// Copyright 2019 Google LLC // // This source code is licensed under the BSD-style license found in the // LICENSE file in the root directory of this source tree. #include "xnnpack/assembly.h" # void xnn_f32_igemm_minmax_ukernel_4x12__asm_aarch64_neonfma_cortex_a53( # size_t mr, x0 # size_t nc, x1 # size_t kc, x2 / x0 # size_t ks, x3 / x9 # const float** restrict a, x4 # const float* restrict w, x5 # float* restrict c, x6 # size_t cm_stride, x7 # size_t cn_stride, [sp] -> (x0) # size_t a_offset, [sp + 8] -> x11 # const float* zero, [sp + 16] -> x12 # const xnn_f32_minmax_params params [sp + 24] -> (x8) # d8-d15, x19-x30 need to be preserved if used. x18 is reserved by the OS. # Register usage # A0 x13 v0 # A1 x14 v0[1] # A2 x15 v1 # A3 x16 v1[1] # A0 x13 v2 # A1 x14 v2[1] # A2 x15 v3 # A3 x16 v3[1] # B v6 v7 v8 # B v9 v10 v11 # B v14 v15 v16 # B v17 v18 v19 # C0 x6 v20 v21 v22 # C1 x17 v23 v24 v25 # C2 x10 v26 v27 v28 # C3 x7 v29 v30 v31 # temporary vector shadow register x8 # Clamp v4 v5 # unused v12 v13 BEGIN_FUNCTION xnn_f32_igemm_minmax_ukernel_4x12__asm_aarch64_neonfma_cortex_a53 # Load a_offset LDR x11, [sp, 8] # Load zero, params pointer LDP x12, x8, [sp, 16] # Save d8-d11,d14,d15 on stack STP d8, d9, [sp, -48]! STP d10, d11, [sp, 16] STP d14, d15, [sp, 32] # Load min/max values LD2R {v4.4s, v5.4s}, [x8] # Clamp C pointers CMP x0, 2 // if mr < 2 ADD x17, x6, x7 // c1 = c0 + cm_stride CSEL x17, x6, x17, LO // c1 = c0 ADD x10, x17, x7 // c2 = c1 + cm_stride // if mr <= 2 CSEL x10, x17, x10, LS // c2 = c1 CMP x0, 4 // if mr < 4 ADD x7, x10, x7 // c3 = c2 + cm_stride CSEL x7, x10, x7, LO // c3 = c2 0: # Load initial bias from w into accumulators LD1 {v20.16b, v21.16b, v22.16b}, [x5], 48 MOV v23.16b, v20.16b PRFM PLDL1KEEP, [x5, 0] // Prefetch B MOV v24.16b, v21.16b PRFM PLDL1KEEP, [x5, 64] MOV v25.16b, v22.16b PRFM PLDL1KEEP, [x5, 128] MOV v26.16b, v20.16b PRFM PLDL1KEEP, [x5, 192] MOV v27.16b, v21.16b PRFM PLDL1KEEP, [x5, 256] MOV v28.16b, v22.16b PRFM PLDL1KEEP, [x5, 320] MOV v29.16b, v20.16b MOV v30.16b, v21.16b MOV v31.16b, v22.16b MOV x9, x3 // p = ks 1: # Load next 4 A pointers LDP x13, x14, [x4], 16 LDP x15, x16, [x4], 16 CMP x13, x12 // if a0 == zero ADD x13, x13, x11 // a0 += a_offset CSEL x13, x12, x13, EQ // a0 = zero, else += a0 + a_offset CMP x14, x12 // if a1 == zero ADD x14, x14, x11 // a1 += a_offset CSEL x14, x12, x14, EQ // a1 = zero, else += a1 + a_offset CMP x15, x12 // if a2 == zero ADD x15, x15, x11 // a2 += a_offset CSEL x15, x12, x15, EQ // a2 = zero, else += a2 + a_offset CMP x16, x12 // if a3 == zero ADD x16, x16, x11 // a3 += a_offset CSEL x16, x12, x16, EQ // a3 = zero, else += a3 + a_offset # Is there at least 4 floats (16 bytes) for prologue + epilogue? SUBS x0, x2, 16 // k = kc - 16 PRFM PLDL1KEEP, [x13, 0] // Prefetch A PRFM PLDL1KEEP, [x13, 64] PRFM PLDL1KEEP, [x14, 0] PRFM PLDL1KEEP, [x14, 64] PRFM PLDL1KEEP, [x15, 0] PRFM PLDL1KEEP, [x15, 64] PRFM PLDL1KEEP, [x16, 0] PRFM PLDL1KEEP, [x16, 64] B.LO 5f SUBS x0, x0, 16 // 4 floats for main loop # Prologue - loads for first group of 24 FMA # Read first block of 4 A. LDR d0, [x13], 8 // a0 LDR d1, [x15], 8 // a2 LD1 {v0.d}[1], [x14], 8 // a1 LD1 {v1.d}[1], [x16], 8 // a3 LD1 {v6.16b, v7.16b, v8.16b}, [x5], 48 LD1 {v9.16b, v10.16b}, [x5], 32 LDR d11, [x5], 8 LDR x8, [x5], 8 # Is there at least 4 floats (16 bytes) for main loop? B.LO 3f # Main loop - 4 floats of A (16 bytes) 2: # First group of 24 fma. 8 blocks of 4 cycles. LDR + 3 FMA # A is loaded for 2nd group into v2/v3 # INS is 4 blocks (16 cycles) after load # BLOCK 0 LDR d2, [x13], 8 // a0 INS v11.d[1], x8 FMLA v20.4s, v6.4s, v0.s[0] LDR x8, [x14], 8 // a1 FMLA v23.4s, v6.4s, v0.s[2] FMLA v26.4s, v6.4s, v1.s[0] PRFM PLDL1KEEP, [x13, 128] // Prefetch A0 # BLOCK 1 LDR d3, [x15], 8 // a2 INS v2.d[1], x8 // a1 was loaded in block 0 FMLA v29.4s, v6.4s, v1.s[2] LDR x8, [x16], 8 // a3 FMLA v21.4s, v7.4s, v0.s[0] FMLA v24.4s, v7.4s, v0.s[2] PRFM PLDL1KEEP, [x14, 128] // Prefetch A1 # BLOCK 2 LDR d14, [x5] // vb0x0123 INS v3.d[1], x8 // a3 was loaded in block 1 FMLA v27.4s, v7.4s, v1.s[0] LDR x8, [x5, 8] FMLA v30.4s, v7.4s, v1.s[2] FMLA v22.4s, v8.4s, v0.s[0] PRFM PLDL1KEEP, [x15, 128] // Prefetch A2 # BLOCK 3 LDR d15, [x5, 16] // vb0x4567 INS v14.d[1], x8 // v14 was loaded in block 2 FMLA v25.4s, v8.4s, v0.s[2] LDR x8, [x5, 24] FMLA v28.4s, v8.4s, v1.s[0] FMLA v31.4s, v8.4s, v1.s[2] PRFM PLDL1KEEP, [x16, 128] // Prefetch A3 # BLOCK 4 LDR d16, [x5, 32] // vb0x89AB INS v15.d[1], x8 FMLA v20.4s, v9.4s, v0.s[1] LDR x8, [x5, 40] FMLA v23.4s, v9.4s, v0.s[3] FMLA v26.4s, v9.4s, v1.s[1] PRFM PLDL1KEEP, [x5, 320] // Prefetch B # BLOCK 5 LDR d17, [x5, 48] // vb1x0123 INS v16.d[1], x8 FMLA v29.4s, v9.4s, v1.s[3] LDR x8, [x5, 56] FMLA v21.4s, v10.4s, v0.s[1] FMLA v24.4s, v10.4s, v0.s[3] PRFM PLDL1KEEP, [x5, 384] // Prefetch B # BLOCK 6 LDR d18, [x5, 64] // vb1x4567 INS v17.d[1], x8 FMLA v27.4s, v10.4s, v1.s[1] LDR x8, [x5, 72] FMLA v30.4s, v10.4s, v1.s[3] FMLA v22.4s, v11.4s, v0.s[1] PRFM PLDL1KEEP, [x5, 448] // Prefetch B # BLOCK 7 LDR