[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PULL 22/25] target/riscv: rvv: Add tail agnostic for vector permutation
From: |
Alistair Francis |
Subject: |
[PULL 22/25] target/riscv: rvv: Add tail agnostic for vector permutation instructions |
Date: |
Fri, 10 Jun 2022 14:26:52 +1000 |
From: eopXD <yueh.ting.chen@gmail.com>
Signed-off-by: eop Chen <eop.chen@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Weiwei Li <liweiwei@iscas.ac.cn>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <165449614532.19704.7000832880482980398-15@git.sr.ht>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
---
target/riscv/vector_helper.c | 40 +++++++++++++++++++++++++
target/riscv/insn_trans/trans_rvv.c.inc | 7 +++--
2 files changed, 45 insertions(+), 2 deletions(-)
diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c
index 75b59cf917..a96fc49c71 100644
--- a/target/riscv/vector_helper.c
+++ b/target/riscv/vector_helper.c
@@ -4931,6 +4931,9 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1,
void *vs2, \
{ \
uint32_t vm = vext_vm(desc); \
uint32_t vl = env->vl; \
+ uint32_t esz = sizeof(ETYPE); \
+ uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
+ uint32_t vta = vext_vta(desc); \
target_ulong offset = s1, i_min, i; \
\
i_min = MAX(env->vstart, offset); \
@@ -4940,6 +4943,8 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1,
void *vs2, \
} \
*((ETYPE *)vd + H(i)) = *((ETYPE *)vs2 + H(i - offset)); \
} \
+ /* set tail elements to 1s */ \
+ vext_set_elems_1s(vd, vta, vl * esz, total_elems * esz); \
}
/* vslideup.vx vd, vs2, rs1, vm # vd[i+rs1] = vs2[i] */
@@ -4955,6 +4960,9 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1,
void *vs2, \
uint32_t vlmax = vext_max_elems(desc, ctzl(sizeof(ETYPE))); \
uint32_t vm = vext_vm(desc); \
uint32_t vl = env->vl; \
+ uint32_t esz = sizeof(ETYPE); \
+ uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
+ uint32_t vta = vext_vta(desc); \
target_ulong i_max, i; \
\
i_max = MAX(MIN(s1 < vlmax ? vlmax - s1 : 0, vl), env->vstart); \
@@ -4971,6 +4979,8 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1,
void *vs2, \
} \
\
env->vstart = 0; \
+ /* set tail elements to 1s */ \
+ vext_set_elems_1s(vd, vta, vl * esz, total_elems * esz); \
}
/* vslidedown.vx vd, vs2, rs1, vm # vd[i] = vs2[i+rs1] */
@@ -4986,6 +4996,9 @@ static void vslide1up_##BITWIDTH(void *vd, void *v0,
target_ulong s1, \
typedef uint##BITWIDTH##_t ETYPE; \
uint32_t vm = vext_vm(desc); \
uint32_t vl = env->vl; \
+ uint32_t esz = sizeof(ETYPE); \
+ uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
+ uint32_t vta = vext_vta(desc); \
uint32_t i; \
\
for (i = env->vstart; i < vl; i++) { \
@@ -4999,6 +5012,8 @@ static void vslide1up_##BITWIDTH(void *vd, void *v0,
target_ulong s1, \
} \
} \
env->vstart = 0; \
+ /* set tail elements to 1s */ \
+ vext_set_elems_1s(vd, vta, vl * esz, total_elems * esz); \
}
GEN_VEXT_VSLIE1UP(8, H1)
@@ -5026,6 +5041,9 @@ static void vslide1down_##BITWIDTH(void *vd, void *v0,
target_ulong s1, \
