RISC-V: Support VLS modes mask operations
Checks
Commit Message
This patch support mask operations (comparison and logical).
This patch reduce these FAILs of "vect" testsuite:
FAIL: gcc.dg/vect/vect-bic-bitmask-12.c -flto -ffat-lto-objects scan-tree-dump dce7 "<=\\s*.+{ 255,.+}"
FAIL: gcc.dg/vect/vect-bic-bitmask-12.c scan-tree-dump dce7 "<=\\s*.+{ 255,.+}"
FAIL: gcc.dg/vect/vect-bic-bitmask-23.c -flto -ffat-lto-objects scan-tree-dump dce7 "<=\\s*.+{ 255, 15, 1, 65535 }"
FAIL: gcc.dg/vect/vect-bic-bitmask-23.c scan-tree-dump dce7 "<=\\s*.+{ 255, 15, 1, 65535 }"
Full regression passed (with reducing 4 FAILs).
gcc/ChangeLog:
* config/riscv/autovec-opt.md: Add VLS mask modes.
* config/riscv/autovec.md (@vcond_mask_<mode><vm>): Remove @.
(vcond_mask_<mode><vm>): Add VLS mask modes.
* config/riscv/vector.md: Ditto.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/autovec/vls/def.h: Add VLS tests.
* gcc.target/riscv/rvv/autovec/vls/cmp-1.c: New test.
* gcc.target/riscv/rvv/autovec/vls/cmp-2.c: New test.
* gcc.target/riscv/rvv/autovec/vls/cmp-3.c: New test.
* gcc.target/riscv/rvv/autovec/vls/cmp-4.c: New test.
* gcc.target/riscv/rvv/autovec/vls/cmp-5.c: New test.
* gcc.target/riscv/rvv/autovec/vls/cmp-6.c: New test.
* gcc.target/riscv/rvv/autovec/vls/mask-1.c: New test.
* gcc.target/riscv/rvv/autovec/vls/mask-2.c: New test.
* gcc.target/riscv/rvv/autovec/vls/mask-3.c: New test.
---
gcc/config/riscv/autovec-opt.md | 18 +--
gcc/config/riscv/autovec.md | 32 +++---
gcc/config/riscv/vector.md | 60 +++++-----
.../gcc.target/riscv/rvv/autovec/vls/cmp-1.c | 106 ++++++++++++++++++
.../gcc.target/riscv/rvv/autovec/vls/cmp-2.c | 106 ++++++++++++++++++
.../gcc.target/riscv/rvv/autovec/vls/cmp-3.c | 106 ++++++++++++++++++
.../gcc.target/riscv/rvv/autovec/vls/cmp-4.c | 106 ++++++++++++++++++
.../gcc.target/riscv/rvv/autovec/vls/cmp-5.c | 106 ++++++++++++++++++
.../gcc.target/riscv/rvv/autovec/vls/cmp-6.c | 106 ++++++++++++++++++
.../gcc.target/riscv/rvv/autovec/vls/def.h | 9 ++
.../gcc.target/riscv/rvv/autovec/vls/mask-1.c | 69 ++++++++++++
.../gcc.target/riscv/rvv/autovec/vls/mask-2.c | 69 ++++++++++++
.../gcc.target/riscv/rvv/autovec/vls/mask-3.c | 69 ++++++++++++
13 files changed, 907 insertions(+), 55 deletions(-)
create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-1.c
create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-2.c
create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-3.c
create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-4.c
create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-5.c
create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-6.c
create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-1.c
create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-2.c
create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-3.c
Comments
LGTM
Juzhe-Zhong <juzhe.zhong@rivai.ai> 於 2023年9月14日 週四 20:44 寫道:
> This patch support mask operations (comparison and logical).
>
> This patch reduce these FAILs of "vect" testsuite:
> FAIL: gcc.dg/vect/vect-bic-bitmask-12.c -flto -ffat-lto-objects
> scan-tree-dump dce7 "<=\\s*.+{ 255,.+}"
> FAIL: gcc.dg/vect/vect-bic-bitmask-12.c scan-tree-dump dce7 "<=\\s*.+{
> 255,.+}"
> FAIL: gcc.dg/vect/vect-bic-bitmask-23.c -flto -ffat-lto-objects
> scan-tree-dump dce7 "<=\\s*.+{ 255, 15, 1, 65535 }"
> FAIL: gcc.dg/vect/vect-bic-bitmask-23.c scan-tree-dump dce7 "<=\\s*.+{
> 255, 15, 1, 65535 }"
>
> Full regression passed (with reducing 4 FAILs).
>
> gcc/ChangeLog:
>
> * config/riscv/autovec-opt.md: Add VLS mask modes.
> * config/riscv/autovec.md (@vcond_mask_<mode><vm>): Remove @.
> (vcond_mask_<mode><vm>): Add VLS mask modes.
> * config/riscv/vector.md: Ditto.
>
> gcc/testsuite/ChangeLog:
>
> * gcc.target/riscv/rvv/autovec/vls/def.h: Add VLS tests.
> * gcc.target/riscv/rvv/autovec/vls/cmp-1.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/cmp-2.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/cmp-3.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/cmp-4.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/cmp-5.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/cmp-6.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/mask-1.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/mask-2.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/mask-3.c: New test.
>
> ---
> gcc/config/riscv/autovec-opt.md | 18 +--
> gcc/config/riscv/autovec.md | 32 +++---
> gcc/config/riscv/vector.md | 60 +++++-----
> .../gcc.target/riscv/rvv/autovec/vls/cmp-1.c | 106 ++++++++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/cmp-2.c | 106 ++++++++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/cmp-3.c | 106 ++++++++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/cmp-4.c | 106 ++++++++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/cmp-5.c | 106 ++++++++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/cmp-6.c | 106 ++++++++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/def.h | 9 ++
> .../gcc.target/riscv/rvv/autovec/vls/mask-1.c | 69 ++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/mask-2.c | 69 ++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/mask-3.c | 69 ++++++++++++
> 13 files changed, 907 insertions(+), 55 deletions(-)
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-1.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-2.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-3.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-4.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-5.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-6.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-1.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-2.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-3.c
>
> diff --git a/gcc/config/riscv/autovec-opt.md
> b/gcc/config/riscv/autovec-opt.md
> index e26c01856ff..22ab8afc994 100644
> --- a/gcc/config/riscv/autovec-opt.md
> +++ b/gcc/config/riscv/autovec-opt.md
> @@ -67,10 +67,10 @@
> ;;
> -------------------------------------------------------------------------
>
> (define_insn_and_split "*<optab>not<mode>"
> - [(set (match_operand:VB 0 "register_operand" "=vr")
> - (bitmanip_bitwise:VB
> - (not:VB (match_operand:VB 2 "register_operand" " vr"))
> - (match_operand:VB 1 "register_operand" " vr")))]
> + [(set (match_operand:VB_VLS 0 "register_operand" "=vr")
> + (bitmanip_bitwise:VB_VLS
> + (not:VB_VLS (match_operand:VB_VLS 2 "register_operand" " vr"))
> + (match_operand:VB_VLS 1 "register_operand" " vr")))]
> "TARGET_VECTOR && can_create_pseudo_p ()"
> "#"
> "&& 1"
> @@ -93,11 +93,11 @@
> ;;
> -------------------------------------------------------------------------
>
> (define_insn_and_split "*n<optab><mode>"
> - [(set (match_operand:VB 0 "register_operand" "=vr")
> - (not:VB
> - (any_bitwise:VB
> - (match_operand:VB 1 "register_operand" " vr")
> - (match_operand:VB 2 "register_operand" " vr"))))]
> + [(set (match_operand:VB_VLS 0 "register_operand" "=vr")
> + (not:VB_VLS
> + (any_bitwise:VB_VLS
