/* * Copyright (C) 2025 Linux Studio Plugins Project * (C) 2025 Vladimir Sadovnikov * * This file is part of lsp-dsp-lib * Created on: 23 февр. 2025 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_AVX512_FFT_BUTTERFLY_H_ #define PRIVATE_DSP_ARCH_X86_AVX512_FFT_BUTTERFLY_H_ #ifndef PRIVATE_DSP_ARCH_X86_AVX512_IMPL #error "This header should not be included directly" #endif /* PRIVATE_DSP_ARCH_X86_AVX512_IMPL */ namespace lsp { namespace avx512 { #define FFT_BUTTERFLY_BODY16(fmadd_b, fmadd_a) \ __IF_32(float *ptr1, *ptr2);\ \ ARCH_X86_ASM( \ /* Prepare angle */ \ __ASM_EMIT32("mov %[fft_a], %[ptr2]") \ __ASM_EMIT32("mov %[dst_re], %[ptr1]") \ __ASM_EMIT("vmovaps 0x00(%[" __IF_32_64("ptr2", "fft_a") "]), %%zmm6") /* zmm6 = x_re */ \ __ASM_EMIT("vmovaps 0x40(%[" __IF_32_64("ptr2", "fft_a") "]), %%zmm7") /* zmm7 = x_im */ \ __ASM_EMIT32("mov %[dst_im], %[ptr2]") \ /* Start loop */ \ __ASM_EMIT("1:") \ __ASM_EMIT("vmovups 0x00(%[" __IF_32_64("ptr1", "dst_re") "], %[off1]), %%zmm0") /* zmm0 = a_re */ \ __ASM_EMIT("vmovups 0x00(%[" __IF_32_64("ptr1", "dst_re") "], %[off2]), %%zmm2") /* zmm2 = b_re */ \ __ASM_EMIT("vmovups 0x00(%[" __IF_32_64("ptr2", "dst_im") "], %[off1]), %%zmm1") /* zmm1 = a_im */ \ __ASM_EMIT("vmovups 0x00(%[" __IF_32_64("ptr2", "dst_im") "], %[off2]), %%zmm3") /* zmm3 = b_im */ \ /* Calculate complex multiplication */ \ __ASM_EMIT("vmulps %%zmm7, %%zmm2, %%zmm4") /* zmm4 = x_im * b_re */ \ __ASM_EMIT("vmulps %%zmm7, %%zmm3, %%zmm5") /* zmm5 = x_im * b_im */ \ __ASM_EMIT(fmadd_b " %%zmm6, %%zmm2, %%zmm5") /* zmm5 = c_re = x_re * b_re +- x_im * b_im */ \ __ASM_EMIT(fmadd_a " %%zmm6, %%zmm3, %%zmm4") /* zmm4 = c_im = x_re * b_im -+ x_im * b_re */ \ /* Perform butterfly */ \ __ASM_EMIT("vsubps %%zmm5, %%zmm0, %%zmm2") /* zmm2 = a_re - c_re */ \ __ASM_EMIT("vsubps %%zmm4, %%zmm1, %%zmm3") /* zmm3 = a_im - c_im */ \ __ASM_EMIT("vaddps %%zmm5, %%zmm0, %%zmm0") /* zmm0 = a_re + c_re */ \ __ASM_EMIT("vaddps %%zmm4, %%zmm1, %%zmm1") /* zmm1 = a_im + c_im */ \ /* Store values */ \ __ASM_EMIT("vmovups %%zmm0, 0x00(%[" __IF_32_64("ptr1", "dst_re") "], %[off1])") \ __ASM_EMIT("vmovups %%zmm2, 0x00(%[" __IF_32_64("ptr1", "dst_re") "], %[off2])") \ __ASM_EMIT("vmovups %%zmm1, 0x00(%[" __IF_32_64("ptr2", "dst_im") "], %[off1])") \ __ASM_EMIT("vmovups %%zmm3, 0x00(%[" __IF_32_64("ptr2", "dst_im") "], %[off2])") \ __ASM_EMIT("add $0x40, %[off1]") \ __ASM_EMIT("add $0x40, %[off2]") \ __ASM_EMIT32("subl $16, %[np]") \ __ASM_EMIT64("subq $16, %[np]") \ __ASM_EMIT("jz 2f") \ /* Rotate angle */ \ __ASM_EMIT32("mov %[fft_w], %[ptr2]") \ __ASM_EMIT("vmovaps 0x00(%[" __IF_32_64("ptr2", "fft_w") "]), %%zmm4") /* zmm4 = w_re */ \ __ASM_EMIT("vmovaps 0x40(%[" __IF_32_64("ptr2", "fft_w") "]), %%zmm5") /* zmm5 = w_im */ \ __ASM_EMIT32("mov %[dst_im], %[ptr2]") \ __ASM_EMIT("vmulps %%zmm5, %%zmm6, %%zmm2") /* zmm2 = w_im * x_re */ \ __ASM_EMIT("vmulps %%zmm5, %%zmm7, %%zmm3") /* zmm3 = w_im * x_im */ \ __ASM_EMIT("vfmsub132ps %%zmm4, %%zmm3, %%zmm6") /* zmm6 = x_re' = w_re * x_re - w_im * x_im */ \ __ASM_EMIT("vfmadd132ps %%zmm4, %%zmm2, %%zmm7") /* zmm7 = x_im' = w_re * x_im + w_im * x_re */ \ /* Repeat loop */ \ __ASM_EMIT("jmp 1b") \ __ASM_EMIT("2:") \ \ : __IF_32([ptr1] "=&r" (ptr1), [ptr2] "=&r" (ptr2), ) \ [off1] "+r" (off1), [off2] "+r" (off2), \ [np] X86_PGREG (np) \ : [dst_re] X86_GREG (dst_re), [dst_im] X86_GREG (dst_im), \ [fft_a] X86_GREG (fft_a), \ [fft_w] X86_GREG (fft_w) \ : "cc", "memory", \ "%xmm0", "%xmm1", "%xmm2", "%xmm3", \ "%xmm4", "%xmm5", "%xmm6", "%xmm7" \ ); static inline void butterfly_direct16p(float *dst_re, float *dst_im, size_t rank, size_t blocks) { size_t pairs = 1 << rank; size_t off1 = 0, shift = sizeof(float) << rank; const float *fft_a = &FFT_A[(rank - 2) << 5]; const float *fft_w = &FFT_DW[(rank - 3) << 5]; for (size_t b=0; b