3' i5'+r7' = i0" i4" i1" i5" */ __ASM_EMIT("vblendps $0xcc, %%xmm1, %%xmm0, %%xmm4") /* xmm4 = r0" i4" r1" r5" */ __ASM_EMIT("vblendps $0xcc, %%xmm0, %%xmm1, %%xmm5") /* xmm5 = r2" r6" r3" r7" */ __ASM_EMIT("vshufps $0x88, %%xmm3, %%xmm2, %%xmm6") /* xmm6 = i0" i1" i2" i3" */ __ASM_EMIT("vshufps $0xdd, %%xmm3, %%xmm2, %%xmm7") /* xmm7 = i4" i5" i6" i7" */ __ASM_EMIT("vshufps $0x88, %%xmm5, %%xmm4, %%xmm2") /* xmm2 = r0" r1" r2" r3" */ __ASM_EMIT("vshufps $0xdd, %%xmm5, %%xmm4, %%xmm3") /* xmm3 = r4" r5" r6" r7" */ /* 3rd-order 8x butterfly */ __ASM_EMIT("vmulps 0x20 + %[FFT_A], %%xmm3, %%xmm4") /* xmm4 = x_im * b_re */ \ __ASM_EMIT("vmulps 0x20 + %[FFT_A], %%xmm7, %%xmm5") /* xmm5 = x_im * b_im */ \ __ASM_EMIT(FFT_FMA("vmulps 0x00 + %[FFT_A], %%xmm3, %%xmm3", "")) /* xmm3 = x_re * b_re */ \ __ASM_EMIT(FFT_FMA("vmulps 0x00 + %[FFT_A], %%xmm7, %%xmm7", "")) /* xmm7 = x_re * b_im */ \ __ASM_EMIT(FFT_FMA("vsubps %%xmm5, %%xmm3, %%xmm5", "vfmsub231ps 0x00 + %[FFT_A], %%xmm3, %%xmm5")) /* xmm5 = c_re = x_re * b_re - x_im * b_im */ \ __ASM_EMIT(FFT_FMA("vaddps %%xmm4, %%xmm7, %%xmm4", "vfmadd231ps 0x00 + %[FFT_A], %%xmm7, %%xmm4")) /* xmm4 = c_im = x_re * b_im + x_im * b_re */ \ __ASM_EMIT("vsubps %%xmm5, %%xmm2, %%xmm0") /* xmm0 = a_re - c_re */ \ __ASM_EMIT("vsubps %%xmm4, %%xmm6, %%xmm1") /* xmm1 = a_im - c_im */ \ __ASM_EMIT("vaddps %%xmm5, %%xmm2, %%xmm2") /* xmm2 = a_re + c_re */ \ __ASM_EMIT("vaddps %%xmm4, %%xmm6, %%xmm3") /* xmm3 = a_im + c_im */ \ /* Store */ __ASM_EMIT("vmovups %%xmm2, 0x00(%[dst_re], %[off])") __ASM_EMIT("vmovups %%xmm0, 0x10(%[dst_re], %[off])") __ASM_EMIT("vmovups %%xmm3, 0x00(%[dst_im], %[off])") __ASM_EMIT("vmovups %%xmm1, 0x10(%[dst_im], %[off])") __ASM_EMIT("4:") : [dst_re] "+r"(dst_re), [dst_im] "+r"(dst_im), [off] "+r" (off), [items] "+r"(items) : [FFT_A] "o" (FFT_A) : "cc", "memory", "%xmm0", "%xmm1", "%xmm2", "%xmm3", "%xmm4", "%xmm5", "%xmm6", "%xmm7" ); } static inline void FFT_SCRAMBLE_COPY_DIRECT_NAME(float *dst_re, float *dst_im, const float *src_re, const float *src_im, size_t rank) { size_t regs = 1 << rank; for (size_t i=0; i