> + (match_operand:VB_VLS 1 "register_operand" " vr")
> + (match_operand:VB_VLS 2 "register_operand" " vr"))))]
> "TARGET_VECTOR && can_create_pseudo_p ()"
> "#"
> "&& 1"
> diff --git a/gcc/config/riscv/autovec.md b/gcc/config/riscv/autovec.md
> index 7121bab1716..d2002a8ee26 100644
> --- a/gcc/config/riscv/autovec.md
> +++ b/gcc/config/riscv/autovec.md
> @@ -495,9 +495,9 @@
> ;;
> -------------------------------------------------------------------------
>
> (define_insn_and_split "<optab><mode>3"
> - [(set (match_operand:VB 0 "register_operand" "=vr")
> - (any_bitwise:VB (match_operand:VB 1 "register_operand" " vr")
> - (match_operand:VB 2 "register_operand" " vr")))]
> + [(set (match_operand:VB_VLS 0 "register_operand" "=vr")
> + (any_bitwise:VB_VLS (match_operand:VB_VLS 1 "register_operand" "
> vr")
> + (match_operand:VB_VLS 2 "register_operand" "
> vr")))]
> "TARGET_VECTOR && can_create_pseudo_p ()"
> "#"
> "&& 1"
> @@ -518,8 +518,8 @@
> ;;
> -------------------------------------------------------------------------
>
> (define_insn_and_split "one_cmpl<mode>2"
> - [(set (match_operand:VB 0 "register_operand" "=vr")
> - (not:VB (match_operand:VB 1 "register_operand" " vr")))]
> + [(set (match_operand:VB_VLS 0 "register_operand" "=vr")
> + (not:VB_VLS (match_operand:VB_VLS 1 "register_operand" " vr")))]
> "TARGET_VECTOR && can_create_pseudo_p ()"
> "#"
> "&& 1"
> @@ -545,12 +545,12 @@
> ;; - vfmerge.vf
> ;;
> -------------------------------------------------------------------------
>
> -(define_insn_and_split "@vcond_mask_<mode><vm>"
> - [(set (match_operand:V 0 "register_operand")
> - (if_then_else:V
> +(define_insn_and_split "vcond_mask_<mode><vm>"
> + [(set (match_operand:V_VLS 0 "register_operand")
> + (if_then_else:V_VLS
> (match_operand:<VM> 3 "register_operand")
> - (match_operand:V 1 "nonmemory_operand")
> - (match_operand:V 2 "register_operand")))]
> + (match_operand:V_VLS 1 "nonmemory_operand")
> + (match_operand:V_VLS 2 "register_operand")))]
> "TARGET_VECTOR && can_create_pseudo_p ()"
> "#"
> "&& 1"
> @@ -609,8 +609,8 @@
> (define_expand "vec_cmp<mode><vm>"
> [(set (match_operand:<VM> 0 "register_operand")
> (match_operator:<VM> 1 "comparison_operator"
> - [(match_operand:VI 2 "register_operand")
> - (match_operand:VI 3 "register_operand")]))]
> + [(match_operand:V_VLSI 2 "register_operand")
> + (match_operand:V_VLSI 3 "register_operand")]))]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_vec_cmp (operands[0], GET_CODE (operands[1]),
> @@ -622,8 +622,8 @@
> (define_expand "vec_cmpu<mode><vm>"
> [(set (match_operand:<VM> 0 "register_operand")
> (match_operator:<VM> 1 "comparison_operator"
> - [(match_operand:VI 2 "register_operand")
> - (match_operand:VI 3 "register_operand")]))]
> + [(match_operand:V_VLSI 2 "register_operand")
> + (match_operand:V_VLSI 3 "register_operand")]))]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_vec_cmp (operands[0], GET_CODE (operands[1]),
> @@ -635,8 +635,8 @@
> (define_expand "vec_cmp<mode><vm>"
> [(set (match_operand:<VM> 0 "register_operand")
> (match_operator:<VM> 1 "comparison_operator"
> - [(match_operand:VF 2 "register_operand")
> - (match_operand:VF 3 "register_operand")]))]
> + [(match_operand:V_VLSF 2 "register_operand")
> + (match_operand:V_VLSF 3 "register_operand")]))]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_vec_cmp_float (operands[0], GET_CODE
> (operands[1]),
> diff --git a/gcc/config/riscv/vector.md b/gcc/config/riscv/vector.md
> index 4630af6cbff..8c0e86576a0 100644
> --- a/gcc/config/riscv/vector.md
> +++ b/gcc/config/riscv/vector.md
> @@ -7177,8 +7177,8 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "signed_order_operator"
> - [(match_operand:VF 4 "register_operand")
> - (match_operand:VF 5 "register_operand")])
> + [(match_operand:V_VLSF 4 "register_operand")
> + (match_operand:V_VLSF 5 "register_operand")])
> (match_operand:<VM> 2 "vector_merge_operand")))]
> "TARGET_VECTOR"
> {})
> @@ -7195,10 +7195,10 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "signed_order_operator"
> - [(match_operand:VF 4 "register_operand" " vr,
> vr")
> - (match_operand:VF 5 "register_operand" " vr,
> vr")])
> + [(match_operand:V_VLSF 4 "register_operand" " vr,
> vr")
> + (match_operand:V_VLSF 5 "register_operand" " vr,
> vr")])
> (match_operand:<VM> 2 "vector_merge_operand" " vu,
> 0")))]
> - "TARGET_VECTOR && known_le (GET_MODE_SIZE (<MODE>mode),
> BYTES_PER_RISCV_VECTOR)"
> + "TARGET_VECTOR && riscv_vector::cmp_lmul_le_one (<MODE>mode)"
> "vmf%B3.vv\t%0,%4,%5%p1"
> [(set_attr "type" "vfcmp")
> (set_attr "mode" "<MODE>")])
> @@ -7214,8 +7214,8 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 2 "signed_order_operator"
> - [(match_operand:VF 3 "register_operand" " vr")
> - (match_operand:VF 4 "register_operand" " vr")])
> + [(match_operand:V_VLSF 3 "register_operand" " vr")
> + (match_operand:V_VLSF 4 "register_operand" " vr")])
> (match_dup 1)))]
> "TARGET_VECTOR"
> "vmf%B2.vv\t%0,%3,%4,v0.t"
> @@ -7238,10 +7238,10 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "signed_order_operator"
> - [(match_operand:VF 4 "register_operand" " vr,
> 0, vr, 0, 0, vr, 0, vr, vr")
> - (match_operand:VF 5 "register_operand" " vr,
> vr, 0, 0, vr, 0, 0, vr, vr")])
> + [(match_operand:V_VLSF 4 "register_operand" " vr,
> 0, vr, 0, 0, vr, 0, vr, vr")
> + (match_operand:V_VLSF 5 "register_operand" " vr,
> vr, 0, 0, vr, 0, 0, vr, vr")])
> (match_operand:<VM> 2 "vector_merge_operand" " vu,
> vu, vu, vu, 0, 0, 0, vu, 0")))]
> - "TARGET_VECTOR && known_gt (GET_MODE_SIZE (<MODE>mode),
> BYTES_PER_RISCV_VECTOR)"
> + "TARGET_VECTOR && riscv_vector::cmp_lmul_gt_one (<MODE>mode)"
> "vmf%B3.vv\t%0,%4,%5%p1"
> [(set_attr "type" "vfcmp")
> (set_attr "mode" "<MODE>")])
> @@ -7257,8 +7257,8 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "signed_order_operator"
> - [(match_operand:VF 4 "register_operand")
> - (vec_duplicate:VF
> + [(match_operand:V_VLSF 4 "register_operand")
> + (vec_duplicate:V_VLSF
> (match_operand:<VEL> 5 "register_operand"))])
> (match_operand:<VM> 2 "vector_merge_operand")))]
> "TARGET_VECTOR"
> @@ -7275,8 +7275,8 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 2 "signed_order_operator"
> - [(match_operand:VF 3 "register_operand" " vr")
> - (vec_duplicate:VF
> + [(match_operand:V_VLSF 3 "register_operand" " vr")
> + (vec_duplicate:V_VLSF
> (match_operand:<VEL> 4 "register_operand" " f"))])
> (match_dup 1)))]
> "TARGET_VECTOR"
> @@ -7300,11 +7300,11 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "signed_order_operator"
> - [(match_operand:VF 4 "register_operand" " vr,
> vr")
> - (vec_duplicate:VF
> + [(match_operand:V_VLSF 4 "register_operand" " vr,
> vr")
> + (vec_duplicate:V_VLSF
> (match_operand:<VEL> 5 "register_operand" " f,
> f"))])
> (match_operand:<VM> 2 "vector_merge_operand" " vu,
> 0")))]
> - "TARGET_VECTOR && known_le (GET_MODE_SIZE (<MODE>mode),
> BYTES_PER_RISCV_VECTOR)"
> + "TARGET_VECTOR && riscv_vector::cmp_lmul_le_one (<MODE>mode)"
> "vmf%B3.vf\t%0,%4,%5%p1"
> [(set_attr "type" "vfcmp")