typedef uint##BITWIDTH##_t ETYPE; \
uint32_t vm = vext_vm(desc); \
uint32_t vl = env->vl; \
+ uint32_t esz = sizeof(ETYPE); \
+ uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
+ uint32_t vta = vext_vta(desc); \
uint32_t i; \
\
for (i = env->vstart; i < vl; i++) { \
@@ -5039,6 +5057,8 @@ static void vslide1down_##BITWIDTH(void *vd, void *v0,
target_ulong s1, \
} \
} \
env->vstart = 0; \
+ /* set tail elements to 1s */ \
+ vext_set_elems_1s(vd, vta, vl * esz, total_elems * esz); \
}
GEN_VEXT_VSLIDE1DOWN(8, H1)
@@ -5092,6 +5112,9 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, void
*vs2, \
uint32_t vlmax = vext_max_elems(desc, ctzl(sizeof(TS2))); \
uint32_t vm = vext_vm(desc); \
uint32_t vl = env->vl; \
+ uint32_t esz = sizeof(TS2); \
+ uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
+ uint32_t vta = vext_vta(desc); \
uint64_t index; \
uint32_t i; \
\
@@ -5107,6 +5130,8 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, void
*vs2, \
} \
} \
env->vstart = 0; \
+ /* set tail elements to 1s */ \
+ vext_set_elems_1s(vd, vta, vl * esz, total_elems * esz); \
}
/* vd[i] = (vs1[i] >= VLMAX) ? 0 : vs2[vs1[i]]; */
@@ -5127,6 +5152,9 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1,
void *vs2, \
uint32_t vlmax = vext_max_elems(desc, ctzl(sizeof(ETYPE))); \
uint32_t vm = vext_vm(desc); \
uint32_t vl = env->vl; \
+ uint32_t esz = sizeof(ETYPE); \
+ uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
+ uint32_t vta = vext_vta(desc); \
uint64_t index = s1; \
uint32_t i; \
\
@@ -5141,6 +5169,8 @@ void HELPER(NAME)(void *vd, void *v0, target_ulong s1,
void *vs2, \
} \
} \
env->vstart = 0; \
+ /* set tail elements to 1s */ \
+ vext_set_elems_1s(vd, vta, vl * esz, total_elems * esz); \
}
/* vd[i] = (x[rs1] >= VLMAX) ? 0 : vs2[rs1] */
@@ -5155,6 +5185,9 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, void
*vs2, \
CPURISCVState *env, uint32_t desc) \
{ \
uint32_t vl = env->vl; \
+ uint32_t esz = sizeof(ETYPE); \
+ uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
+ uint32_t vta = vext_vta(desc); \
uint32_t num = 0, i; \
\
for (i = env->vstart; i < vl; i++) { \
@@ -5165,6 +5198,8 @@ void HELPER(NAME)(void *vd, void *v0, void *vs1, void
*vs2, \
num++; \
} \
env->vstart = 0; \
+ /* set tail elements to 1s */ \
+ vext_set_elems_1s(vd, vta, vl * esz, total_elems * esz); \
}
/* Compress into vd elements of vs2 where vs1 is enabled */
@@ -5196,6 +5231,9 @@ void HELPER(NAME)(void *vd, void *v0, void *vs2,
\
{ \
uint32_t vl = env->vl; \
uint32_t vm = vext_vm(desc); \
+ uint32_t esz = sizeof(ETYPE); \
+ uint32_t total_elems = vext_get_total_elems(env, desc, esz); \
+ uint32_t vta = vext_vta(desc); \
uint32_t i; \
\
for (i = env->vstart; i < vl; i++) { \
@@ -5205,6 +5243,8 @@ void HELPER(NAME)(void *vd, void *v0, void *vs2,
\
*((ETYPE *)vd + HD(i)) = *((DTYPE *)vs2 + HS1(i)); \
} \
env->vstart = 0; \
+ /* set tail elements to 1s */ \
+ vext_set_elems_1s(vd, vta, vl * esz, total_elems * esz); \
}
GEN_VEXT_INT_EXT(vzext_vf2_h, uint16_t, uint8_t, H2, H1)