d19, [x5, 80] // vb1x89AB INS v18.d[1], x8 FMLA v25.4s, v11.4s, v0.s[3] LDR x8, [x5, 88] FMLA v28.4s, v11.4s, v1.s[1] FMLA v31.4s, v11.4s, v1.s[3] # Second group of 24 fma. 8 blocks of 4 cycles. LDR + 3 FMA # A is loaded for 1st group into v0/v1 # BLOCK 0 LDR d0, [x13], 8 // a0 INS v19.d[1], x8 FMLA v20.4s, v14.4s, v2.s[0] LDR x8, [x14], 8 // a1 FMLA v23.4s, v14.4s, v2.s[2] FMLA v26.4s, v14.4s, v3.s[0] # BLOCK 1 LDR d1, [x15], 8 // a2 INS v0.d[1], x8 // a1 FMLA v29.4s, v14.4s, v3.s[2] LDR x8, [x16], 8 // a3 FMLA v21.4s, v15.4s, v2.s[0] FMLA v24.4s, v15.4s, v2.s[2] # BLOCK 2 LDR d6, [x5, 96] // vb0x0123 INS v1.d[1], x8 // a3 FMLA v27.4s, v15.4s, v3.s[0] LDR x8, [x5, 104] FMLA v30.4s, v15.4s, v3.s[2] FMLA v22.4s, v16.4s, v2.s[0] # BLOCK 3 LDR d7, [x5, 112] // vb0x4567 INS v6.d[1], x8 FMLA v25.4s, v16.4s, v2.s[2] LDR x8, [x5, 120] FMLA v28.4s, v16.4s, v3.s[0] FMLA v31.4s, v16.4s, v3.s[2] # BLOCK 4 LDR d8, [x5, 128] // vb0x89AB INS v7.d[1], x8 FMLA v20.4s, v17.4s, v2.s[1] LDR x8, [x5, 136] FMLA v23.4s, v17.4s, v2.s[3] FMLA v26.4s, v17.4s, v3.s[1] # BLOCK 5 LDR d9, [x5, 144] // vb1x0123 INS v8.d[1], x8 FMLA v29.4s, v17.4s, v3.s[3] LDR x8, [x5, 152] FMLA v21.4s, v18.4s, v2.s[1] FMLA v24.4s, v18.4s, v2.s[3] # BLOCK 6 LDR d10, [x5, 160] // vb1x4567 INS v9.d[1], x8 FMLA v27.4s, v18.4s, v3.s[1] LDR x8, [x5, 168] FMLA v30.4s, v18.4s, v3.s[3] SUBS x0, x0, 16 FMLA v22.4s, v19.4s, v2.s[1] # BLOCK 7 LDR d11, [x5, 176] // vb1x89AB INS v10.d[1], x8 FMLA v25.4s, v19.4s, v2.s[3] LDR x8, [x5, 184] FMLA v28.4s, v19.4s, v3.s[1] ADD x5, x5, 192 FMLA v31.4s, v19.4s, v3.s[3] B.HS 2b # Epilogue # First block same as main loop. Second block has no loads. 3: # BLOCK 0 LDR d2, [x13], 8 // a0 INS v11.d[1], x8 FMLA v20.4s, v6.4s, v0.s[0] LDR x8, [x14], 8 // a1 FMLA v23.4s, v6.4s, v0.s[2] FMLA v26.4s, v6.4s, v1.s[0] # BLOCK 1 LDR d3, [x15], 8 // a2 INS v2.d[1], x8 // a1 was loaded in block 0 FMLA v29.4s, v6.4s, v1.s[2] LDR x8, [x16], 8 // a3 FMLA v21.4s, v7.4s, v0.s[0] FMLA v24.4s, v7.4s, v0.s[2] # BLOCK 2 LDR d14, [x5] // vb0x0123 INS v3.d[1], x8 // a3 was loaded in block 1 FMLA v27.4s, v7.4s, v1.s[0] LDR x8, [x5, 8] FMLA v30.4s, v7.4s, v1.s[2] FMLA v22.4s, v8.4s, v0.s[0] # BLOCK 3 LDR d15, [x5, 16] // vb0x4567 INS v14.d[1], x8 // v14 was loaded in block 2 FMLA v25.4s, v8.4s, v0.s[2] LDR x8, [x5, 24] FMLA v28.4s, v8.4s, v1.s[0] FMLA v31.4s, v8.4s, v1.s[2] # BLOCK 4 LDR d16, [x5, 32] // vb0x89AB INS v15.d[1], x8 FMLA v20.4s, v9.4s, v0.s[1] LDR x8, [x5, 40] FMLA v23.4s, v9.4s, v0.s[3] FMLA v26.4s, v9.4s, v1.s[1] # BLOCK 