> (set_attr "mode" "<MODE>")])
> @@ -7321,11 +7321,11 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "signed_order_operator"
> - [(match_operand:VF 4 "register_operand" " vr,
> 0, 0, vr, vr")
> - (vec_duplicate:VF
> + [(match_operand:V_VLSF 4 "register_operand" " vr,
> 0, 0, vr, vr")
> + (vec_duplicate:V_VLSF
> (match_operand:<VEL> 5 "register_operand" " f,
> f, f, f, f"))])
> (match_operand:<VM> 2 "vector_merge_operand" " vu,
> vu, 0, vu, 0")))]
> - "TARGET_VECTOR && known_gt (GET_MODE_SIZE (<MODE>mode),
> BYTES_PER_RISCV_VECTOR)"
> + "TARGET_VECTOR && riscv_vector::cmp_lmul_gt_one (<MODE>mode)"
> "vmf%B3.vf\t%0,%4,%5%p1"
> [(set_attr "type" "vfcmp")
> (set_attr "mode" "<MODE>")])
> @@ -7341,9 +7341,9 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "equality_operator"
> - [(vec_duplicate:VF
> + [(vec_duplicate:V_VLSF
> (match_operand:<VEL> 5 "register_operand"))
> - (match_operand:VF 4 "register_operand")])
> + (match_operand:V_VLSF 4 "register_operand")])
> (match_operand:<VM> 2 "vector_merge_operand")))]
> "TARGET_VECTOR"
> {})
> @@ -7359,9 +7359,9 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 2 "equality_operator"
> - [(vec_duplicate:VF
> + [(vec_duplicate:V_VLSF
> (match_operand:<VEL> 4 "register_operand" " f"))
> - (match_operand:VF 3 "register_operand" " vr")])
> + (match_operand:V_VLSF 3 "register_operand" " vr")])
> (match_dup 1)))]
> "TARGET_VECTOR"
> "vmf%B2.vf\t%0,%3,%4,v0.t"
> @@ -7384,11 +7384,11 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "equality_operator"
> - [(vec_duplicate:VF
> + [(vec_duplicate:V_VLSF
> (match_operand:<VEL> 5 "register_operand" " f,
> f"))
> - (match_operand:VF 4 "register_operand" " vr,
> vr")])
> + (match_operand:V_VLSF 4 "register_operand" " vr,
> vr")])
> (match_operand:<VM> 2 "vector_merge_operand" " vu,
> 0")))]
> - "TARGET_VECTOR && known_le (GET_MODE_SIZE (<MODE>mode),
> BYTES_PER_RISCV_VECTOR)"
> + "TARGET_VECTOR && riscv_vector::cmp_lmul_le_one (<MODE>mode)"
> "vmf%B3.vf\t%0,%4,%5%p1"
> [(set_attr "type" "vfcmp")
> (set_attr "mode" "<MODE>")])
> @@ -7405,11 +7405,11 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "equality_operator"
> - [(vec_duplicate:VF
> + [(vec_duplicate:V_VLSF
> (match_operand:<VEL> 5 "register_operand" " f,
> f, f, f, f"))
> - (match_operand:VF 4 "register_operand" " vr,
> 0, 0, vr, vr")])
> + (match_operand:V_VLSF 4 "register_operand" " vr,
> 0, 0, vr, vr")])
> (match_operand:<VM> 2 "vector_merge_operand" " vu,
> vu, 0, vu, 0")))]
> - "TARGET_VECTOR && known_gt (GET_MODE_SIZE (<MODE>mode),
> BYTES_PER_RISCV_VECTOR)"
> + "TARGET_VECTOR && riscv_vector::cmp_lmul_gt_one (<MODE>mode)"
> "vmf%B3.vf\t%0,%4,%5%p1"
> [(set_attr "type" "vfcmp")
> (set_attr "mode" "<MODE>")])
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-1.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-1.c
> new file mode 100644
> index 00000000000..05742b90fd5
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-1.c
> @@ -0,0 +1,106 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_OP_VV (eq, 1, int8_t, ==)
> +DEF_OP_VV (eq, 2, int8_t, ==)
> +DEF_OP_VV (eq, 4, int8_t, ==)
> +DEF_OP_VV (eq, 8, int8_t, ==)
> +DEF_OP_VV (eq, 16, int8_t, ==)
> +DEF_OP_VV (eq, 32, int8_t, ==)
> +DEF_OP_VV (eq, 64, int8_t, ==)
> +DEF_OP_VV (eq, 128, int8_t, ==)
> +DEF_OP_VV (eq, 256, int8_t, ==)
> +DEF_OP_VV (eq, 512, int8_t, ==)
> +DEF_OP_VV (eq, 1024, int8_t, ==)
> +DEF_OP_VV (eq, 2048, int8_t, ==)
> +DEF_OP_VV (eq, 4096, int8_t, ==)
> +
> +DEF_OP_VV (eq, 1, int16_t, ==)
> +DEF_OP_VV (eq, 2, int16_t, ==)
> +DEF_OP_VV (eq, 4, int16_t, ==)
> +DEF_OP_VV (eq, 8, int16_t, ==)
> +DEF_OP_VV (eq, 16, int16_t, ==)
> +DEF_OP_VV (eq, 32, int16_t, ==)
> +DEF_OP_VV (eq, 64, int16_t, ==)
> +DEF_OP_VV (eq, 128, int16_t, ==)
> +DEF_OP_VV (eq, 256, int16_t, ==)
> +DEF_OP_VV (eq, 512, int16_t, ==)
> +DEF_OP_VV (eq, 1024, int16_t, ==)
> +DEF_OP_VV (eq, 2048, int16_t, ==)
> +
> +DEF_OP_VV (eq, 1, int32_t, ==)
> +DEF_OP_VV (eq, 2, int32_t, ==)
> +DEF_OP_VV (eq, 4, int32_t, ==)
> +DEF_OP_VV (eq, 8, int32_t, ==)
> +DEF_OP_VV (eq, 16, int32_t, ==)
> +DEF_OP_VV (eq, 32, int32_t, ==)
> +DEF_OP_VV (eq, 64, int32_t, ==)
> +DEF_OP_VV (eq, 128, int32_t, ==)
> +DEF_OP_VV (eq, 256, int32_t, ==)
> +DEF_OP_VV (eq, 512, int32_t, ==)
> +DEF_OP_VV (eq, 1024, int32_t, ==)
> +
> +DEF_OP_VV (eq, 1, int64_t, ==)
> +DEF_OP_VV (eq, 2, int64_t, ==)
> +DEF_OP_VV (eq, 4, int64_t, ==)
> +DEF_OP_VV (eq, 8, int64_t, ==)
> +DEF_OP_VV (eq, 16, int64_t, ==)
> +DEF_OP_VV (eq, 32, int64_t, ==)
> +DEF_OP_VV (eq, 64, int64_t, ==)
> +DEF_OP_VV (eq, 128, int64_t, ==)
> +DEF_OP_VV (eq, 256, int64_t, ==)
> +DEF_OP_VV (eq, 512, int64_t, ==)
> +
> +DEF_OP_VV (eq, 1, _Float16, ==)
> +DEF_OP_VV (eq, 2, _Float16, ==)
> +DEF_OP_VV (eq, 4, _Float16, ==)
> +DEF_OP_VV (eq, 8, _Float16, ==)
> +DEF_OP_VV (eq, 16, _Float16, ==)
> +DEF_OP_VV (eq, 32, _Float16, ==)
> +DEF_OP_VV (eq, 64, _Float16, ==)
> +DEF_OP_VV (eq, 128, _Float16, ==)
> +DEF_OP_VV (eq, 256, _Float16, ==)
> +DEF_OP_VV (eq, 512, _Float16, ==)
> +DEF_OP_VV (eq, 1024, _Float16, ==)
> +DEF_OP_VV (eq, 2048, _Float16, ==)
> +
> +DEF_OP_VV (eq, 1, float, ==)
> +DEF_OP_VV (eq, 2, float, ==)
> +DEF_OP_VV (eq, 4, float, ==)
> +DEF_OP_VV (eq, 8, float, ==)
> +DEF_OP_VV (eq, 16, float, ==)
> +DEF_OP_VV (eq, 32, float, ==)
> +DEF_OP_VV (eq, 64, float, ==)
> +DEF_OP_VV (eq, 128, float, ==)
> +DEF_OP_VV (eq, 256, float, ==)
> +DEF_OP_VV (eq, 512, float, ==)
> +DEF_OP_VV (eq, 1024, float, ==)
> +
> +DEF_OP_VV (eq, 1, double, ==)
> +DEF_OP_VV (eq, 2, double, ==)
> +DEF_OP_VV (eq, 4, double, ==)
> +DEF_OP_VV (eq, 8, double, ==)
> +DEF_OP_VV (eq, 16, double, ==)
> +DEF_OP_VV (eq, 32, double, ==)
> +DEF_OP_VV (eq, 64, double, ==)
> +DEF_OP_VV (eq, 128, double, ==)
> +DEF_OP_VV (eq, 256, double, ==)
> +DEF_OP_VV (eq, 512, double, ==)
> +
> +/* { dg-final { scan-assembler-times {vmseq\.vv} 42 } } */
> +/* { dg-final { scan-assembler-times {vmfeq\.vv} 27 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-2.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-2.c
> new file mode 100644
> index 00000000000..39a56025818
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-2.c
> @@ -0,0 +1,106 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_OP_VV (ne, 1, int8_t, !=)
> +DEF_OP_VV (ne, 2, int8_t, !=)
> +DEF_OP_VV (ne, 4, int8_t, !=)
> +DEF_OP_VV (ne, 8, int8_t, !=)
> +DEF_OP_VV (ne, 16, int8_t, !=)
> +DEF_OP_VV (ne, 32, int8_t, !=)
> +DEF_OP_VV (ne, 64, int8_t, !=)
> +DEF_OP_VV (ne, 128, int8_t, !=)
> +DEF_OP_VV (ne, 256, int8_t, !=)
> +DEF_OP_VV (ne, 512, int8_t, !=)
> +DEF_OP_VV (ne, 1024, int8_t, !=)
> +DEF_OP_VV (ne, 2048, int8_t, !=)
> +DEF_OP_VV (ne, 4096, int8_t, !=)
> +
> +DEF_OP_VV (ne, 1, int16_t, !=)
> +DEF_OP_VV (ne, 2, int16_t, !=)
> +DEF_OP_VV (ne, 4, int16_t, !=)
> +DEF_OP_VV (ne, 8, int16_t, !=)
> +DEF_OP_VV (ne, 16, int16_t, !=)
> +DEF_OP_VV (ne, 32, int16_t, !=)
> +DEF_OP_VV (ne, 64, int16_t, !=)
> +DEF_OP_VV (ne, 128, int16_t, !=)
> +DEF_OP_VV (ne, 256, int16_t, !=)
> +DEF_OP_VV (ne, 512, int16_t, !=)
> +DEF_OP_VV (ne, 1024, int16_t, !=)
