diff --git a/erasure_code/gf_2vect_dot_prod_avx2_gfni.asm b/erasure_code/gf_2vect_dot_prod_avx2_gfni.asm index ca5dc12..5b69a2b 100644 --- a/erasure_code/gf_2vect_dot_prod_avx2_gfni.asm +++ b/erasure_code/gf_2vect_dot_prod_avx2_gfni.asm @@ -80,36 +80,42 @@ %define tmp3 r13 ; must be saved and restored %define tmp4 r14 ; must be saved and restored %define tmp5 rdi ; must be saved and restored - %define stack_size 4*16 + 5*8 ; must be an odd multiple of 8 + %define stack_size 7*16 + 5*8 ; must be an odd multiple of 8 %define arg(x) [rsp + stack_size + 8 + 8*x] %define func(x) proc_frame x %macro FUNC_SAVE 0 - alloc_stack stack_size - vmovdqa [rsp + 0*16], xmm6 - vmovdqa [rsp + 1*16], xmm7 - vmovdqa [rsp + 2*16], xmm8 - vmovdqa [rsp + 3*16], xmm9 - mov [rsp + 4*16 + 0*8], r12 - mov [rsp + 4*16 + 1*8], r13 - mov [rsp + 4*16 + 2*8], r14 - mov [rsp + 4*16 + 3*8], r15 - mov [rsp + 4*16 + 4*8], rdi - end_prolog - mov arg4, arg(4) + alloc_stack stack_size + vmovdqa [rsp + 0*16], xmm6 + vmovdqa [rsp + 1*16], xmm7 + vmovdqa [rsp + 2*16], xmm8 + vmovdqa [rsp + 3*16], xmm9 + vmovdqa [rsp + 4*16], xmm10 + vmovdqa [rsp + 5*16], xmm11 + vmovdqa [rsp + 6*16], xmm12 + mov [rsp + 7*16 + 0*8], r12 + mov [rsp + 7*16 + 1*8], r13 + mov [rsp + 7*16 + 2*8], r14 + mov [rsp + 7*16 + 3*8], r15 + mov [rsp + 7*16 + 4*8], rdi + end_prolog + mov arg4, arg(4) %endmacro %macro FUNC_RESTORE 0 - vmovdqa xmm6, [rsp + 0*16] - vmovdqa xmm7, [rsp + 1*16] - vmovdqa xmm8, [rsp + 2*16] - vmovdqa xmm9, [rsp + 3*16] - mov r12, [rsp + 4*16 + 0*8] - mov r13, [rsp + 4*16 + 1*8] - mov r14, [rsp + 4*16 + 2*8] - mov r15, [rsp + 4*16 + 3*8] - mov rdi, [rsp + 4*16 + 4*8] - add rsp, stack_size + vmovdqa xmm6, [rsp + 0*16] + vmovdqa xmm7, [rsp + 1*16] + vmovdqa xmm8, [rsp + 2*16] + vmovdqa xmm9, [rsp + 3*16] + vmovdqa xmm10, [rsp + 4*16] + vmovdqa xmm11, [rsp + 5*16] + vmovdqa xmm12, [rsp + 6*16] + mov r12, [rsp + 7*16 + 0*8] + mov r13, [rsp + 7*16 + 1*8] + mov r14, [rsp + 7*16 + 2*8] + mov r15, [rsp + 7*16 + 3*8] + mov rdi, [rsp + 7*16 + 4*8] + add rsp, stack_size %endmacro %endif @@ -142,17 +148,22 @@ %define x0l ymm0 %define x0h ymm1 +%define x0x ymm2 -%define xgft1 ymm2 -%define xgft2 ymm3 +%define xgft1 ymm3 +%define xgft2 ymm4 -%define xtmp1 ymm4 -%define xtmp2 ymm5 +%define xtmp1 ymm5 +%define xtmp2 ymm6 -%define xp1l ymm6 -%define xp2l ymm7 -%define xp1h ymm8 -%define xp2h ymm9 +%define xp1l ymm7 +%define xp2l ymm8 + +%define xp1h ymm9 +%define xp2h ymm10 + +%define xp1x ymm11 +%define xp2x ymm12 %define x0 x0l %define xp1 xp1l @@ -163,64 +174,105 @@ default rel section .text +;; +;; Encodes 96 bytes of all "k" sources into 2x 96 bytes (parity disk) +;; +%macro ENCODE_96B_2 0 + vpxor xp1l, xp1l, xp1l + vpxor xp1h, xp1h, xp1h + vpxor xp1x, xp1x, xp1x + + vpxor