5 LDR d17, [x5, 48] // vb1x0123 INS v16.d[1], x8 FMLA v29.4s, v9.4s, v1.s[3] LDR x8, [x5, 56] FMLA v21.4s, v10.4s, v0.s[1] FMLA v24.4s, v10.4s, v0.s[3] # BLOCK 6 LDR d18, [x5, 64] // vb1x4567 INS v17.d[1], x8 FMLA v27.4s, v10.4s, v1.s[1] LDR x8, [x5, 72] FMLA v30.4s, v10.4s, v1.s[3] FMLA v22.4s, v11.4s, v0.s[1] # BLOCK 7 LDR d19, [x5, 80] // vb1x89AB INS v18.d[1], x8 FMLA v25.4s, v11.4s, v0.s[3] LDR x8, [x5, 88] FMLA v28.4s, v11.4s, v1.s[1] FMLA v31.4s, v11.4s, v1.s[3] # Second group of 24 fma. 8 blocks of 4 cycles. LDR + 3 FMA # A is loaded for 1st group into v0/v1 # BLOCK 0 INS v19.d[1], x8 FMLA v20.4s, v14.4s, v2.s[0] FMLA v23.4s, v14.4s, v2.s[2] FMLA v26.4s, v14.4s, v3.s[0] # BLOCK 1 FMLA v29.4s, v14.4s, v3.s[2] FMLA v21.4s, v15.4s, v2.s[0] FMLA v24.4s, v15.4s, v2.s[2] # BLOCK 2 FMLA v27.4s, v15.4s, v3.s[0] FMLA v30.4s, v15.4s, v3.s[2] FMLA v22.4s, v16.4s, v2.s[0] # BLOCK 3 FMLA v25.4s, v16.4s, v2.s[2] FMLA v28.4s, v16.4s, v3.s[0] FMLA v31.4s, v16.4s, v3.s[2] # BLOCK 4 FMLA v20.4s, v17.4s, v2.s[1] FMLA v23.4s, v17.4s, v2.s[3] FMLA v26.4s, v17.4s, v3.s[1] # BLOCK 5 FMLA v29.4s, v17.4s, v3.s[3] FMLA v21.4s, v18.4s, v2.s[1] FMLA v24.4s, v18.4s, v2.s[3] # BLOCK 6 FMLA v27.4s, v18.4s, v3.s[1] FMLA v30.4s, v18.4s, v3.s[3] FMLA v22.4s, v19.4s, v2.s[1] TST x0, 15 # BLOCK 7 FMLA v25.4s, v19.4s, v2.s[3] FMLA v28.4s, v19.4s, v3.s[1] ADD x5, x5, 96 FMLA v31.4s, v19.4s, v3.s[3] # Is there a remainder?- 2 floats of A (8 bytes) or less B.NE 5f 4: # ks loop SUBS x9, x9, 32 // ks -= MR * sizeof(void*) B.HI 1b # Clamp FMAX v20.4s, v20.4s, v4.4s # Load cn_stride LDR x0, [sp, 48] FMAX v21.4s, v21.4s, v4.4s FMAX v22.4s, v22.4s, v4.4s FMAX v23.4s, v23.4s, v4.4s FMAX v24.4s, v24.4s, v4.4s FMAX v25.4s, v25.4s, v4.4s FMAX v26.4s, v26.4s, v4.4s FMAX v27.4s, v27.4s, v4.4s FMAX v28.4s, v28.4s, v4.4s FMAX v29.4s, v29.4s, v4.4s FMAX v30.4s, v30.4s, v4.4s FMAX v31.4s, v31.4s, v4.4s SUBS x1, x1, 12 FMIN v20.4s, v20.4s, v5.4s FMIN v21.4s, v21.4s, v5.4s FMIN v22.4s, v22.4s, v5.4s FMIN v23.4s, v23.4s, v5.4s FMIN v24.4s, v24.4s, v5.4s FMIN v25.4s, v25.4s, v5.4s FMIN v26.4s, v26.4s, v5.4s FMIN v27.4s, v27.4s, v5.4s FMIN v28.4s, v28.4s, v5.4s FMIN v29.4s, v29.4s, v5.4s FMIN v30.4s, v30.4s, v5.4s FMIN v31.4s, v31.4s, v5.4s # Store full 4 x 12 B.LO 7f ST1 {v29.16b, v30.16b, v31.16b}, [x7], x0 ST1 {v26.16b, v27.16b, v28.16b}, [x10], x0 ST1 {v23.16b, v24.16b, v25.16b}, [x17], x0 ST1 {v20.16b, v21.16b, v22.16b}, [x6], x0 SUB x4, x4, x3 // a -= ks # nc loop B.HI 0b # Restore d8-d11,d14,d15 from stack LDP d14, d15, [sp, 32] LDP d10, d11, [sp, 16] LDP d8, d9, [sp], 48 RET 5: # Is there a remainder?