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc
b/target/riscv/insn_trans/trans_rvv.c.inc
index a94e634a6b..4f84d4878a 100644
--- a/target/riscv/insn_trans/trans_rvv.c.inc
+++ b/target/riscv/insn_trans/trans_rvv.c.inc
@@ -3669,7 +3669,7 @@ static bool trans_vrgather_vx(DisasContext *s, arg_rmrr
*a)
return false;
}
- if (a->vm && s->vl_eq_vlmax) {
+ if (a->vm && s->vl_eq_vlmax && !(s->vta && s->lmul < 0)) {
int scale = s->lmul - (s->sew + 3);
int vlmax = s->cfg_ptr->vlen >> -scale;
TCGv_i64 dest = tcg_temp_new_i64();
@@ -3701,7 +3701,7 @@ static bool trans_vrgather_vi(DisasContext *s, arg_rmrr
*a)
return false;
}
- if (a->vm && s->vl_eq_vlmax) {
+ if (a->vm && s->vl_eq_vlmax && !(s->vta && s->lmul < 0)) {
int scale = s->lmul - (s->sew + 3);
int vlmax = s->cfg_ptr->vlen >> -scale;
if (a->rs1 >= vlmax) {
@@ -3753,6 +3753,7 @@ static bool trans_vcompress_vm(DisasContext *s, arg_r *a)
tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_vl, 0, over);
data = FIELD_DP32(data, VDATA, LMUL, s->lmul);
+ data = FIELD_DP32(data, VDATA, VTA, s->vta);
tcg_gen_gvec_4_ptr(vreg_ofs(s, a->rd), vreg_ofs(s, 0),
vreg_ofs(s, a->rs1), vreg_ofs(s, a->rs2),
cpu_env, s->cfg_ptr->vlen / 8,
@@ -3853,6 +3854,8 @@ static bool int_ext_op(DisasContext *s, arg_rmr *a,
uint8_t seq)
}
data = FIELD_DP32(data, VDATA, VM, a->vm);
+ data = FIELD_DP32(data, VDATA, LMUL, s->lmul);
+ data = FIELD_DP32(data, VDATA, VTA, s->vta);
tcg_gen_gvec_3_ptr(vreg_ofs(s, a->rd), vreg_ofs(s, 0),
vreg_ofs(s, a->rs2), cpu_env,
--
2.36.1
- [PULL 10/25] target/riscv: rvv: Rename ambiguous esz, (continued)
- [PULL 10/25] target/riscv: rvv: Rename ambiguous esz, Alistair Francis, 2022/06/10
- [PULL 16/25] target/riscv: rvv: Add tail agnostic for vector integer comparison instructions, Alistair Francis, 2022/06/10
- [PULL 14/25] target/riscv: rvv: Add tail agnostic for vx, vvm, vxm instructions, Alistair Francis, 2022/06/10
- [PULL 15/25] target/riscv: rvv: Add tail agnostic for vector integer shift instructions, Alistair Francis, 2022/06/10
- [PULL 18/25] target/riscv: rvv: Add tail agnostic for vector fix-point arithmetic instructions, Alistair Francis, 2022/06/10
- [PULL 19/25] target/riscv: rvv: Add tail agnostic for vector floating-point instructions, Alistair Francis, 2022/06/10
- [PULL 20/25] target/riscv: rvv: Add tail agnostic for vector reduction instructions, Alistair Francis, 2022/06/10
- [PULL 21/25] target/riscv: rvv: Add tail agnostic for vector mask instructions, Alistair Francis, 2022/06/10
- [PULL 25/25] target/riscv: trans_rvv: Avoid assert for RV32 and e64, Alistair Francis, 2022/06/10
- [PULL 24/25] target/riscv: Don't expose the CPU properties on names CPUs, Alistair Francis, 2022/06/10
- [PULL 22/25] target/riscv: rvv: Add tail agnostic for vector permutation instructions,
Alistair Francis <=
- [PULL 23/25] target/riscv: rvv: Add option 'rvv_ta_all_1s' to enable optional tail agnostic behavior, Alistair Francis, 2022/06/10
- Re: [PULL 00/25] riscv-to-apply queue, Richard Henderson, 2022/06/10