> +DEF_OP_VV (ne, 2048, int16_t, !=)
> +
> +DEF_OP_VV (ne, 1, int32_t, !=)
> +DEF_OP_VV (ne, 2, int32_t, !=)
> +DEF_OP_VV (ne, 4, int32_t, !=)
> +DEF_OP_VV (ne, 8, int32_t, !=)
> +DEF_OP_VV (ne, 16, int32_t, !=)
> +DEF_OP_VV (ne, 32, int32_t, !=)
> +DEF_OP_VV (ne, 64, int32_t, !=)
> +DEF_OP_VV (ne, 128, int32_t, !=)
> +DEF_OP_VV (ne, 256, int32_t, !=)
> +DEF_OP_VV (ne, 512, int32_t, !=)
> +DEF_OP_VV (ne, 1024, int32_t, !=)
> +
> +DEF_OP_VV (ne, 1, int64_t, !=)
> +DEF_OP_VV (ne, 2, int64_t, !=)
> +DEF_OP_VV (ne, 4, int64_t, !=)
> +DEF_OP_VV (ne, 8, int64_t, !=)
> +DEF_OP_VV (ne, 16, int64_t, !=)
> +DEF_OP_VV (ne, 32, int64_t, !=)
> +DEF_OP_VV (ne, 64, int64_t, !=)
> +DEF_OP_VV (ne, 128, int64_t, !=)
> +DEF_OP_VV (ne, 256, int64_t, !=)
> +DEF_OP_VV (ne, 512, int64_t, !=)
> +
> +DEF_OP_VV (ne, 1, _Float16, !=)
> +DEF_OP_VV (ne, 2, _Float16, !=)
> +DEF_OP_VV (ne, 4, _Float16, !=)
> +DEF_OP_VV (ne, 8, _Float16, !=)
> +DEF_OP_VV (ne, 16, _Float16, !=)
> +DEF_OP_VV (ne, 32, _Float16, !=)
> +DEF_OP_VV (ne, 64, _Float16, !=)
> +DEF_OP_VV (ne, 128, _Float16, !=)
> +DEF_OP_VV (ne, 256, _Float16, !=)
> +DEF_OP_VV (ne, 512, _Float16, !=)
> +DEF_OP_VV (ne, 1024, _Float16, !=)
> +DEF_OP_VV (ne, 2048, _Float16, !=)
> +
> +DEF_OP_VV (ne, 1, float, !=)
> +DEF_OP_VV (ne, 2, float, !=)
> +DEF_OP_VV (ne, 4, float, !=)
> +DEF_OP_VV (ne, 8, float, !=)
> +DEF_OP_VV (ne, 16, float, !=)
> +DEF_OP_VV (ne, 32, float, !=)
> +DEF_OP_VV (ne, 64, float, !=)
> +DEF_OP_VV (ne, 128, float, !=)
> +DEF_OP_VV (ne, 256, float, !=)
> +DEF_OP_VV (ne, 512, float, !=)
> +DEF_OP_VV (ne, 1024, float, !=)
> +
> +DEF_OP_VV (ne, 1, double, !=)
> +DEF_OP_VV (ne, 2, double, !=)
> +DEF_OP_VV (ne, 4, double, !=)
> +DEF_OP_VV (ne, 8, double, !=)
> +DEF_OP_VV (ne, 16, double, !=)
> +DEF_OP_VV (ne, 32, double, !=)
> +DEF_OP_VV (ne, 64, double, !=)
> +DEF_OP_VV (ne, 128, double, !=)
> +DEF_OP_VV (ne, 256, double, !=)
> +DEF_OP_VV (ne, 512, double, !=)
> +
> +/* { dg-final { scan-assembler-times {vmsne\.vv} 42 } } */
> +/* { dg-final { scan-assembler-times {vmfne\.vv} 27 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-3.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-3.c
> new file mode 100644
> index 00000000000..387157d9be6
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-3.c
> @@ -0,0 +1,106 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_OP_VV (gt, 1, int8_t, >)
> +DEF_OP_VV (gt, 2, int8_t, >)
> +DEF_OP_VV (gt, 4, int8_t, >)
> +DEF_OP_VV (gt, 8, int8_t, >)
> +DEF_OP_VV (gt, 16, int8_t, >)
> +DEF_OP_VV (gt, 32, int8_t, >)
> +DEF_OP_VV (gt, 64, int8_t, >)
> +DEF_OP_VV (gt, 128, int8_t, >)
> +DEF_OP_VV (gt, 256, int8_t, >)
> +DEF_OP_VV (gt, 512, int8_t, >)
> +DEF_OP_VV (gt, 1024, int8_t, >)
> +DEF_OP_VV (gt, 2048, int8_t, >)
> +DEF_OP_VV (gt, 4096, int8_t, >)
> +
> +DEF_OP_VV (gt, 1, int16_t, >)
> +DEF_OP_VV (gt, 2, int16_t, >)
> +DEF_OP_VV (gt, 4, int16_t, >)
> +DEF_OP_VV (gt, 8, int16_t, >)
> +DEF_OP_VV (gt, 16, int16_t, >)
> +DEF_OP_VV (gt, 32, int16_t, >)
> +DEF_OP_VV (gt, 64, int16_t, >)
> +DEF_OP_VV (gt, 128, int16_t, >)
> +DEF_OP_VV (gt, 256, int16_t, >)
> +DEF_OP_VV (gt, 512, int16_t, >)
> +DEF_OP_VV (gt, 1024, int16_t, >)
> +DEF_OP_VV (gt, 2048, int16_t, >)
> +
> +DEF_OP_VV (gt, 1, int32_t, >)
> +DEF_OP_VV (gt, 2, int32_t, >)
> +DEF_OP_VV (gt, 4, int32_t, >)
> +DEF_OP_VV (gt, 8, int32_t, >)
> +DEF_OP_VV (gt, 16, int32_t, >)
> +DEF_OP_VV (gt, 32, int32_t, >)
> +DEF_OP_VV (gt, 64, int32_t, >)
> +DEF_OP_VV (gt, 128, int32_t, >)
> +DEF_OP_VV (gt, 256, int32_t, >)
> +DEF_OP_VV (gt, 512, int32_t, >)
> +DEF_OP_VV (gt, 1024, int32_t, >)
> +
> +DEF_OP_VV (gt, 1, int64_t, >)
> +DEF_OP_VV (gt, 2, int64_t, >)
> +DEF_OP_VV (gt, 4, int64_t, >)
> +DEF_OP_VV (gt, 8, int64_t, >)
> +DEF_OP_VV (gt, 16, int64_t, >)
> +DEF_OP_VV (gt, 32, int64_t, >)
> +DEF_OP_VV (gt, 64, int64_t, >)
> +DEF_OP_VV (gt, 128, int64_t, >)
> +DEF_OP_VV (gt, 256, int64_t, >)
> +DEF_OP_VV (gt, 512, int64_t, >)
> +
> +DEF_OP_VV (gt, 1, _Float16, >)
> +DEF_OP_VV (gt, 2, _Float16, >)
> +DEF_OP_VV (gt, 4, _Float16, >)
> +DEF_OP_VV (gt, 8, _Float16, >)
> +DEF_OP_VV (gt, 16, _Float16, >)
> +DEF_OP_VV (gt, 32, _Float16, >)
> +DEF_OP_VV (gt, 64, _Float16, >)
> +DEF_OP_VV (gt, 128, _Float16, >)
> +DEF_OP_VV (gt, 256, _Float16, >)
> +DEF_OP_VV (gt, 512, _Float16, >)
> +DEF_OP_VV (gt, 1024, _Float16, >)
> +DEF_OP_VV (gt, 2048, _Float16, >)
> +
> +DEF_OP_VV (gt, 1, float, >)
> +DEF_OP_VV (gt, 2, float, >)
> +DEF_OP_VV (gt, 4, float, >)
> +DEF_OP_VV (gt, 8, float, >)
> +DEF_OP_VV (gt, 16, float, >)
> +DEF_OP_VV (gt, 32, float, >)
> +DEF_OP_VV (gt, 64, float, >)
> +DEF_OP_VV (gt, 128, float, >)
> +DEF_OP_VV (gt, 256, float, >)
> +DEF_OP_VV (gt, 512, float, >)
> +DEF_OP_VV (gt, 1024, float, >)
> +
> +DEF_OP_VV (gt, 1, double, >)
> +DEF_OP_VV (gt, 2, double, >)
> +DEF_OP_VV (gt, 4, double, >)
> +DEF_OP_VV (gt, 8, double, >)
> +DEF_OP_VV (gt, 16, double, >)
> +DEF_OP_VV (gt, 32, double, >)
> +DEF_OP_VV (gt, 64, double, >)
> +DEF_OP_VV (gt, 128, double, >)
> +DEF_OP_VV (gt, 256, double, >)
> +DEF_OP_VV (gt, 512, double, >)
> +
> +/* { dg-final { scan-assembler-times {vmsgt\.vv} 40 } } */
> +/* { dg-final { scan-assembler-times {vmfgt\.vv} 27 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-4.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-4.c
> new file mode 100644
> index 00000000000..40b8871ea3a
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-4.c
> @@ -0,0 +1,106 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_OP_VV (ge, 1, int8_t, >=)
> +DEF_OP_VV (ge, 2, int8_t, >=)
> +DEF_OP_VV (ge, 4, int8_t, >=)
> +DEF_OP_VV (ge, 8, int8_t, >=)
> +DEF_OP_VV (ge, 16, int8_t, >=)
> +DEF_OP_VV (ge, 32, int8_t, >=)
> +DEF_OP_VV (ge, 64, int8_t, >=)
> +DEF_OP_VV (ge, 128, int8_t, >=)
> +DEF_OP_VV (ge, 256, int8_t, >=)
> +DEF_OP_VV (ge, 512, int8_t, >=)
> +DEF_OP_VV (ge, 1024, int8_t, >=)
> +DEF_OP_VV (ge, 2048, int8_t, >=)
> +DEF_OP_VV (ge, 4096, int8_t, >=)
> +
> +DEF_OP_VV (ge, 1, int16_t, >=)
> +DEF_OP_VV (ge, 2, int16_t, >=)
> +DEF_OP_VV (ge, 4, int16_t, >=)
> +DEF_OP_VV (ge, 8, int16_t, >=)
> +DEF_OP_VV (ge, 16, int16_t, >=)
> +DEF_OP_VV (ge, 32, int16_t, >=)
> +DEF_OP_VV (ge, 64, int16_t, >=)
> +DEF_OP_VV (ge, 128, int16_t, >=)
> +DEF_OP_VV (ge, 256, int16_t, >=)
> +DEF_OP_VV (ge, 512, int16_t, >=)
> +DEF_OP_VV (ge, 1024, int16_t, >=)
> +DEF_OP_VV (ge, 2048, int16_t, >=)
> +
> +DEF_OP_VV (ge, 1, int32_t, >=)
> +DEF_OP_VV (ge, 2, int32_t, >=)
> +DEF_OP_VV (ge, 4, int32_t, >=)
> +DEF_OP_VV (ge, 8, int32_t, >=)
> +DEF_OP_VV (ge, 16, int32_t, >=)
> +DEF_OP_VV (ge, 32, int32_t, >=)
> +DEF_OP_VV (ge, 64, int32_t, >=)
> +DEF_OP_VV (ge, 128, int32_t, >=)
> +DEF_OP_VV (ge, 256, int32_t, >=)
> +DEF_OP_VV (ge, 512, int32_t, >=)
> +DEF_OP_VV (ge, 1024, int32_t, >=)
> +
> +DEF_OP_VV (ge, 1, int64_t, >=)
> +DEF_OP_VV (ge, 2, int64_t, >=)
> +DEF_OP_VV (ge, 4, int64_t, >=)
> +DEF_OP_VV (ge, 8, int64_t, >=)
> +DEF_OP_VV (ge, 16, int64_t, >=)
> +DEF_OP_VV (ge, 32, int64_t, >=)
> +DEF_OP_VV (ge, 64, int64_t, >=)
> +DEF_OP_VV (ge, 128, int64_t, >=)
> +DEF_OP_VV (ge, 256, int64_t, >=)
