/* * Copyright (C) 2023 Linux Studio Plugins Project * (C) 2023 Vladimir Sadovnikov * * This file is part of lsp-dsp-lib * Created on: 20 окт. 2023 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_DYNAMICS_GATE_H_ #define PRIVATE_DSP_ARCH_AARCH64_ASIMD_DYNAMICS_GATE_H_ #ifndef PRIVATE_DSP_ARCH_AARCH64_ASIMD_IMPL #error "This header should not be included directly" #endif /* PRIVATE_DSP_ARCH_AARCH64_ASIMD_IMPL */ #include #include namespace lsp { namespace asimd { #define PROCESS_KNEE_SINGLE_X8 \ /* in: v0 = lx0, v1 = lx1 */ \ __ASM_EMIT("add %[off], %[gate], #0x10") \ __ASM_EMIT("ld4r {v8.4s, v9.4s, v10.4s, v11.4s}, [%[off]]" ) /* v8=herm[0], v9=herm[1], v10=herm[2], v11=herm[3] */ \ __ASM_EMIT("fmul v2.4s, v0.4s, v8.4s") /* v2 = herm[0]*lx0 */ \ __ASM_EMIT("fmul v3.4s, v1.4s, v8.4s") \ __ASM_EMIT("fadd v2.4s, v2.4s, v9.4s") /* v2 = herm[0]*lx0+herm[1] */ \ __ASM_EMIT("fadd v3.4s, v3.4s, v9.4s") \ __ASM_EMIT("fmul v2.4s, v2.4s, v0.4s") /* v2 = (herm[0]*lx0+herm[1])*lx0 */ \ __ASM_EMIT("fmul v3.4s, v3.4s, v1.4s") \ __ASM_EMIT("fadd v2.4s, v2.4s, v10.4s") /* v2 = (herm[0]*lx0+herm[1])*lx0+herm[2] */ \ __ASM_EMIT("fadd v3.4s, v3.4s, v10.4s") \ __ASM_EMIT("fmul v2.4s, v2.4s, v0.4s") /* v2 = ((herm[0]*lx0+herm[1])*lx0+herm[2])*lx0 */ \ __ASM_EMIT("fmul v3.4s, v3.4s, v1.4s") \ __ASM_EMIT("fadd v0.4s, v2.4s, v11.4s") /* v0 = TV = ((herm[0]*lx0+herm[1])*lx0+herm[2])*lx0+herm[3] */ \ __ASM_EMIT("fadd v1.4s, v3.4s, v11.4s") \ EXP_CORE_X8 /* v0 = KV = expf(TV) */ \ __ASM_EMIT("ldp q6, q7, [%[mem], #0x00]") /* v6 = x0 */ \ __ASM_EMIT("ld4r {v2.4s, v3.4s, v4.4s, v5.4s}, [%[gate]]" ) /* v2=start, v3=end, v4=gain_start, v5=gain_end */ \ __ASM_EMIT("fcmgt v8.4s, v6.4s, v2.4s") /* v8 = [x0 > start] */ \ __ASM_EMIT("fcmgt v9.4s, v7.4s, v2.4s") \ __ASM_EMIT("fcmge v6.4s, v6.4s, v3.4s") /* v6 = [x0 >= end] */ \ __ASM_EMIT("fcmge v7.4s, v7.4s, v3.4s") \ __ASM_EMIT("bif v0.16b, v4.16b, v8.16b") /* v0 = [x0 <= start] ? gain_start : KV */ \ __ASM_EMIT("bif v1.16b, v4.16b, v9.16b") \ __ASM_EMIT("bit v0.16b, v5.16b, v6.16b") /* v0 = [x0 >= end] ? gain_end : [x0 <= start] ? gain_start : KV */ \ __ASM_EMIT("bit v1.16b, v5.16b, v7.16b") \ /* out: v0 = g0, v1 = g1 */ #define PROCESS_KNEE_SINGLE_X4 \ /* in: v0 = lx0 */ \ __ASM_EMIT("add %[off], %[gate], #0x10") \ __ASM_EMIT("ld4r {v8.4s, v9.4s, v10.4s, v11.4s}, [%[off]]" ) /* v8=herm[0], v9=herm[1], v10=herm[2], v11=herm[3] */ \ __ASM_EMIT("fmul v2.4s, v0.4s, v8.4s") /* v2 = herm[0]*lx0 */ \ __ASM_EMIT("fadd v2.4s, v2.4s, v9.4s") /* v2 = herm[0]*lx0+herm[1] */ \ __ASM_EMIT("fmul v2.4s, v2.4s, v0.4s") /* v2 = (herm[0]*lx0+herm[1])*lx0 */ \ __ASM_EMIT("fadd v2.4s, v2.4s, v10.4s") /* v2 = (herm[0]*lx0+herm[1])*lx0+herm[2] */ \ __ASM_EMIT("fmul v2.4s, v2.4s, v0.4s") /* v2 = ((herm[0]*lx0+herm[1])*lx0+herm[2])*lx0 */ \ __ASM_EMIT("fadd v0.4s, v2.4s, v11.4s") /* v0 = TV = ((herm[0]*lx0+herm[1])*lx0+herm[2])*lx0+herm[3] */ \ EXP_CORE_X4 /* v0 = KV = expf(TV) */ \ __ASM_EMIT("ldp q6, q7, [%[mem], #0x00]") /* v6 = x0 */ \ __ASM_EMIT("ld4r {v2.4s, v3.4s, v4.4s, v5.4s}, [%[gate]]" ) /* v2=start, v3=end, v4=gain_start, v5=gain_end */ \ __ASM_EMIT("fcmgt v8.4s, v6.4s, v2.4s") /* v8 = [x0 > start] */ \ __ASM_EMIT("fcmge v6.4s, v6.4s, v3.4s") /* v6 = [x0 >= end] */ \ __ASM_EMIT("bif v0.16b, v4.16b, v8.16b") /* v0 = [x0 <= start] ? gain_start : KV */ \ __ASM_EMIT("bit v0.16b, v5.16b, v6.16b") /* v0 = [x0 >= end] ? gain_end : [x0 <= start] ? gain_start : KV */ \ /* out: v0 = g0 */ #define PROCESS_GATE_FULL_X8 \ /* in: v0 = x0, v1 = x1 */ \ __ASM_EMIT("ld2r {v12.4s, v13.4s}, [%[gate]]") /* v12 = begin, v13 = end */ \ __ASM_EMIT("fabs v0.4s, v0.4s") /* v0 = fabsf(x0) */ \ __ASM_EMIT("fabs v1.4s, v1.4s") \ __ASM_EMIT("fcmgt v2.4s, v0.4s, v12.4s") /* v2 = [fabs(x0) > begin] */ \ __ASM_EMIT("fcmgt v3.4s, v1.4s, v12.4s") \ __ASM_EMIT("fcmgt v4.4s, v13.4s, v0.4s") /* v4 = [fabs(x0) < end] */ \ __ASM_EMIT("fcmgt v5.4s, v13.4s, v1.4s") \ __ASM_EMIT("and v6.16b, v2.16b, v4.16b") /* v6 = [fabs(x0) > begin] && [fabs(x0) < end] */ \ __ASM_EMIT("and v7.16b, v3.16b, v5.16b") \ __ASM_EMIT("orr v6.16b, v6.16b, v7.16b") \ __ASM_EMIT("ext v7.16b, v6.16b, v6.16b, #8") \ __ASM_EMIT("orr v6.16b, v6.16b, v7.16b") \ __ASM_EMIT("ext v7.16b, v6.16b, v6.16b, #4") \ __ASM_EMIT("orr v6.16b, v6.16b, v7.16b") \ __ASM_EMIT("mov %w[off], v6.s[0]") \ __ASM_EMIT("cmp %w[off], #0") \ __ASM_EMIT("b.ne 100f") \ __ASM_EMIT("add %[off], %[gate], #0x08") \ __ASM_EMIT("ld1r {v0.4s}, [%[off]]") /* v0 = gain_begin */ \ __ASM_EMIT("add %[off], %[gate], #0x0c") \ __ASM_EMIT("ld1r {v4.4s}, [%[off]]") /* v4 = gain_end */ \ __ASM_EMIT("mov v1.16b, v0.16b") \ __ASM_EMIT("bit v0.16b, v4.16b, v2.16b") /* v0 = [fabs(x0) > start] ? gain_end : gain_start */ \ __ASM_EMIT("bit v1.16b, v4.16b, v3.16b") \ __ASM_EMIT("b 200f") \ __ASM_EMIT("100:") \ __ASM_EMIT("stp