- 2 floats of A (8 bytes) TBZ x0, 3, 6f # Remainder- 2 floats of A (8 bytes) LDR d0, [x13], 8 // a0 LD1 {v6.16b, v7.16b, v8.16b}, [x5], 48 LDR d1, [x14], 8 // a1 LDR d2, [x15], 8 // a2 LDR d3, [x16], 8 // a3 LD1 {v9.16b, v10.16b, v11.16b}, [x5], 48 # First block of 3 B FMLA v20.4s, v6.4s, v0.s[0] FMLA v23.4s, v6.4s, v1.s[0] FMLA v26.4s, v6.4s, v2.s[0] FMLA v29.4s, v6.4s, v3.s[0] FMLA v21.4s, v7.4s, v0.s[0] FMLA v24.4s, v7.4s, v1.s[0] FMLA v27.4s, v7.4s, v2.s[0] FMLA v30.4s, v7.4s, v3.s[0] FMLA v22.4s, v8.4s, v0.s[0] FMLA v25.4s, v8.4s, v1.s[0] FMLA v28.4s, v8.4s, v2.s[0] FMLA v31.4s, v8.4s, v3.s[0] # Second block of 3 B FMLA v20.4s, v9.4s, v0.s[1] FMLA v23.4s, v9.4s, v1.s[1] FMLA v26.4s, v9.4s, v2.s[1] FMLA v29.4s, v9.4s, v3.s[1] FMLA v21.4s, v10.4s, v0.s[1] FMLA v24.4s, v10.4s, v1.s[1] FMLA v27.4s, v10.4s, v2.s[1] FMLA v30.4s, v10.4s, v3.s[1] FMLA v22.4s, v11.4s, v0.s[1] FMLA v25.4s, v11.4s, v1.s[1] FMLA v28.4s, v11.4s, v2.s[1] FMLA v31.4s, v11.4s, v3.s[1] # Is there a remainder?- 1 float of A (4 bytes) TBZ x0, 2, 4b 6: # Remainder- 1 float of A (4 bytes) LDR s0, [x13], 4 // a0 LD1 {v6.16b, v7.16b, v8.16b}, [x5], 48 LDR s1, [x14], 4 // a1 LDR s2, [x15], 4 // a2 LDR s3, [x16], 4 // a3 FMLA v20.4s, v6.4s, v0.s[0] FMLA v23.4s, v6.4s, v1.s[0] FMLA v26.4s, v6.4s, v2.s[0] FMLA v29.4s, v6.4s, v3.s[0] FMLA v21.4s, v7.4s, v0.s[0] FMLA v24.4s, v7.4s, v1.s[0] FMLA v27.4s, v7.4s, v2.s[0] FMLA v30.4s, v7.4s, v3.s[0] FMLA v22.4s, v8.4s, v0.s[0] FMLA v25.4s, v8.4s, v1.s[0] FMLA v28.4s, v8.4s, v2.s[0] FMLA v31.4s, v8.4s, v3.s[0] B 4b 7: ADD x1, x1, 12 # Store odd channels TBZ x1, 3, 8f STP q29, q30, [x7], 32 MOV v29.16b, v31.16b STP q26, q27, [x10], 32 MOV v26.16b, v28.16b STP q23, q24, [x17], 32 MOV v23.16b, v25.16b STP q20, q21, [x6], 32 MOV v20.16b, v22.16b 8: TBZ x1, 2, 9f STR q29, [x7], 16 MOV v29.16b, v30.16b STR q26, [x10], 16 MOV v26.16b, v27.16b STR q23, [x17], 16 MOV v23.16b, v24.16b STR q20, [x6], 16 MOV v20.16b, v21.16b 9: TBZ x1, 1, 10f STR d29, [x7], 8 DUP d29, v29.d[1] STR d26, [x10], 8 DUP d26, v26.d[1] STR d23, [x17], 8 DUP d23, v23.d[1] STR d20, [x6], 8 DUP d20, v20.d[1] 10: TBZ x1, 0, 11f STR s29, [x7] STR s26, [x10] STR s23, [x17] STR s20, [x6] 11: # Restore d8-d11,d14,d15 from stack LDP d14, d15, [sp, 32] LDP d10, d11, [sp, 16] LDP d8, d9, [sp], 48 RET END_FUNCTION xnn_f32_igemm_minmax_ukernel_4x12__asm_aarch64_neonfma_cortex_a53 #ifdef __ELF__ .section ".note.GNU-stack","",%progbits #endif