> +DEF_OP_VV (ge, 512, int64_t, >=)
> +
> +DEF_OP_VV (ge, 1, _Float16, >=)
> +DEF_OP_VV (ge, 2, _Float16, >=)
> +DEF_OP_VV (ge, 4, _Float16, >=)
> +DEF_OP_VV (ge, 8, _Float16, >=)
> +DEF_OP_VV (ge, 16, _Float16, >=)
> +DEF_OP_VV (ge, 32, _Float16, >=)
> +DEF_OP_VV (ge, 64, _Float16, >=)
> +DEF_OP_VV (ge, 128, _Float16, >=)
> +DEF_OP_VV (ge, 256, _Float16, >=)
> +DEF_OP_VV (ge, 512, _Float16, >=)
> +DEF_OP_VV (ge, 1024, _Float16, >=)
> +DEF_OP_VV (ge, 2048, _Float16, >=)
> +
> +DEF_OP_VV (ge, 1, float, >=)
> +DEF_OP_VV (ge, 2, float, >=)
> +DEF_OP_VV (ge, 4, float, >=)
> +DEF_OP_VV (ge, 8, float, >=)
> +DEF_OP_VV (ge, 16, float, >=)
> +DEF_OP_VV (ge, 32, float, >=)
> +DEF_OP_VV (ge, 64, float, >=)
> +DEF_OP_VV (ge, 128, float, >=)
> +DEF_OP_VV (ge, 256, float, >=)
> +DEF_OP_VV (ge, 512, float, >=)
> +DEF_OP_VV (ge, 1024, float, >=)
> +
> +DEF_OP_VV (ge, 1, double, >=)
> +DEF_OP_VV (ge, 2, double, >=)
> +DEF_OP_VV (ge, 4, double, >=)
> +DEF_OP_VV (ge, 8, double, >=)
> +DEF_OP_VV (ge, 16, double, >=)
> +DEF_OP_VV (ge, 32, double, >=)
> +DEF_OP_VV (ge, 64, double, >=)
> +DEF_OP_VV (ge, 128, double, >=)
> +DEF_OP_VV (ge, 256, double, >=)
> +DEF_OP_VV (ge, 512, double, >=)
> +
> +/* { dg-final { scan-assembler-times {vmsge\.vv} 40 } } */
> +/* { dg-final { scan-assembler-times {vmfge\.vv} 27 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-5.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-5.c
> new file mode 100644
> index 00000000000..378b704d360
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-5.c
> @@ -0,0 +1,106 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_OP_VV (lt, 1, int8_t, <)
> +DEF_OP_VV (lt, 2, int8_t, <)
> +DEF_OP_VV (lt, 4, int8_t, <)
> +DEF_OP_VV (lt, 8, int8_t, <)
> +DEF_OP_VV (lt, 16, int8_t, <)
> +DEF_OP_VV (lt, 32, int8_t, <)
> +DEF_OP_VV (lt, 64, int8_t, <)
> +DEF_OP_VV (lt, 128, int8_t, <)
> +DEF_OP_VV (lt, 256, int8_t, <)
> +DEF_OP_VV (lt, 512, int8_t, <)
> +DEF_OP_VV (lt, 1024, int8_t, <)
> +DEF_OP_VV (lt, 2048, int8_t, <)
> +DEF_OP_VV (lt, 4096, int8_t, <)
> +
> +DEF_OP_VV (lt, 1, int16_t, <)
> +DEF_OP_VV (lt, 2, int16_t, <)
> +DEF_OP_VV (lt, 4, int16_t, <)
> +DEF_OP_VV (lt, 8, int16_t, <)
> +DEF_OP_VV (lt, 16, int16_t, <)
> +DEF_OP_VV (lt, 32, int16_t, <)
> +DEF_OP_VV (lt, 64, int16_t, <)
> +DEF_OP_VV (lt, 128, int16_t, <)
> +DEF_OP_VV (lt, 256, int16_t, <)
> +DEF_OP_VV (lt, 512, int16_t, <)
> +DEF_OP_VV (lt, 1024, int16_t, <)
> +DEF_OP_VV (lt, 2048, int16_t, <)
> +
> +DEF_OP_VV (lt, 1, int32_t, <)
> +DEF_OP_VV (lt, 2, int32_t, <)
> +DEF_OP_VV (lt, 4, int32_t, <)
> +DEF_OP_VV (lt, 8, int32_t, <)
> +DEF_OP_VV (lt, 16, int32_t, <)
> +DEF_OP_VV (lt, 32, int32_t, <)
> +DEF_OP_VV (lt, 64, int32_t, <)
> +DEF_OP_VV (lt, 128, int32_t, <)
> +DEF_OP_VV (lt, 256, int32_t, <)
> +DEF_OP_VV (lt, 512, int32_t, <)
> +DEF_OP_VV (lt, 1024, int32_t, <)
> +
> +DEF_OP_VV (lt, 1, int64_t, <)
> +DEF_OP_VV (lt, 2, int64_t, <)
> +DEF_OP_VV (lt, 4, int64_t, <)
> +DEF_OP_VV (lt, 8, int64_t, <)
> +DEF_OP_VV (lt, 16, int64_t, <)
> +DEF_OP_VV (lt, 32, int64_t, <)
> +DEF_OP_VV (lt, 64, int64_t, <)
> +DEF_OP_VV (lt, 128, int64_t, <)
> +DEF_OP_VV (lt, 256, int64_t, <)
> +DEF_OP_VV (lt, 512, int64_t, <)
> +
> +DEF_OP_VV (lt, 1, _Float16, <)
> +DEF_OP_VV (lt, 2, _Float16, <)
> +DEF_OP_VV (lt, 4, _Float16, <)
> +DEF_OP_VV (lt, 8, _Float16, <)
> +DEF_OP_VV (lt, 16, _Float16, <)
> +DEF_OP_VV (lt, 32, _Float16, <)
> +DEF_OP_VV (lt, 64, _Float16, <)
> +DEF_OP_VV (lt, 128, _Float16, <)
> +DEF_OP_VV (lt, 256, _Float16, <)
> +DEF_OP_VV (lt, 512, _Float16, <)
> +DEF_OP_VV (lt, 1024, _Float16, <)
> +DEF_OP_VV (lt, 2048, _Float16, <)
> +
> +DEF_OP_VV (lt, 1, float, <)
> +DEF_OP_VV (lt, 2, float, <)
> +DEF_OP_VV (lt, 4, float, <)
> +DEF_OP_VV (lt, 8, float, <)
> +DEF_OP_VV (lt, 16, float, <)
> +DEF_OP_VV (lt, 32, float, <)
> +DEF_OP_VV (lt, 64, float, <)
> +DEF_OP_VV (lt, 128, float, <)
> +DEF_OP_VV (lt, 256, float, <)
> +DEF_OP_VV (lt, 512, float, <)
> +DEF_OP_VV (lt, 1024, float, <)
> +
> +DEF_OP_VV (lt, 1, double, <)
> +DEF_OP_VV (lt, 2, double, <)
> +DEF_OP_VV (lt, 4, double, <)
> +DEF_OP_VV (lt, 8, double, <)
> +DEF_OP_VV (lt, 16, double, <)
> +DEF_OP_VV (lt, 32, double, <)
> +DEF_OP_VV (lt, 64, double, <)
> +DEF_OP_VV (lt, 128, double, <)
> +DEF_OP_VV (lt, 256, double, <)
> +DEF_OP_VV (lt, 512, double, <)
> +
> +/* { dg-final { scan-assembler-times {vmslt\.vv} 40 } } */
> +/* { dg-final { scan-assembler-times {vmflt\.vv} 27 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-6.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-6.c
> new file mode 100644
> index 00000000000..f0351e0baf8
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-6.c
> @@ -0,0 +1,106 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_OP_VV (le, 1, int8_t, <=)
> +DEF_OP_VV (le, 2, int8_t, <=)
> +DEF_OP_VV (le, 4, int8_t, <=)
> +DEF_OP_VV (le, 8, int8_t, <=)
> +DEF_OP_VV (le, 16, int8_t, <=)
> +DEF_OP_VV (le, 32, int8_t, <=)
> +DEF_OP_VV (le, 64, int8_t, <=)
> +DEF_OP_VV (le, 128, int8_t, <=)
> +DEF_OP_VV (le, 256, int8_t, <=)
> +DEF_OP_VV (le, 512, int8_t, <=)
> +DEF_OP_VV (le, 1024, int8_t, <=)
> +DEF_OP_VV (le, 2048, int8_t, <=)
> +DEF_OP_VV (le, 4096, int8_t, <=)
> +
> +DEF_OP_VV (le, 1, int16_t, <=)
> +DEF_OP_VV (le, 2, int16_t, <=)
> +DEF_OP_VV (le, 4, int16_t, <=)
> +DEF_OP_VV (le, 8, int16_t, <=)
> +DEF_OP_VV (le, 16, int16_t, <=)
> +DEF_OP_VV (le, 32, int16_t, <=)
> +DEF_OP_VV (le, 64, int16_t, <=)
> +DEF_OP_VV (le, 128, int16_t, <=)
> +DEF_OP_VV (le, 256, int16_t, <=)
> +DEF_OP_VV (le, 512, int16_t, <=)
> +DEF_OP_VV (le, 1024, int16_t, <=)
> +DEF_OP_VV (le, 2048, int16_t, <=)
> +
> +DEF_OP_VV (le, 1, int32_t, <=)
> +DEF_OP_VV (le, 2, int32_t, <=)
> +DEF_OP_VV (le, 4, int32_t, <=)
> +DEF_OP_VV (le, 8, int32_t, <=)
> +DEF_OP_VV (le, 16, int32_t, <=)
> +DEF_OP_VV (le, 32, int32_t, <=)
> +DEF_OP_VV (le, 64, int32_t, <=)
> +DEF_OP_VV (le, 128, int32_t, <=)
> +DEF_OP_VV (le, 256, int32_t, <=)
> +DEF_OP_VV (le, 512, int32_t, <=)
> +DEF_OP_VV (le, 1024, int32_t, <=)
> +
> +DEF_OP_VV (le, 1, int64_t, <=)
> +DEF_OP_VV (le, 2, int64_t, <=)
> +DEF_OP_VV (le, 4, int64_t, <=)
> +DEF_OP_VV (le, 8, int64_t, <=)
> +DEF_OP_VV (le, 16, int64_t, <=)
> +DEF_OP_VV (le, 32, int64_t, <=)
> +DEF_OP_VV (le, 64, int64_t, <=)
> +DEF_OP_VV (le, 128, int64_t, <=)
> +DEF_OP_VV (le, 256, int64_t, <=)
> +DEF_OP_VV (le, 512, int64_t, <=)
> +
> +DEF_OP_VV (le, 1, _Float16, <=)
> +DEF_OP_VV (le, 2, _Float16, <=)
> +DEF_OP_VV (le, 4, _Float16, <=)
> +DEF_OP_VV (le, 8, _Float16, <=)
> +DEF_OP_VV (le, 16, _Float16, <=)
> +DEF_OP_VV (le, 32, _Float16, <=)
> +DEF_OP_VV (le, 64, _Float16, <=)
> +DEF_OP_VV (le, 128, _Float16, <=)
> +DEF_OP_VV (le, 256, _Float16, <=)
> +DEF_OP_VV (le, 512, _Float16, <=)
> +DEF_OP_VV (le, 1024, _Float16, <=)
> +DEF_OP_VV (le, 2048, _Float16, <=)
> +
> +DEF_OP_VV (le, 1, float, <=)
> +DEF_OP_VV (le, 2, float, <=)
> +DEF_OP_VV (le, 4, float, <=)
> +DEF_OP_VV (le, 8, float, <=)
> +DEF_OP_VV (le, 16, float, <=)
> +DEF_OP_VV (le, 32, float, <=)
> +DEF_OP_VV (le, 64, float, <=)
