diff --git a/src/fdr/teddy.c b/src/fdr/teddy.c index 9f8b5104..e7a0fccd 100644 --- a/src/fdr/teddy.c +++ b/src/fdr/teddy.c @@ -125,36 +125,34 @@ do { \ #endif static really_inline -m128 prep_conf_teddy_m1(const m128 *maskBase, m128 p_mask, m128 val) { +m128 prep_conf_teddy_m1(const m128 *maskBase, m128 val) { m128 mask = set16x8(0xf); m128 lo = and128(val, mask); m128 hi = and128(rshift64_m128(val, 4), mask); - return and128(and128(pshufb(maskBase[0*2], lo), - pshufb(maskBase[0*2+1], hi)), p_mask); + return and128(pshufb(maskBase[0*2], lo), pshufb(maskBase[0*2+1], hi)); } static really_inline -m128 prep_conf_teddy_m2(const m128 *maskBase, m128 *old_1, m128 p_mask, - m128 val) { +m128 prep_conf_teddy_m2(const m128 *maskBase, m128 *old_1, m128 val) { m128 mask = set16x8(0xf); m128 lo = and128(val, mask); m128 hi = and128(rshift64_m128(val, 4), mask); - m128 r = prep_conf_teddy_m1(maskBase, p_mask, val); + m128 r = prep_conf_teddy_m1(maskBase, val); m128 res_1 = and128(pshufb(maskBase[1*2], lo), pshufb(maskBase[1*2+1], hi)); m128 res_shifted_1 = palignr(res_1, *old_1, 16-1); *old_1 = res_1; - return and128(and128(r, p_mask), res_shifted_1); + return and128(r, res_shifted_1); } static really_inline m128 prep_conf_teddy_m3(const m128 *maskBase, m128 *old_1, m128 *old_2, - m128 p_mask, m128 val) { + m128 val) { m128 mask = set16x8(0xf); m128 lo = and128(val, mask); m128 hi = and128(rshift64_m128(val, 4), mask); - m128 r = prep_conf_teddy_m2(maskBase, old_1, p_mask, val); + m128 r = prep_conf_teddy_m2(maskBase, old_1, val); m128 res_2 = and128(pshufb(maskBase[2*2], lo), pshufb(maskBase[2*2+1], hi)); @@ -165,11 +163,11 @@ m128 prep_conf_teddy_m3(const m128 *maskBase, m128 *old_1, m128 *old_2, static really_inline m128 prep_conf_teddy_m4(const m128 *maskBase, m128 *old_1, m128 *old_2, - m128 *old_3, m128 p_mask, m128 val) { + m128 *old_3, m128 val) { m128 mask = set16x8(0xf); m128 lo = and128(val, mask); m128 hi = and128(rshift64_m128(val, 4), mask); - m128 r = prep_conf_teddy_m3(maskBase, old_1, old_2, p_mask, val); + m128 r = prep_conf_teddy_m3(maskBase, old_1, old_2, val); m128 res_3 = and128(pshufb(maskBase[3*2], lo), pshufb(maskBase[3*2+1], hi)); @@ -201,13 +199,14 @@ hwlm_error_t fdr_exec_teddy_msks1(const struct FDR *fdr, m128 p_mask; m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 1); - m128 r_0 = prep_conf_teddy_m1(maskBase, p_mask, val_0); + m128 r_0 = prep_conf_teddy_m1(maskBase, val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit1_teddy); ptr += 16; } if (ptr + 16 < buf_end) { - m128 r_0 = prep_conf_teddy_m1(maskBase, ones128(), load128(ptr)); + m128 r_0 = prep_conf_teddy_m1(maskBase, load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit1_teddy); ptr += 16; } @@ -215,9 +214,9 @@ hwlm_error_t fdr_exec_teddy_msks1(const struct FDR *fdr, for (; ptr + iterBytes <= buf_end; ptr += iterBytes) { __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; - m128 r_0 = prep_conf_teddy_m1(maskBase, ones128(), load128(ptr)); + m128 r_0 = prep_conf_teddy_m1(maskBase, load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, NOT_CAUTIOUS, do_confWithBit1_teddy); - m128 r_1 = prep_conf_teddy_m1(maskBase, ones128(), load128(ptr + 16)); + m128 r_1 = prep_conf_teddy_m1(maskBase, load128(ptr + 16)); CONFIRM_TEDDY(r_1, 8, 16, NOT_CAUTIOUS, do_confWithBit1_teddy); } @@ -225,7 +224,8 @@ hwlm_error_t fdr_exec_teddy_msks1(const struct FDR *fdr, m128 p_mask; m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 1); - m128 r_0 = prep_conf_teddy_m1(maskBase, p_mask, val_0); + m128 r_0 = prep_conf_teddy_m1(maskBase, val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit1_teddy); } @@ -255,13 +255,14 @@ hwlm_error_t fdr_exec_teddy_msks1_pck(const struct FDR *fdr, m128 p_mask; m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 1); - m128 r_0 = prep_conf_teddy_m1(maskBase, p_mask, val_0); + m128 r_0 = prep_conf_teddy_m1(maskBase, val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } if (ptr + 16 < buf_end) { - m128 r_0 = prep_conf_teddy_m1(maskBase, ones128(), load128(ptr)); + m128 r_0 = prep_conf_teddy_m1(maskBase, load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } @@ -269,9 +270,9 @@ hwlm_error_t fdr_exec_teddy_msks1_pck(const struct FDR *fdr, for (; ptr + iterBytes <= buf_end; ptr += iterBytes) { __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; - m128 r_0 = prep_conf_teddy_m1(maskBase, ones128(), load128(ptr)); + m128 r_0 = prep_conf_teddy_m1(maskBase, load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, NOT_CAUTIOUS, do_confWithBit_teddy); - m128 r_1 = prep_conf_teddy_m1(maskBase, ones128(), load128(ptr + 16)); + m128 r_1 = prep_conf_teddy_m1(maskBase, load128(ptr + 16)); CONFIRM_TEDDY(r_1, 8, 16, NOT_CAUTIOUS, do_confWithBit_teddy); } @@ -279,7 +280,8 @@ hwlm_error_t fdr_exec_teddy_msks1_pck(const struct FDR *fdr, m128 p_mask; m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 1); - m128 r_0 = prep_conf_teddy_m1(maskBase, p_mask, val_0); + m128 r_0 = prep_conf_teddy_m1(maskBase, val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit_teddy); } @@ -310,14 +312,14 @@ hwlm_error_t fdr_exec_teddy_msks2(const struct FDR *fdr, m128 p_mask; m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 2); - m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, p_mask, val_0); + m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBitMany_teddy); ptr += 16; } if (ptr + 16 < buf_end) { - m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, ones128(), - load128(ptr)); + m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBitMany_teddy); ptr += 16; } @@ -325,11 +327,9 @@ hwlm_error_t fdr_exec_teddy_msks2(const struct FDR *fdr, for (; ptr + iterBytes <= buf_end; ptr += iterBytes) { __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; - m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, ones128(), - load128(ptr)); + m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, NOT_CAUTIOUS, do_confWithBitMany_teddy); - m128 r_1 = prep_conf_teddy_m2(maskBase, &res_old_1, ones128(), - load128(ptr + 16)); + m128 r_1 = prep_conf_teddy_m2(maskBase, &res_old_1, load128(ptr + 16)); CONFIRM_TEDDY(r_1, 8, 16, NOT_CAUTIOUS, do_confWithBitMany_teddy); } @@ -337,7 +337,8 @@ hwlm_error_t fdr_exec_teddy_msks2(const struct FDR *fdr, m128 p_mask; m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 2); - m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, p_mask, val_0); + m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBitMany_teddy); } @@ -368,14 +369,14 @@ hwlm_error_t fdr_exec_teddy_msks2_pck(const struct FDR *fdr, m128 p_mask; m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 2); - m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, p_mask, val_0); + m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } if (ptr + 16 < buf_end) { - m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, ones128(), - load128(ptr)); + m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } @@ -383,11 +384,9 @@ hwlm_error_t fdr_exec_teddy_msks2_pck(const struct FDR *fdr, for (; ptr + iterBytes <= buf_end; ptr += iterBytes) { __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; - m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, ones128(), - load128(ptr)); + m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, NOT_CAUTIOUS, do_confWithBit_teddy); - m128 r_1 = prep_conf_teddy_m2(maskBase, &res_old_1, ones128(), - load128(ptr + 16)); + m128 r_1 = prep_conf_teddy_m2(maskBase, &res_old_1, load128(ptr + 16)); CONFIRM_TEDDY(r_1, 8, 16, NOT_CAUTIOUS, do_confWithBit_teddy); } @@ -395,7 +394,8 @@ hwlm_error_t fdr_exec_teddy_msks2_pck(const struct FDR *fdr, m128 p_mask; m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 2); - m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, p_mask, val_0); + m128 r_0 = prep_conf_teddy_m2(maskBase, &res_old_1, val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit_teddy); } @@ -428,14 +428,15 @@ hwlm_error_t fdr_exec_teddy_msks3(const struct FDR *fdr, m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 3); m128 r_0 = prep_conf_teddy_m3(maskBase, &res_old_1, &res_old_2, - p_mask, val_0); + val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBitMany_teddy); ptr += 16; } if (ptr + 16 < buf_end) { m128 r_0 = prep_conf_teddy_m3(maskBase, &res_old_1, &res_old_2, - ones128(), load128(ptr)); + load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBitMany_teddy); ptr += 16; } @@ -444,10 +445,10 @@ hwlm_error_t fdr_exec_teddy_msks3(const struct FDR *fdr, __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; m128 r_0 = prep_conf_teddy_m3(maskBase, &res_old_1, &res_old_2, - ones128(), load128(ptr)); + load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, NOT_CAUTIOUS, do_confWithBitMany_teddy); m128 r_1 = prep_conf_teddy_m3(maskBase, &res_old_1, &res_old_2, - ones128(), load128(ptr + 16)); + load128(ptr + 16)); CONFIRM_TEDDY(r_1, 8, 16, NOT_CAUTIOUS, do_confWithBitMany_teddy); } @@ -455,8 +456,8 @@ hwlm_error_t fdr_exec_teddy_msks3(const struct FDR *fdr, m128 p_mask; m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 3); - m128 r_0 = prep_conf_teddy_m3(maskBase, &res_old_1, &res_old_2, - p_mask, val_0); + m128 r_0 = prep_conf_teddy_m3(maskBase, &res_old_1, &res_old_2, val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBitMany_teddy); } @@ -489,14 +490,15 @@ hwlm_error_t fdr_exec_teddy_msks3_pck(const struct FDR *fdr, m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 3); m128 r_0 = prep_conf_teddy_m3(maskBase, &res_old_1, &res_old_2, - p_mask, val_0); + val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } if (ptr + 16 < buf_end) { m128 r_0 = prep_conf_teddy_m3(maskBase, &res_old_1, &res_old_2, - ones128(), load128(ptr)); + load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } @@ -505,10 +507,10 @@ hwlm_error_t fdr_exec_teddy_msks3_pck(const struct FDR *fdr, __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; m128 r_0 = prep_conf_teddy_m3(maskBase, &res_old_1, &res_old_2, - ones128(), load128(ptr)); + load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, NOT_CAUTIOUS, do_confWithBit_teddy); m128 r_1 = prep_conf_teddy_m3(maskBase, &res_old_1, &res_old_2, - ones128(), load128(ptr + 16)); + load128(ptr + 16)); CONFIRM_TEDDY(r_1, 8, 16, NOT_CAUTIOUS, do_confWithBit_teddy); } @@ -516,8 +518,8 @@ hwlm_error_t fdr_exec_teddy_msks3_pck(const struct FDR *fdr, m128 p_mask; m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 3); - m128 r_0 = prep_conf_teddy_m3(maskBase, &res_old_1, &res_old_2, - p_mask, val_0); + m128 r_0 = prep_conf_teddy_m3(maskBase, &res_old_1, &res_old_2, val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit_teddy); } @@ -551,14 +553,15 @@ hwlm_error_t fdr_exec_teddy_msks4(const struct FDR *fdr, m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 4); m128 r_0 = prep_conf_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, p_mask, val_0); + &res_old_3, val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBitMany_teddy); ptr += 16; } if (ptr + 16 < buf_end) { m128 r_0 = prep_conf_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, ones128(), load128(ptr)); + &res_old_3, load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBitMany_teddy); ptr += 16; } @@ -567,10 +570,10 @@ hwlm_error_t fdr_exec_teddy_msks4(const struct FDR *fdr, __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; m128 r_0 = prep_conf_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, ones128(), load128(ptr)); + &res_old_3, load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, NOT_CAUTIOUS, do_confWithBitMany_teddy); m128 r_1 = prep_conf_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, ones128(), load128(ptr + 16)); + &res_old_3, load128(ptr + 16)); CONFIRM_TEDDY(r_1, 8, 16, NOT_CAUTIOUS, do_confWithBitMany_teddy); } @@ -579,7 +582,8 @@ hwlm_error_t fdr_exec_teddy_msks4(const struct FDR *fdr, m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 4); m128 r_0 = prep_conf_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, p_mask, val_0); + &res_old_3, val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBitMany_teddy); } @@ -613,14 +617,15 @@ hwlm_error_t fdr_exec_teddy_msks4_pck(const struct FDR *fdr, m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 4); m128 r_0 = prep_conf_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, p_mask, val_0); + &res_old_3, val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } if (ptr + 16 < buf_end) { m128 r_0 = prep_conf_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, ones128(), load128(ptr)); + &res_old_3, load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } @@ -629,10 +634,10 @@ hwlm_error_t fdr_exec_teddy_msks4_pck(const struct FDR *fdr, __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; m128 r_0 = prep_conf_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, ones128(), load128(ptr)); + &res_old_3, load128(ptr)); CONFIRM_TEDDY(r_0, 8, 0, NOT_CAUTIOUS, do_confWithBit_teddy); m128 r_1 = prep_conf_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, ones128(), load128(ptr + 16)); + &res_old_3, load128(ptr + 16)); CONFIRM_TEDDY(r_1, 8, 16, NOT_CAUTIOUS, do_confWithBit_teddy); } @@ -641,7 +646,8 @@ hwlm_error_t fdr_exec_teddy_msks4_pck(const struct FDR *fdr, m128 val_0 = vectoredLoad128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 4); m128 r_0 = prep_conf_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, p_mask, val_0); + &res_old_3, val_0); + r_0 = and128(r_0, p_mask); CONFIRM_TEDDY(r_0, 8, 0, VECTORING, do_confWithBit_teddy); } diff --git a/src/fdr/teddy_avx2.c b/src/fdr/teddy_avx2.c index 428c9446..e4a836d4 100644 --- a/src/fdr/teddy_avx2.c +++ b/src/fdr/teddy_avx2.c @@ -407,36 +407,35 @@ void bit_array_fast_teddy(m128 var, u16 *bitArr, u32 *arrCnt, u32 offset) { } static really_inline -m256 prep_conf_fat_teddy_m1(const m256 *maskBase, m256 p_mask, m256 val) { +m256 prep_conf_fat_teddy_m1(const m256 *maskBase, m256 val) { m256 mask = set32x8(0xf); m256 lo = and256(val, mask); m256 hi = and256(rshift64_m256(val, 4), mask); - return and256(and256(vpshufb(maskBase[0*2], lo), - vpshufb(maskBase[0*2+1], hi)), p_mask); + return and256(vpshufb(maskBase[0*2], lo), + vpshufb(maskBase[0*2+1], hi)); } static really_inline -m256 prep_conf_fat_teddy_m2(const m256 *maskBase, m256 *old_1, m256 p_mask, - m256 val) { +m256 prep_conf_fat_teddy_m2(const m256 *maskBase, m256 *old_1, m256 val) { m256 mask = set32x8(0xf); m256 lo = and256(val, mask); m256 hi = and256(rshift64_m256(val, 4), mask); - m256 r = prep_conf_fat_teddy_m1(maskBase, p_mask, val); + m256 r = prep_conf_fat_teddy_m1(maskBase, val); m256 res_1 = and256(vpshufb(maskBase[1*2], lo), vpshufb(maskBase[1*2+1], hi)); m256 res_shifted_1 = vpalignr(res_1, *old_1, 16-1); *old_1 = res_1; - return and256(and256(r, p_mask), res_shifted_1); + return and256(r, res_shifted_1); } static really_inline m256 prep_conf_fat_teddy_m3(const m256 *maskBase, m256 *old_1, m256 *old_2, - m256 p_mask, m256 val) { + m256 val) { m256 mask = set32x8(0xf); m256 lo = and256(val, mask); m256 hi = and256(rshift64_m256(val, 4), mask); - m256 r = prep_conf_fat_teddy_m2(maskBase, old_1, p_mask, val); + m256 r = prep_conf_fat_teddy_m2(maskBase, old_1, val); m256 res_2 = and256(vpshufb(maskBase[2*2], lo), vpshufb(maskBase[2*2+1], hi)); @@ -447,11 +446,11 @@ m256 prep_conf_fat_teddy_m3(const m256 *maskBase, m256 *old_1, m256 *old_2, static really_inline m256 prep_conf_fat_teddy_m4(const m256 *maskBase, m256 *old_1, m256 *old_2, - m256 *old_3, m256 p_mask, m256 val) { + m256 *old_3, m256 val) { m256 mask = set32x8(0xf); m256 lo = and256(val, mask); m256 hi = and256(rshift64_m256(val, 4), mask); - m256 r = prep_conf_fat_teddy_m3(maskBase, old_1, old_2, p_mask, val); + m256 r = prep_conf_fat_teddy_m3(maskBase, old_1, old_2, val); m256 res_3 = and256(vpshufb(maskBase[3*2], lo), vpshufb(maskBase[3*2+1], hi)); @@ -461,12 +460,10 @@ m256 prep_conf_fat_teddy_m4(const m256 *maskBase, m256 *old_1, m256 *old_2, } static really_inline -m256 prep_conf_fast_teddy_m1(m256 val, m256 mask, m256 maskLo, m256 maskHi, - m256 p_mask) { +m256 prep_conf_fast_teddy_m1(m256 val, m256 mask, m256 maskLo, m256 maskHi) { m256 lo = and256(val, mask); m256 hi = and256(rshift64_m256(val, 4), mask); - m256 res = and256(vpshufb(maskLo, lo), vpshufb(maskHi, hi)); - return and256(res, p_mask); + return and256(vpshufb(maskLo, lo), vpshufb(maskHi, hi)); } static really_inline @@ -503,13 +500,14 @@ hwlm_error_t fdr_exec_teddy_avx2_msks1_fat(const struct FDR *fdr, m256 p_mask; m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 1); - m256 r_0 = prep_conf_fat_teddy_m1(maskBase, p_mask, val_0); + m256 r_0 = prep_conf_fat_teddy_m1(maskBase, val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit1_teddy); ptr += 16; } if (ptr + 16 < buf_end) { - m256 r_0 = prep_conf_fat_teddy_m1(maskBase, ones256(), load2x128(ptr)); + m256 r_0 = prep_conf_fat_teddy_m1(maskBase, load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit1_teddy); ptr += 16; } @@ -517,10 +515,9 @@ hwlm_error_t fdr_exec_teddy_avx2_msks1_fat(const struct FDR *fdr, for ( ; ptr + iterBytes <= buf_end; ptr += iterBytes) { __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; - m256 r_0 = prep_conf_fat_teddy_m1(maskBase, ones256(), load2x128(ptr)); + m256 r_0 = prep_conf_fat_teddy_m1(maskBase, load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, NOT_CAUTIOUS, do_confWithBit1_teddy); - m256 r_1 = prep_conf_fat_teddy_m1(maskBase, ones256(), - load2x128(ptr + 16)); + m256 r_1 = prep_conf_fat_teddy_m1(maskBase, load2x128(ptr + 16)); CONFIRM_FAT_TEDDY(r_1, 16, 16, NOT_CAUTIOUS, do_confWithBit1_teddy); } @@ -528,7 +525,8 @@ hwlm_error_t fdr_exec_teddy_avx2_msks1_fat(const struct FDR *fdr, m256 p_mask; m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 1); - m256 r_0 = prep_conf_fat_teddy_m1(maskBase, p_mask, val_0); + m256 r_0 = prep_conf_fat_teddy_m1(maskBase, val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit1_teddy); } @@ -558,13 +556,14 @@ hwlm_error_t fdr_exec_teddy_avx2_msks1_pck_fat(const struct FDR *fdr, m256 p_mask; m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 1); - m256 r_0 = prep_conf_fat_teddy_m1(maskBase, p_mask, val_0); + m256 r_0 = prep_conf_fat_teddy_m1(maskBase, val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } if (ptr + 16 < buf_end) { - m256 r_0 = prep_conf_fat_teddy_m1(maskBase, ones256(), load2x128(ptr)); + m256 r_0 = prep_conf_fat_teddy_m1(maskBase, load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } @@ -572,10 +571,9 @@ hwlm_error_t fdr_exec_teddy_avx2_msks1_pck_fat(const struct FDR *fdr, for ( ; ptr + iterBytes <= buf_end; ptr += iterBytes) { __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; - m256 r_0 = prep_conf_fat_teddy_m1(maskBase, ones256(), load2x128(ptr)); + m256 r_0 = prep_conf_fat_teddy_m1(maskBase, load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, NOT_CAUTIOUS, do_confWithBit_teddy); - m256 r_1 = prep_conf_fat_teddy_m1(maskBase, ones256(), - load2x128(ptr + 16)); + m256 r_1 = prep_conf_fat_teddy_m1(maskBase, load2x128(ptr + 16)); CONFIRM_FAT_TEDDY(r_1, 16, 16, NOT_CAUTIOUS, do_confWithBit_teddy); } @@ -583,7 +581,8 @@ hwlm_error_t fdr_exec_teddy_avx2_msks1_pck_fat(const struct FDR *fdr, m256 p_mask; m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 1); - m256 r_0 = prep_conf_fat_teddy_m1(maskBase, p_mask, val_0); + m256 r_0 = prep_conf_fat_teddy_m1(maskBase, val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit_teddy); } @@ -614,14 +613,14 @@ hwlm_error_t fdr_exec_teddy_avx2_msks2_fat(const struct FDR *fdr, m256 p_mask; m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 2); - m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, p_mask, val_0); + m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBitMany_teddy); ptr += 16; } if (ptr + 16 < buf_end) { - m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, ones256(), - load2x128(ptr)); + m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBitMany_teddy); ptr += 16; } @@ -629,10 +628,9 @@ hwlm_error_t fdr_exec_teddy_avx2_msks2_fat(const struct FDR *fdr, for ( ; ptr + iterBytes <= buf_end; ptr += iterBytes) { __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; - m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, ones256(), - load2x128(ptr)); + m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, NOT_CAUTIOUS, do_confWithBitMany_teddy); - m256 r_1 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, ones256(), + m256 r_1 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, load2x128(ptr + 16)); CONFIRM_FAT_TEDDY(r_1, 16, 16, NOT_CAUTIOUS, do_confWithBitMany_teddy); } @@ -641,7 +639,8 @@ hwlm_error_t fdr_exec_teddy_avx2_msks2_fat(const struct FDR *fdr, m256 p_mask; m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 2); - m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, p_mask, val_0); + m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBitMany_teddy); } @@ -672,25 +671,24 @@ hwlm_error_t fdr_exec_teddy_avx2_msks2_pck_fat(const struct FDR *fdr, m256 p_mask; m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 2); - m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, p_mask, val_0); + m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } if (ptr + 16 < buf_end) { - m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, ones256(), - load2x128(ptr)); + m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit_teddy); - ptr += 16; + ptr += 16; } for ( ; ptr + iterBytes <= buf_end; ptr += iterBytes) { __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; - m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, ones256(), - load2x128(ptr)); + m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, NOT_CAUTIOUS, do_confWithBit_teddy); - m256 r_1 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, ones256(), + m256 r_1 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, load2x128(ptr + 16)); CONFIRM_FAT_TEDDY(r_1, 16, 16, NOT_CAUTIOUS, do_confWithBit_teddy); } @@ -699,7 +697,8 @@ hwlm_error_t fdr_exec_teddy_avx2_msks2_pck_fat(const struct FDR *fdr, m256 p_mask; m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 2); - m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, p_mask, val_0); + m256 r_0 = prep_conf_fat_teddy_m2(maskBase, &res_old_1, val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit_teddy); } @@ -732,14 +731,15 @@ hwlm_error_t fdr_exec_teddy_avx2_msks3_fat(const struct FDR *fdr, m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 3); m256 r_0 = prep_conf_fat_teddy_m3(maskBase, &res_old_1, &res_old_2, - p_mask, val_0); + val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBitMany_teddy); ptr += 16; } if (ptr + 16 < buf_end) { m256 r_0 = prep_conf_fat_teddy_m3(maskBase, &res_old_1, &res_old_2, - ones256(), load2x128(ptr)); + load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBitMany_teddy); ptr += 16; } @@ -748,10 +748,10 @@ hwlm_error_t fdr_exec_teddy_avx2_msks3_fat(const struct FDR *fdr, __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; m256 r_0 = prep_conf_fat_teddy_m3(maskBase, &res_old_1, &res_old_2, - ones256(), load2x128(ptr)); + load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, NOT_CAUTIOUS, do_confWithBitMany_teddy); m256 r_1 = prep_conf_fat_teddy_m3(maskBase, &res_old_1, &res_old_2, - ones256(), load2x128(ptr + 16)); + load2x128(ptr + 16)); CONFIRM_FAT_TEDDY(r_1, 16, 16, NOT_CAUTIOUS, do_confWithBitMany_teddy); } @@ -760,7 +760,8 @@ hwlm_error_t fdr_exec_teddy_avx2_msks3_fat(const struct FDR *fdr, m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 3); m256 r_0 = prep_conf_fat_teddy_m3(maskBase, &res_old_1, &res_old_2, - p_mask, val_0); + val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBitMany_teddy); } @@ -793,14 +794,15 @@ hwlm_error_t fdr_exec_teddy_avx2_msks3_pck_fat(const struct FDR *fdr, m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 3); m256 r_0 = prep_conf_fat_teddy_m3(maskBase, &res_old_1, &res_old_2, - p_mask, val_0); + val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } if (ptr + 16 < buf_end) { m256 r_0 = prep_conf_fat_teddy_m3(maskBase, &res_old_1, &res_old_2, - ones256(), load2x128(ptr)); + load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } @@ -809,10 +811,10 @@ hwlm_error_t fdr_exec_teddy_avx2_msks3_pck_fat(const struct FDR *fdr, __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; m256 r_0 = prep_conf_fat_teddy_m3(maskBase, &res_old_1, &res_old_2, - ones256(), load2x128(ptr)); + load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, NOT_CAUTIOUS, do_confWithBit_teddy); m256 r_1 = prep_conf_fat_teddy_m3(maskBase, &res_old_1, &res_old_2, - ones256(), load2x128(ptr + 16)); + load2x128(ptr + 16)); CONFIRM_FAT_TEDDY(r_1, 16, 16, NOT_CAUTIOUS, do_confWithBit_teddy); } @@ -821,7 +823,8 @@ hwlm_error_t fdr_exec_teddy_avx2_msks3_pck_fat(const struct FDR *fdr, m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 3); m256 r_0 = prep_conf_fat_teddy_m3(maskBase, &res_old_1, &res_old_2, - p_mask, val_0); + val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit_teddy); } @@ -855,15 +858,15 @@ hwlm_error_t fdr_exec_teddy_avx2_msks4_fat(const struct FDR *fdr, m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 4); m256 r_0 = prep_conf_fat_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, p_mask, val_0); + &res_old_3, val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBitMany_teddy); ptr += 16; } if (ptr + 16 < buf_end) { m256 r_0 = prep_conf_fat_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, ones256(), - load2x128(ptr)); + &res_old_3, load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBitMany_teddy); ptr += 16; } @@ -872,12 +875,10 @@ hwlm_error_t fdr_exec_teddy_avx2_msks4_fat(const struct FDR *fdr, __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; m256 r_0 = prep_conf_fat_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, ones256(), - load2x128(ptr)); + &res_old_3, load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, NOT_CAUTIOUS, do_confWithBitMany_teddy); m256 r_1 = prep_conf_fat_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, ones256(), - load2x128(ptr + 16)); + &res_old_3, load2x128(ptr + 16)); CONFIRM_FAT_TEDDY(r_1, 16, 16, NOT_CAUTIOUS, do_confWithBitMany_teddy); } @@ -886,7 +887,8 @@ hwlm_error_t fdr_exec_teddy_avx2_msks4_fat(const struct FDR *fdr, m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 4); m256 r_0 = prep_conf_fat_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, p_mask, val_0); + &res_old_3, val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBitMany_teddy); } @@ -920,15 +922,15 @@ hwlm_error_t fdr_exec_teddy_avx2_msks4_pck_fat(const struct FDR *fdr, m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 4); m256 r_0 = prep_conf_fat_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, p_mask, val_0); + &res_old_3, val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } if (ptr + 16 < buf_end) { m256 r_0 = prep_conf_fat_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, ones256(), - load2x128(ptr)); + &res_old_3, load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit_teddy); ptr += 16; } @@ -937,12 +939,10 @@ hwlm_error_t fdr_exec_teddy_avx2_msks4_pck_fat(const struct FDR *fdr, __builtin_prefetch(ptr + (iterBytes*4)); CHECK_FLOOD; m256 r_0 = prep_conf_fat_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, ones256(), - load2x128(ptr)); + &res_old_3, load2x128(ptr)); CONFIRM_FAT_TEDDY(r_0, 16, 0, NOT_CAUTIOUS, do_confWithBit_teddy); m256 r_1 = prep_conf_fat_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, ones256(), - load2x128(ptr + 16)); + &res_old_3, load2x128(ptr + 16)); CONFIRM_FAT_TEDDY(r_1, 16, 16, NOT_CAUTIOUS, do_confWithBit_teddy); } @@ -951,7 +951,8 @@ hwlm_error_t fdr_exec_teddy_avx2_msks4_pck_fat(const struct FDR *fdr, m256 val_0 = vectoredLoad2x128(&p_mask, ptr, a->buf, buf_end, a->buf_history, a->len_history, 4); m256 r_0 = prep_conf_fat_teddy_m4(maskBase, &res_old_1, &res_old_2, - &res_old_3, p_mask, val_0); + &res_old_3, val_0); + r_0 = and256(r_0, p_mask); CONFIRM_FAT_TEDDY(r_0, 16, 0, VECTORING, do_confWithBit_teddy); } @@ -986,16 +987,15 @@ hwlm_error_t fdr_exec_teddy_avx2_msks1_fast(const struct FDR *fdr, m256 p_mask; m256 val_0 = vectoredLoad256(&p_mask, ptr, a->buf + a->start_offset, buf_end, a->buf_history, a->len_history); - m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi, - p_mask); + m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi); + res_0 = and256(res_0, p_mask); CONFIRM_FAST_TEDDY(res_0, 0, VECTORING, do_confWithBit1_fast_teddy); ptr += 32; } if (ptr + 32 < buf_end) { m256 val_0 = load256(ptr + 0); - m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi, - ones256()); + m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi); CONFIRM_FAST_TEDDY(res_0, 0, VECTORING, do_confWithBit1_fast_teddy); ptr += 32; } @@ -1005,13 +1005,11 @@ hwlm_error_t fdr_exec_teddy_avx2_msks1_fast(const struct FDR *fdr, CHECK_FLOOD; m256 val_0 = load256(ptr + 0); - m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi, - ones256()); + m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi); CONFIRM_FAST_TEDDY(res_0, 0, NOT_CAUTIOUS, do_confWithBit1_fast_teddy); m256 val_1 = load256(ptr + 32); - m256 res_1 = prep_conf_fast_teddy_m1(val_1, mask, maskLo, maskHi, - ones256()); + m256 res_1 = prep_conf_fast_teddy_m1(val_1, mask, maskLo, maskHi); CONFIRM_FAST_TEDDY(res_1, 4, NOT_CAUTIOUS, do_confWithBit1_fast_teddy); } @@ -1019,8 +1017,8 @@ hwlm_error_t fdr_exec_teddy_avx2_msks1_fast(const struct FDR *fdr, m256 p_mask; m256 val_0 = vectoredLoad256(&p_mask, ptr, a->buf + a->start_offset, buf_end, a->buf_history, a->len_history); - m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi, - p_mask); + m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi); + res_0 = and256(res_0, p_mask); CONFIRM_FAST_TEDDY(res_0, 0, VECTORING, do_confWithBit1_fast_teddy); } @@ -1055,16 +1053,15 @@ hwlm_error_t fdr_exec_teddy_avx2_msks1_pck_fast(const struct FDR *fdr, m256 p_mask; m256 val_0 = vectoredLoad256(&p_mask, ptr, a->buf + a->start_offset, buf_end, a->buf_history, a->len_history); - m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi, - p_mask); + m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi); + res_0 = and256(res_0, p_mask); CONFIRM_FAST_TEDDY(res_0, 0, VECTORING, do_confWithBit_fast_teddy); ptr += 32; } if (ptr + 32 < buf_end) { m256 val_0 = load256(ptr + 0); - m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi, - ones256()); + m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi); CONFIRM_FAST_TEDDY(res_0, 0, VECTORING, do_confWithBit_fast_teddy); ptr += 32; } @@ -1074,13 +1071,11 @@ hwlm_error_t fdr_exec_teddy_avx2_msks1_pck_fast(const struct FDR *fdr, CHECK_FLOOD; m256 val_0 = load256(ptr + 0); - m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi, - ones256()); + m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi); CONFIRM_FAST_TEDDY(res_0, 0, NOT_CAUTIOUS, do_confWithBit_fast_teddy); m256 val_1 = load256(ptr + 32); - m256 res_1 = prep_conf_fast_teddy_m1(val_1, mask, maskLo, maskHi, - ones256()); + m256 res_1 = prep_conf_fast_teddy_m1(val_1, mask, maskLo, maskHi); CONFIRM_FAST_TEDDY(res_1, 4, NOT_CAUTIOUS, do_confWithBit_fast_teddy); } @@ -1088,8 +1083,8 @@ hwlm_error_t fdr_exec_teddy_avx2_msks1_pck_fast(const struct FDR *fdr, m256 p_mask; m256 val_0 = vectoredLoad256(&p_mask, ptr, a->buf + a->start_offset, buf_end, a->buf_history, a->len_history); - m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi, - p_mask); + m256 res_0 = prep_conf_fast_teddy_m1(val_0, mask, maskLo, maskHi); + res_0 = and256(res_0, p_mask); CONFIRM_FAST_TEDDY(res_0, 0, VECTORING, do_confWithBit_fast_teddy); }