xp2l, xp2l, xp2l + vpxor xp2h, xp2h, xp2h + vpxor xp2x, xp2x, xp2x + mov tmp, mul_array + xor vec_i, vec_i + +%%next_vect: + ;; load next source vector + mov ptr, [src + vec_i] + XLDR x0l, [ptr + pos] + XLDR x0h, [ptr + pos + 32] + XLDR x0x, [ptr + pos + 64] + add vec_i, 8 + + vbroadcastsd xgft1, [tmp] + vbroadcastsd xgft2, [tmp + vec] + + GF_MUL_XOR VEX, x0l, xgft1, xtmp1, xp1l, xgft2, xtmp2, xp2l + GF_MUL_XOR VEX, x0h, xgft1, xtmp1, xp1h, xgft2, xtmp2, xp2h + GF_MUL_XOR VEX, x0x, xgft1, xtmp1, xp1x, xgft2, xtmp2, xp2x + add tmp, 8 + + cmp vec_i, vec + jl %%next_vect + + XSTR [dest1 + pos], xp1l + XSTR [dest1 + pos + 32], xp1h + XSTR [dest1 + pos + 64], xp1x + XSTR [dest2 + pos], xp2l + XSTR [dest2 + pos + 32], xp2h + XSTR [dest2 + pos + 64], xp2x +%endmacro + ;; ;; Encodes 64 bytes of all "k" sources into 2x 64 bytes (parity disks) ;; %macro ENCODE_64B_2 0 - vpxor xp1l, xp1l, xp1l - vpxor xp1h, xp1h, xp1h - vpxor xp2l, xp2l, xp2l - vpxor xp2h, xp2h, xp2h - mov tmp, mul_array - xor vec_i, vec_i + vpxor xp1l, xp1l, xp1l + vpxor xp1h, xp1h, xp1h + vpxor xp2l, xp2l, xp2l + vpxor xp2h, xp2h, xp2h + mov tmp, mul_array + xor vec_i, vec_i %%next_vect: - mov ptr, [src + vec_i] - XLDR x0l, [ptr + pos] ;; Get next source vector low 32 bytes - XLDR x0h, [ptr + pos + 32] ;; Get next source vector high 32 bytes - add vec_i, 8 + mov ptr, [src + vec_i] + XLDR x0l, [ptr + pos] ;; Get next source vector low 32 bytes + XLDR x0h, [ptr + pos + 32] ;; Get next source vector high 32 bytes + add vec_i, 8 - vbroadcastsd xgft1, [tmp] - vbroadcastsd xgft2, [tmp + vec] - add tmp, 8 + vbroadcastsd xgft1, [tmp] + vbroadcastsd xgft2, [tmp + vec] + add tmp, 8 - GF_MUL_XOR VEX, x0l, xgft1, xtmp1, xp1l, xgft2, xtmp2, xp2l - GF_MUL_XOR VEX, x0h, xgft1, xgft1, xp1h, xgft2, xgft2, xp2h + GF_MUL_XOR VEX, x0l, xgft1, xtmp1, xp1l, xgft2, xtmp2, xp2l + GF_MUL_XOR VEX, x0h, xgft1, xgft1, xp1h, xgft2, xgft2, xp2h - cmp vec_i, vec - jl %%next_vect + cmp vec_i, vec + jl %%next_vect - XSTR [dest1 + pos], xp1l - XSTR [dest1 + pos + 32], xp1h - XSTR [dest2 + pos], xp2l - XSTR [dest2 + pos + 32], xp2h + XSTR [dest1 + pos], xp1l + XSTR [dest1 + pos + 32], xp1h + XSTR [dest2 + pos], xp2l + XSTR [dest2 + pos + 32], xp2h %endmacro ;; ;; Encodes 32 bytes of all "k" sources into 2x 32 bytes (parity disks) ;; %macro ENCODE_32B_2 0 - vpxor xp1, xp1, xp1 - vpxor xp2, xp2, xp2 - mov tmp, mul_array - xor vec_i, vec_i + vpxor xp1, xp1, xp1 + vpxor xp2, xp2, xp2 + mov tmp, mul_array + xor vec_i, vec_i %%next_vect: - mov ptr, [src + vec_i] - XLDR x0, [ptr + pos] ;Get next source vector (32 bytes) - add vec_i, 8 + mov ptr, [src + vec_i] + XLDR x0, [ptr + pos] ;Get next source vector (32 bytes) + add vec_i, 8 - vbroadcastsd xgft1, [tmp] - vbroadcastsd xgft2, [tmp + vec] - add tmp, 8 + vbroadcastsd xgft1, [tmp] + vbroadcastsd xgft2, [tmp + vec] + add tmp, 8 - GF_MUL_XOR VEX, x0, xgft1, xgft1, xp1, xgft2, xgft2, xp2 + GF_MUL_XOR VEX, x0, xgft1, xgft1, xp1, xgft2, xgft2, xp2 - cmp vec_i, vec - jl %%next_vect + cmp vec_i, vec + jl %%next_vect - XSTR [dest1 + pos], xp1 - XSTR [dest2 + pos], xp2 + XSTR [dest1 + pos], xp1 + XSTR [dest2 + pos], xp2 %endmacro ;; @@ -229,73 +281,82 @@ section .text %macro ENCODE_LT_32B_2 1 %define %%LEN %1 - vpxor xp1, xp1, xp1 - vpxor xp2, xp2, xp2 - xor vec_i, vec_i + vpxor xp1, xp1, xp1 + vpxor xp2, xp2, xp2 + xor vec_i, vec_i %%next_vect: - mov ptr, [src + vec_i] - simd_load_avx2 x0, ptr + pos, %%LEN, tmp, tmp4 ;Get next source vector - add vec_i, 8 + mov ptr, [src + vec_i] + simd_load_avx2 x0, ptr + pos, %%LEN, tmp, tmp4 ;Get next source vector + add vec_i, 8 - vbroadcastsd xgft1, [mul_array] - vbroadcastsd xgft2, [mul_array + vec] - add mul_array, 8 + vbroadcastsd xgft1, [mul_array] + vbroadcastsd xgft2, [mul_array + vec] + add mul_array, 8 - GF_MUL_XOR VEX, x0, xgft1, xgft1, xp1, xgft2, xgft2, xp2 + GF_MUL_XOR VEX, x0, xgft1, xgft1, xp1, xgft2, xgft2, xp2 - cmp vec_i, vec - jl %%next_vect + cmp vec_i, vec + jl %%next_vect - ;Store updated encoded data - lea ptr, [dest1 + pos] - simd_store_avx2 ptr, xp1, %%LEN, tmp, vec_i + ;Store updated encoded data + lea ptr, [dest1 + pos] + simd_store_avx2 ptr, xp1, %%LEN, tmp, tmp4 - lea ptr, [dest2 + pos] - simd_store_avx2 ptr, xp2, %%LEN, tmp, vec_i + lea ptr, [dest2 + pos] + simd_store_avx2 ptr, xp2, %%LEN, tmp, tmp4 %endmacro align 16 mk_global gf_2vect_dot_prod_avx2_gfni, function func(gf_2vect_dot_prod_avx2_gfni) - FUNC_SAVE + FUNC_SAVE - xor pos, pos - shl vec, 3 ;; vec *= 8. Make vec_i count by 8 - mov dest1, [dest] - mov dest2, [dest + 8] + xor pos, pos + shl vec, 3 ;; vec *= 8. Make vec_i count by 8 + mov dest1, [dest] + mov dest2, [dest + 8] - cmp len, 64 - jb .len_lt_64 + cmp len, 96 + jl .len_lt_96 -.loop64: - ENCODE_64B_2 +.loop96: + ENCODE_96B_2 - add pos, 64 ;; Loop on 64 bytes at a time first - sub len, 64 - cmp len, 64 - jge .loop64 + add pos, 96 ;; Loop on 96 bytes at a time first + sub len, 96 + cmp len, 96 + jge .loop96 + +.len_lt_96: + cmp len, 64 + jl .len_lt_64 + + ENCODE_64B_2 + + add pos, 64 ;; encode next 64 bytes + sub len, 64 .len_lt_64: - cmp len, 32 - jb .len_lt_32 + cmp len, 32 + jl .len_lt_32 - ENCODE_32B_2 + ENCODE_32B_2 - add pos, 32 ;; encode next 32 bytes - sub len, 32 + add pos, 32 ;; encode next 32 bytes + sub len, 32 .len_lt_32: - cmp len, 0 - jle .exit + cmp len, 0 + jle .exit - ENCODE_LT_32B_2 len ;; encode remaining bytes + ENCODE_LT_32B_2 len ;; encode remaining bytes .exit: - vzeroupper + vzeroupper - FUNC_RESTORE - ret + FUNC_RESTORE + ret endproc_frame %endif ; if AS_FEATURE_LEVEL >= 10