> +DEF_OP_VV (le, 128, float, <=)
> +DEF_OP_VV (le, 256, float, <=)
> +DEF_OP_VV (le, 512, float, <=)
> +DEF_OP_VV (le, 1024, float, <=)
> +
> +DEF_OP_VV (le, 1, double, <=)
> +DEF_OP_VV (le, 2, double, <=)
> +DEF_OP_VV (le, 4, double, <=)
> +DEF_OP_VV (le, 8, double, <=)
> +DEF_OP_VV (le, 16, double, <=)
> +DEF_OP_VV (le, 32, double, <=)
> +DEF_OP_VV (le, 64, double, <=)
> +DEF_OP_VV (le, 128, double, <=)
> +DEF_OP_VV (le, 256, double, <=)
> +DEF_OP_VV (le, 512, double, <=)
> +
> +/* { dg-final { scan-assembler-times {vmsle\.vv} 40 } } */
> +/* { dg-final { scan-assembler-times {vmfle\.vv} 27 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/def.h
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/def.h
> index 79b4fbc6d93..1edc1910920 100644
> --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/def.h
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/def.h
> @@ -249,3 +249,12 @@ typedef double v512df __attribute__ ((vector_size
> (4096)));
> {
> \
> return v[INDEX];
> \
> }
> +
> +#define DEF_MASK_LOGIC(PREFIX, NUM, TYPE, OP)
> \
> + void __attribute__ ((noinline, noclone))
> \
> + PREFIX##_##TYPE##NUM (TYPE *restrict a, TYPE *restrict b, TYPE
> *restrict c, \
> + TYPE *restrict d, TYPE *restrict e)
> \
> + {
> \
> + for (int i = 0; i < NUM; ++i)
> \
> + a[i] = (b[i] > c[i]) OP (d[i] < e[i]);
> \
> + }
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-1.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-1.c
> new file mode 100644
> index 00000000000..d5b65ff5500
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-1.c
> @@ -0,0 +1,69 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_MASK_LOGIC (and, 1, int8_t, &)
> +DEF_MASK_LOGIC (and, 2, int8_t, &)
> +DEF_MASK_LOGIC (and, 4, int8_t, &)
> +DEF_MASK_LOGIC (and, 8, int8_t, &)
> +DEF_MASK_LOGIC (and, 16, int8_t, &)
> +DEF_MASK_LOGIC (and, 32, int8_t, &)
> +DEF_MASK_LOGIC (and, 64, int8_t, &)
> +DEF_MASK_LOGIC (and, 128, int8_t, &)
> +DEF_MASK_LOGIC (and, 256, int8_t, &)
> +DEF_MASK_LOGIC (and, 512, int8_t, &)
> +DEF_MASK_LOGIC (and, 1024, int8_t, &)
> +DEF_MASK_LOGIC (and, 2048, int8_t, &)
> +DEF_MASK_LOGIC (and, 4096, int8_t, &)
> +
> +DEF_MASK_LOGIC (and, 1, int16_t, &)
> +DEF_MASK_LOGIC (and, 2, int16_t, &)
> +DEF_MASK_LOGIC (and, 4, int16_t, &)
> +DEF_MASK_LOGIC (and, 8, int16_t, &)
> +DEF_MASK_LOGIC (and, 16, int16_t, &)
> +DEF_MASK_LOGIC (and, 32, int16_t, &)
> +DEF_MASK_LOGIC (and, 64, int16_t, &)
> +DEF_MASK_LOGIC (and, 128, int16_t, &)
> +DEF_MASK_LOGIC (and, 256, int16_t, &)
> +DEF_MASK_LOGIC (and, 512, int16_t, &)
> +DEF_MASK_LOGIC (and, 1024, int16_t, &)
> +DEF_MASK_LOGIC (and, 2048, int16_t, &)
> +
> +DEF_MASK_LOGIC (and, 1, int32_t, &)
> +DEF_MASK_LOGIC (and, 2, int32_t, &)
> +DEF_MASK_LOGIC (and, 4, int32_t, &)
> +DEF_MASK_LOGIC (and, 8, int32_t, &)
> +DEF_MASK_LOGIC (and, 16, int32_t, &)
> +DEF_MASK_LOGIC (and, 32, int32_t, &)
> +DEF_MASK_LOGIC (and, 64, int32_t, &)
> +DEF_MASK_LOGIC (and, 128, int32_t, &)
> +DEF_MASK_LOGIC (and, 256, int32_t, &)
> +DEF_MASK_LOGIC (and, 512, int32_t, &)
> +DEF_MASK_LOGIC (and, 1024, int32_t, &)
> +
> +DEF_MASK_LOGIC (and, 1, int64_t, &)
> +DEF_MASK_LOGIC (and, 2, int64_t, &)
> +DEF_MASK_LOGIC (and, 4, int64_t, &)
> +DEF_MASK_LOGIC (and, 8, int64_t, &)
> +DEF_MASK_LOGIC (and, 16, int64_t, &)
> +DEF_MASK_LOGIC (and, 32, int64_t, &)
> +DEF_MASK_LOGIC (and, 64, int64_t, &)
> +DEF_MASK_LOGIC (and, 128, int64_t, &)
> +DEF_MASK_LOGIC (and, 256, int64_t, &)
> +DEF_MASK_LOGIC (and, 512, int64_t, &)
> +
> +/* { dg-final { scan-assembler-times {vmand\.mm} 42 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-2.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-2.c
> new file mode 100644
> index 00000000000..7fe4ec96dc8
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-2.c
> @@ -0,0 +1,69 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_MASK_LOGIC (and, 1, int8_t, |)
> +DEF_MASK_LOGIC (and, 2, int8_t, |)
> +DEF_MASK_LOGIC (and, 4, int8_t, |)
> +DEF_MASK_LOGIC (and, 8, int8_t, |)
> +DEF_MASK_LOGIC (and, 16, int8_t, |)
> +DEF_MASK_LOGIC (and, 32, int8_t, |)
> +DEF_MASK_LOGIC (and, 64, int8_t, |)
> +DEF_MASK_LOGIC (and, 128, int8_t, |)
> +DEF_MASK_LOGIC (and, 256, int8_t, |)
> +DEF_MASK_LOGIC (and, 512, int8_t, |)
> +DEF_MASK_LOGIC (and, 1024, int8_t, |)
> +DEF_MASK_LOGIC (and, 2048, int8_t, |)
> +DEF_MASK_LOGIC (and, 4096, int8_t, |)
> +
> +DEF_MASK_LOGIC (and, 1, int16_t, |)
> +DEF_MASK_LOGIC (and, 2, int16_t, |)
> +DEF_MASK_LOGIC (and, 4, int16_t, |)
> +DEF_MASK_LOGIC (and, 8, int16_t, |)
> +DEF_MASK_LOGIC (and, 16, int16_t, |)
> +DEF_MASK_LOGIC (and, 32, int16_t, |)
> +DEF_MASK_LOGIC (and, 64, int16_t, |)
> +DEF_MASK_LOGIC (and, 128, int16_t, |)
> +DEF_MASK_LOGIC (and, 256, int16_t, |)
> +DEF_MASK_LOGIC (and, 512, int16_t, |)
> +DEF_MASK_LOGIC (and, 1024, int16_t, |)
> +DEF_MASK_LOGIC (and, 2048, int16_t, |)
> +
> +DEF_MASK_LOGIC (and, 1, int32_t, |)
> +DEF_MASK_LOGIC (and, 2, int32_t, |)
> +DEF_MASK_LOGIC (and, 4, int32_t, |)
> +DEF_MASK_LOGIC (and, 8, int32_t, |)
> +DEF_MASK_LOGIC (and, 16, int32_t, |)
> +DEF_MASK_LOGIC (and, 32, int32_t, |)
> +DEF_MASK_LOGIC (and, 64, int32_t, |)
> +DEF_MASK_LOGIC (and, 128, int32_t, |)
> +DEF_MASK_LOGIC (and, 256, int32_t, |)
> +DEF_MASK_LOGIC (and, 512, int32_t, |)
> +DEF_MASK_LOGIC (and, 1024, int32_t, |)
> +
> +DEF_MASK_LOGIC (and, 1, int64_t, |)
> +DEF_MASK_LOGIC (and, 2, int64_t, |)
> +DEF_MASK_LOGIC (and, 4, int64_t, |)
> +DEF_MASK_LOGIC (and, 8, int64_t, |)
> +DEF_MASK_LOGIC (and, 16, int64_t, |)
> +DEF_MASK_LOGIC (and, 32, int64_t, |)
> +DEF_MASK_LOGIC (and, 64, int64_t, |)
> +DEF_MASK_LOGIC (and, 128, int64_t, |)
> +DEF_MASK_LOGIC (and, 256, int64_t, |)
> +DEF_MASK_LOGIC (and, 512, int64_t, |)
> +
> +/* { dg-final { scan-assembler-times {vmor\.mm} 42 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-3.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-3.c
> new file mode 100644
> index 00000000000..7d2254b326d
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-3.c
> @@ -0,0 +1,69 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_MASK_LOGIC (and, 1, int8_t, ^)
> +DEF_MASK_LOGIC (and, 2, int8_t, ^)
> +DEF_MASK_LOGIC (and, 4, int8_t, ^)
> +DEF_MASK_LOGIC (and, 8, int8_t, ^)
> +DEF_MASK_LOGIC (and, 16, int8_t, ^)
> +DEF_MASK_LOGIC (and, 32, int8_t, ^)
> +DEF_MASK_LOGIC (and, 64, int8_t, ^)
> +DEF_MASK_LOGIC (and, 128, int8_t, ^)
> +DEF_MASK_LOGIC (and, 256, int8_t, ^)
> +DEF_MASK_LOGIC (and, 512, int8_t, ^)
> +DEF_MASK_LOGIC (and, 1024, int8_t, ^)
> +DEF_MASK_LOGIC (and, 2048, int8_t, ^)
> +DEF_MASK_LOGIC (and, 4096, int8_t, ^)
> +
> +DEF_MASK_LOGIC (and, 1, int16_t, ^)
> +DEF_MASK_LOGIC (and, 2, int16_t, ^)
> +DEF_MASK_LOGIC (and, 4, int16_t, ^)
> +DEF_MASK_LOGIC (and, 8, int16_t, ^)
> +DEF_MASK_LOGIC (and, 16, int16_t, ^)
> +DEF_MASK_LOGIC (and, 32, int16_t, ^)
> +DEF_MASK_LOGIC (and, 64, int16_t, ^)
> +DEF_MASK_LOGIC (and, 128, int16_t, ^)
> +DEF_MASK_LOGIC (and, 256, int16_t, ^)
> +DEF_MASK_LOGIC (and, 512, int16_t, ^)
> +DEF_MASK_LOGIC (and, 1024, int16_t, ^)
> +DEF_MASK_LOGIC (and, 2048, int16_t, ^)