q0, q1, [%[mem], #0x00]") /* mem[0x00] = fabfs(x0) */ \ LOGE_CORE_X8 /* v0= lx0 = logf(fabsf(x0)) */ \ PROCESS_KNEE_SINGLE_X8 /* apply knee 0 */ \ __ASM_EMIT("200:") \ /* out: v0 = G0, v1= G1 */ #define PROCESS_GATE_FULL_X4 \ /* in: v0 = x0 */ \ __ASM_EMIT("fabs v0.4s, v0.4s") /* v0 = fabsf(x0) */ \ __ASM_EMIT("str q0, [%[mem], #0x00]") /* mem[0x00] = fabfs(x0) */ \ LOGE_CORE_X4 /* v0= lx0 = logf(fabsf(x0)) */ \ PROCESS_KNEE_SINGLE_X4 /* apply knee 0 */ \ /* out: v0 = G0 */ void gate_x1_gain(float *dst, const float *src, const dsp::gate_knee_t *c, size_t count) { IF_ARCH_AARCH64( float mem[8] __lsp_aligned16; size_t off; ); ARCH_AARCH64_ASM( // x8 blocks __ASM_EMIT("subs %[count], %[count], #8") __ASM_EMIT("b.lo 2f") __ASM_EMIT("1:") __ASM_EMIT("ldp q0, q1, [%[src], #0x00]") PROCESS_GATE_FULL_X8 __ASM_EMIT("subs %[count], %[count], #8") __ASM_EMIT("stp q0, q1, [%[dst], #0x00]") __ASM_EMIT("add %[src], %[src], 0x20") __ASM_EMIT("add %[dst], %[dst], 0x20") __ASM_EMIT("b.hs 1b") __ASM_EMIT("2:") // x4 block __ASM_EMIT("adds %[count], %[count], #4") __ASM_EMIT("b.lt 4f") __ASM_EMIT("ldr q0, [%[src], #0x00]") PROCESS_GATE_FULL_X4 __ASM_EMIT("sub %[count], %[count], #4") __ASM_EMIT("str q0, [%[dst], #0x00]") __ASM_EMIT("add %[src], %[src], 0x10") __ASM_EMIT("add %[dst], %[dst], 0x10") __ASM_EMIT("4:") // Tail: 1x-3x block __ASM_EMIT("adds %[count], %[count], #4") __ASM_EMIT("b.ls 12f") __ASM_EMIT("tst %[count], #1") __ASM_EMIT("b.eq 6f") __ASM_EMIT("ld1 {v0.s}[0], [%[src]]") __ASM_EMIT("add %[src], %[src], #0x04") __ASM_EMIT("6:") __ASM_EMIT("tst %[count], #2") __ASM_EMIT("b.eq 8f") __ASM_EMIT("ld1 {v0.d}[1], [%[src]]") __ASM_EMIT("8:") PROCESS_GATE_FULL_X4 __ASM_EMIT("tst %[count], #1") __ASM_EMIT("b.eq 10f") __ASM_EMIT("st1 {v0.s}[0], [%[dst]]") __ASM_EMIT("add %[dst], %[dst], #0x04") __ASM_EMIT("10:") __ASM_EMIT("tst %[count], #2") __ASM_EMIT("b.eq 12f") __ASM_EMIT("st1 {v0.d}[1], [%[dst]]") __ASM_EMIT("12:") : [dst] "+r" (dst), [src] "+r" (src), [count] "+r" (count), [off] "=&r" (off) : [gate] "r" (c), [L2C] "r" (&LOG2_CONST[0]), [LOGC] "r" (&LOGE_C[0]), [mem] "r" (&mem[0]), [E2C] "r" (&EXP2_CONST[0]), [LOG2E] "r" (&EXP_LOG2E[0]) : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v8", "v9", "v10", "v11", "v12", "v13", "v14", "v15", "v16", "v17", "v18", "v19", "v20", "v21", "v22", "v23", "v24", "v25", "v26", "v27", "v28", "v29", "v30", "v31" ); } void gate_x1_curve(float *dst, const float *src, const dsp::gate_knee_t *c, size_t count) { IF_ARCH_AARCH64( float mem[8] __lsp_aligned16; size_t off; ); ARCH_AARCH64_ASM( // x8 blocks __ASM_EMIT("subs %[count], %[count], #8") __ASM_EMIT("b.lo 2f") __ASM_EMIT("1:") __ASM_EMIT("ldp q0, q1, [%[src], #0x00]") PROCESS_GATE_FULL_X8 __ASM_EMIT("ldp q2, q3, [%[src], #0x00]") __ASM_EMIT("subs %[count], %[count], #8") __ASM_EMIT("fmul v0.4s, v0.4s, v2.4s") __ASM_EMIT("fmul v1.4s, v1.4s, v3.4s") __ASM_EMIT("stp q0, q1, [%[dst], #0x00]") __ASM_EMIT("add %[src], %[src], 0x20") __ASM_EMIT("add %[dst], %[dst], 0x20") __ASM_EMIT("b.hs 1b") __ASM_EMIT("2:") // x4 block __ASM_EMIT("adds %[count], %[count], #4") __ASM_EMIT("b.lt 4f") __ASM_EMIT("ldr q0, [%[src], #0x00]") PROCESS_GATE_FULL_X4 __ASM_EMIT("ldr q2, [%[src], #0x00]") __ASM_EMIT("sub %[count], %[count], #4") __ASM_EMIT("fmul v0.4s, v0.4s, v2.4s") __ASM_EMIT("str q0, [%[dst], #0x00]") __ASM_EMIT("add %[src], %[src], 0x10") __ASM_EMIT("add %[dst], %[dst], 0x10") __ASM_EMIT("4:") // Tail: 1x-3x block __ASM_EMIT("adds %[count], %[count], #4") __ASM_EMIT("b.ls 12f") __ASM_EMIT("tst %[count], #1") __ASM_EMIT("b.eq 6f") __ASM_EMIT("ld1 {v0.s}[0], [%[src]]") __ASM_EMIT("add %[src], %[src], #0x04") __ASM_EMIT("6:") __ASM_EMIT("tst %[count], #2") __ASM_EMIT("b.eq 8f") __ASM_EMIT("ld1 {v0.d}[1], [%[src]]") __ASM_EMIT("8:") __ASM_EMIT("mov v1.16b, v0.16b") PROCESS_GATE_FULL_X4 __ASM_EMIT("fmul v0.4s, v0.4s, v1.4s") __ASM_EMIT("tst %[count], #1") __ASM_EMIT("b.eq 10f") __ASM_EMIT("st1 {v0.s}[0], [%[dst]]") __ASM_EMIT("add %[dst], %[dst], #0x04") __ASM_EMIT("10:") __ASM_EMIT("tst %[count], #2") __ASM_EMIT("b.eq 12f") __ASM_EMIT("st1 {v0.d}[1], [%[dst]]") __ASM_EMIT("12:") : [dst] "+r" (dst), [src] "+r" (src), [count] "+r" (count), [off] "=&r" (off) : [gate] "r" (c), [L2C] "r" (&LOG2_CONST[0]), [LOGC] "r" (&LOGE_C[0]), [mem] "r" (&mem[0]), [E2C] "r" (&EXP2_CONST[0]), [LOG2E] "r" (&EXP_LOG2E[0]) : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v8", "v9", "v10", "v11", "v12", "v13", "v14", "v15", "v16", "v17", "v18", "v19", "v20", "v21", "v22", "v23", "v24", "v25", "v26", "v27", "v28", "v29", "v30", "v31" ); } #undef KNEE_LOAD #undef PROCESS_KNEE_SINGLE_X4 #undef PROCESS_KNEE_SINGLE_X8 #undef PROCESS_GATE_FULL_X8 #undef PROCESS_GATE_FULL_X4 } /* namespace asimd */ } /* namespace lsp */ #endif /* PRIVATE_DSP_ARCH_AARCH64_ASIMD_DYNAMICS_GATE_H_ */