> +
> +DEF_MASK_LOGIC (and, 1, int32_t, ^)
> +DEF_MASK_LOGIC (and, 2, int32_t, ^)
> +DEF_MASK_LOGIC (and, 4, int32_t, ^)
> +DEF_MASK_LOGIC (and, 8, int32_t, ^)
> +DEF_MASK_LOGIC (and, 16, int32_t, ^)
> +DEF_MASK_LOGIC (and, 32, int32_t, ^)
> +DEF_MASK_LOGIC (and, 64, int32_t, ^)
> +DEF_MASK_LOGIC (and, 128, int32_t, ^)
> +DEF_MASK_LOGIC (and, 256, int32_t, ^)
> +DEF_MASK_LOGIC (and, 512, int32_t, ^)
> +DEF_MASK_LOGIC (and, 1024, int32_t, ^)
> +
> +DEF_MASK_LOGIC (and, 1, int64_t, ^)
> +DEF_MASK_LOGIC (and, 2, int64_t, ^)
> +DEF_MASK_LOGIC (and, 4, int64_t, ^)
> +DEF_MASK_LOGIC (and, 8, int64_t, ^)
> +DEF_MASK_LOGIC (and, 16, int64_t, ^)
> +DEF_MASK_LOGIC (and, 32, int64_t, ^)
> +DEF_MASK_LOGIC (and, 64, int64_t, ^)
> +DEF_MASK_LOGIC (and, 128, int64_t, ^)
> +DEF_MASK_LOGIC (and, 256, int64_t, ^)
> +DEF_MASK_LOGIC (and, 512, int64_t, ^)
> +
> +/* { dg-final { scan-assembler-times {vmxor\.mm} 42 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> --
> 2.36.3
>
>
Committed, thanks Kito.
Pan
-----Original Message-----
From: Gcc-patches <gcc-patches-bounces+pan2.li=intel.com@gcc.gnu.org> On Behalf Of Kito Cheng via Gcc-patches
Sent: Thursday, September 14, 2023 10:23 PM
To: Juzhe-Zhong <juzhe.zhong@rivai.ai>
Cc: GCC Patches <gcc-patches@gcc.gnu.org>; Kito Cheng <kito.cheng@gmail.com>; Jeff Law <jeffreyalaw@gmail.com>; Robin Dapp <rdapp.gcc@gmail.com>
Subject: Re: [PATCH] RISC-V: Support VLS modes mask operations
LGTM
Juzhe-Zhong <juzhe.zhong@rivai.ai> 於 2023年9月14日 週四 20:44 寫道:
> This patch support mask operations (comparison and logical).
>
> This patch reduce these FAILs of "vect" testsuite:
> FAIL: gcc.dg/vect/vect-bic-bitmask-12.c -flto -ffat-lto-objects
> scan-tree-dump dce7 "<=\\s*.+{ 255,.+}"
> FAIL: gcc.dg/vect/vect-bic-bitmask-12.c scan-tree-dump dce7 "<=\\s*.+{
> 255,.+}"
> FAIL: gcc.dg/vect/vect-bic-bitmask-23.c -flto -ffat-lto-objects
> scan-tree-dump dce7 "<=\\s*.+{ 255, 15, 1, 65535 }"
> FAIL: gcc.dg/vect/vect-bic-bitmask-23.c scan-tree-dump dce7 "<=\\s*.+{
> 255, 15, 1, 65535 }"
>
> Full regression passed (with reducing 4 FAILs).
>
> gcc/ChangeLog:
>
> * config/riscv/autovec-opt.md: Add VLS mask modes.
> * config/riscv/autovec.md (@vcond_mask_<mode><vm>): Remove @.
> (vcond_mask_<mode><vm>): Add VLS mask modes.
> * config/riscv/vector.md: Ditto.
>
> gcc/testsuite/ChangeLog:
>
> * gcc.target/riscv/rvv/autovec/vls/def.h: Add VLS tests.
> * gcc.target/riscv/rvv/autovec/vls/cmp-1.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/cmp-2.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/cmp-3.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/cmp-4.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/cmp-5.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/cmp-6.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/mask-1.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/mask-2.c: New test.
> * gcc.target/riscv/rvv/autovec/vls/mask-3.c: New test.
>
> ---
> gcc/config/riscv/autovec-opt.md | 18 +--
> gcc/config/riscv/autovec.md | 32 +++---
> gcc/config/riscv/vector.md | 60 +++++-----
> .../gcc.target/riscv/rvv/autovec/vls/cmp-1.c | 106 ++++++++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/cmp-2.c | 106 ++++++++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/cmp-3.c | 106 ++++++++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/cmp-4.c | 106 ++++++++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/cmp-5.c | 106 ++++++++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/cmp-6.c | 106 ++++++++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/def.h | 9 ++
> .../gcc.target/riscv/rvv/autovec/vls/mask-1.c | 69 ++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/mask-2.c | 69 ++++++++++++
> .../gcc.target/riscv/rvv/autovec/vls/mask-3.c | 69 ++++++++++++
> 13 files changed, 907 insertions(+), 55 deletions(-)
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-1.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-2.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-3.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-4.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-5.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-6.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-1.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-2.c
> create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-3.c
>
> diff --git a/gcc/config/riscv/autovec-opt.md
> b/gcc/config/riscv/autovec-opt.md
> index e26c01856ff..22ab8afc994 100644
> --- a/gcc/config/riscv/autovec-opt.md
> +++ b/gcc/config/riscv/autovec-opt.md
> @@ -67,10 +67,10 @@
> ;;
> -------------------------------------------------------------------------
>
> (define_insn_and_split "*<optab>not<mode>"
> - [(set (match_operand:VB 0 "register_operand" "=vr")
> - (bitmanip_bitwise:VB
> - (not:VB (match_operand:VB 2 "register_operand" " vr"))
> - (match_operand:VB 1 "register_operand" " vr")))]
> + [(set (match_operand:VB_VLS 0 "register_operand" "=vr")
> + (bitmanip_bitwise:VB_VLS
> + (not:VB_VLS (match_operand:VB_VLS 2 "register_operand" " vr"))
> + (match_operand:VB_VLS 1 "register_operand" " vr")))]
> "TARGET_VECTOR && can_create_pseudo_p ()"
> "#"
> "&& 1"
> @@ -93,11 +93,11 @@
> ;;
> -------------------------------------------------------------------------
>
> (define_insn_and_split "*n<optab><mode>"
> - [(set (match_operand:VB 0 "register_operand" "=vr")
> - (not:VB
> - (any_bitwise:VB
> - (match_operand:VB 1 "register_operand" " vr")
> - (match_operand:VB 2 "register_operand" " vr"))))]
> + [(set (match_operand:VB_VLS 0 "register_operand" "=vr")
> + (not:VB_VLS
> + (any_bitwise:VB_VLS
> + (match_operand:VB_VLS 1 "register_operand" " vr")
> + (match_operand:VB_VLS 2 "register_operand" " vr"))))]
> "TARGET_VECTOR && can_create_pseudo_p ()"
> "#"
> "&& 1"
> diff --git a/gcc/config/riscv/autovec.md b/gcc/config/riscv/autovec.md
> index 7121bab1716..d2002a8ee26 100644
> --- a/gcc/config/riscv/autovec.md
> +++ b/gcc/config/riscv/autovec.md
> @@ -495,9 +495,9 @@
> ;;
> -------------------------------------------------------------------------
>
> (define_insn_and_split "<optab><mode>3"
> - [(set (match_operand:VB 0 "register_operand" "=vr")
> - (any_bitwise:VB (match_operand:VB 1 "register_operand" " vr")
> - (match_operand:VB 2 "register_operand" " vr")))]
> + [(set (match_operand:VB_VLS 0 "register_operand" "=vr")
> + (any_bitwise:VB_VLS (match_operand:VB_VLS 1 "register_operand" "
> vr")
> + (match_operand:VB_VLS 2 "register_operand" "
> vr")))]
> "TARGET_VECTOR && can_create_pseudo_p ()"
> "#"
> "&& 1"
> @@ -518,8 +518,8 @@
> ;;
> -------------------------------------------------------------------------
>
> (define_insn_and_split "one_cmpl<mode>2"
> - [(set (match_operand:VB 0 "register_operand" "=vr")
> - (not:VB (match_operand:VB 1 "register_operand" " vr")))]
> + [(set (match_operand:VB_VLS 0 "register_operand" "=vr")
> + (not:VB_VLS (match_operand:VB_VLS 1 "register_operand" " vr")))]
> "TARGET_VECTOR && can_create_pseudo_p ()"
> "#"
> "&& 1"
> @@ -545,12 +545,12 @@
> ;; - vfmerge.vf
> ;;
> -------------------------------------------------------------------------
>
> -(define_insn_and_split "@vcond_mask_<mode><vm>"
> - [(set (match_operand:V 0 "register_operand")
> - (if_then_else:V
> +(define_insn_and_split "vcond_mask_<mode><vm>"
> + [(set (match_operand:V_VLS 0 "register_operand")
> + (if_then_else:V_VLS
> (match_operand:<VM> 3 "register_operand")
> - (match_operand:V 1 "nonmemory_operand")
> - (match_operand:V 2 "register_operand")))]
> + (match_operand:V_VLS 1 "nonmemory_operand")
> + (match_operand:V_VLS 2 "register_operand")))]
> "TARGET_VECTOR && can_create_pseudo_p ()"
> "#"
> "&& 1"
> @@ -609,8 +609,8 @@
> (define_expand "vec_cmp<mode><vm>"
> [(set (match_operand:<VM> 0 "register_operand")
> (match_operator:<VM> 1 "comparison_operator"
> - [(match_operand:VI 2 "register_operand")
> - (match_operand:VI 3 "register_operand")]))]
> + [(match_operand:V_VLSI 2 "register_operand")
> + (match_operand:V_VLSI 3 "register_operand")]))]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_vec_cmp (operands[0], GET_CODE (operands[1]),
> @@ -622,8 +622,8 @@
> (define_expand "vec_cmpu<mode><vm>"
> [(set (match_operand:<VM> 0 "register_operand")
> (match_operator:<VM> 1 "comparison_operator"
> - [(match_operand:VI 2 "register_operand")
> - (match_operand:VI 3 "register_operand")]))]
> + [(match_operand:V_VLSI 2 "register_operand")
> + (match_operand:V_VLSI 3 "register_operand")]))]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_vec_cmp (operands[0], GET_CODE (operands[1]),
> @@ -635,8 +635,8 @@
> (define_expand "vec_cmp<mode><vm>"
> [(set (match_operand:<VM> 0 "register_operand")
> (match_operator:<VM> 1 "comparison_operator"
> - [(match_operand:VF 2 "register_operand")
> - (match_operand:VF 3 "register_operand")]))]
> + [(match_operand:V_VLSF 2 "register_operand")
> + (match_operand:V_VLSF 3 "register_operand")]))]
> "TARGET_VECTOR"
> {
> riscv_vector::expand_vec_cmp_float (operands[0], GET_CODE
> (operands[1]),
> diff --git a/gcc/config/riscv/vector.md b/gcc/config/riscv/vector.md
> index 4630af6cbff..8c0e86576a0 100644
> --- a/gcc/config/riscv/vector.md
> +++ b/gcc/config/riscv/vector.md
> @@ -7177,8 +7177,8 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "signed_order_operator"
> - [(match_operand:VF 4 "register_operand")
> - (match_operand:VF 5 "register_operand")])
> + [(match_operand:V_VLSF 4 "register_operand")
> + (match_operand:V_VLSF 5 "register_operand")])
> (match_operand:<VM> 2 "vector_merge_operand")))]
> "TARGET_VECTOR"
> {})
> @@ -7195,10 +7195,10 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "signed_order_operator"
> - [(match_operand:VF 4 "register_operand" " vr,
> vr")
> - (match_operand:VF 5 "register_operand" " vr,
> vr")])
> + [(match_operand:V_VLSF 4 "register_operand" " vr,
> vr")
> + (match_operand:V_VLSF 5 "register_operand" " vr,
> vr")])
> (match_operand:<VM> 2 "vector_merge_operand" " vu,
> 0")))]
> - "TARGET_VECTOR && known_le (GET_MODE_SIZE (<MODE>mode),
> BYTES_PER_RISCV_VECTOR)"
> + "TARGET_VECTOR && riscv_vector::cmp_lmul_le_one (<MODE>mode)"
> "vmf%B3.vv\t%0,%4,%5%p1"
> [(set_attr "type" "vfcmp")
> (set_attr "mode" "<MODE>")])
> @@ -7214,8 +7214,8 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 2 "signed_order_operator"
> - [(match_operand:VF 3 "register_operand" " vr")
> - (match_operand:VF 4 "register_operand" " vr")])
> + [(match_operand:V_VLSF 3 "register_operand" " vr")
> + (match_operand:V_VLSF 4 "register_operand" " vr")])
> (match_dup 1)))]
> "TARGET_VECTOR"
> "vmf%B2.vv\t%0,%3,%4,v0.t"
> @@ -7238,10 +7238,10 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "signed_order_operator"
> - [(match_operand:VF 4 "register_operand" " vr,
> 0, vr, 0, 0, vr, 0, vr, vr")
> - (match_operand:VF 5 "register_operand" " vr,
> vr, 0, 0, vr, 0, 0, vr, vr")])
> + [(match_operand:V_VLSF 4 "register_operand" " vr,
> 0, vr, 0, 0, vr, 0, vr, vr")
> + (match_operand:V_VLSF 5 "register_operand" " vr,
> vr, 0, 0, vr, 0, 0, vr, vr")])
> (match_operand:<VM> 2 "vector_merge_operand" " vu,
> vu, vu, vu, 0, 0, 0, vu, 0")))]
> - "TARGET_VECTOR && known_gt (GET_MODE_SIZE (<MODE>mode),
> BYTES_PER_RISCV_VECTOR)"
> + "TARGET_VECTOR && riscv_vector::cmp_lmul_gt_one (<MODE>mode)"
> "vmf%B3.vv\t%0,%4,%5%p1"
> [(set_attr "type" "vfcmp")
> (set_attr "mode" "<MODE>")])
> @@ -7257,8 +7257,8 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "signed_order_operator"
> - [(match_operand:VF 4 "register_operand")
> - (vec_duplicate:VF
> + [(match_operand:V_VLSF 4 "register_operand")
> + (vec_duplicate:V_VLSF
> (match_operand:<VEL> 5 "register_operand"))])
> (match_operand:<VM> 2 "vector_merge_operand")))]
> "TARGET_VECTOR"
> @@ -7275,8 +7275,8 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 2 "signed_order_operator"
> - [(match_operand:VF 3 "register_operand" " vr")
> - (vec_duplicate:VF
> + [(match_operand:V_VLSF 3 "register_operand" " vr")
> + (vec_duplicate:V_VLSF
> (match_operand:<VEL> 4 "register_operand" " f"))])
> (match_dup 1)))]
> "TARGET_VECTOR"
> @@ -7300,11 +7300,11 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "signed_order_operator"
> - [(match_operand:VF 4 "register_operand" " vr,
> vr")
> - (vec_duplicate:VF
> + [(match_operand:V_VLSF 4 "register_operand" " vr,
> vr")
> + (vec_duplicate:V_VLSF
> (match_operand:<VEL> 5 "register_operand" " f,
> f"))])
> (match_operand:<VM> 2 "vector_merge_operand" " vu,
> 0")))]
> - "TARGET_VECTOR && known_le (GET_MODE_SIZE (<MODE>mode),
> BYTES_PER_RISCV_VECTOR)"
> + "TARGET_VECTOR && riscv_vector::cmp_lmul_le_one (<MODE>mode)"
> "vmf%B3.vf\t%0,%4,%5%p1"
> [(set_attr "type" "vfcmp")
> (set_attr "mode" "<MODE>")])
> @@ -7321,11 +7321,11 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "signed_order_operator"
> - [(match_operand:VF 4 "register_operand" " vr,
> 0, 0, vr, vr")
> - (vec_duplicate:VF
> + [(match_operand:V_VLSF 4 "register_operand" " vr,
> 0, 0, vr, vr")
> + (vec_duplicate:V_VLSF
> (match_operand:<VEL> 5 "register_operand" " f,
> f, f, f, f"))])
> (match_operand:<VM> 2 "vector_merge_operand" " vu,
> vu, 0, vu, 0")))]
> - "TARGET_VECTOR && known_gt (GET_MODE_SIZE (<MODE>mode),
> BYTES_PER_RISCV_VECTOR)"
> + "TARGET_VECTOR && riscv_vector::cmp_lmul_gt_one (<MODE>mode)"
> "vmf%B3.vf\t%0,%4,%5%p1"
> [(set_attr "type" "vfcmp")
> (set_attr "mode" "<MODE>")])
> @@ -7341,9 +7341,9 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "equality_operator"
> - [(vec_duplicate:VF
> + [(vec_duplicate:V_VLSF
> (match_operand:<VEL> 5 "register_operand"))
> - (match_operand:VF 4 "register_operand")])
> + (match_operand:V_VLSF 4 "register_operand")])
> (match_operand:<VM> 2 "vector_merge_operand")))]
> "TARGET_VECTOR"
> {})
> @@ -7359,9 +7359,9 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 2 "equality_operator"
> - [(vec_duplicate:VF
> + [(vec_duplicate:V_VLSF
> (match_operand:<VEL> 4 "register_operand" " f"))
> - (match_operand:VF 3 "register_operand" " vr")])
> + (match_operand:V_VLSF 3 "register_operand" " vr")])
> (match_dup 1)))]
> "TARGET_VECTOR"
> "vmf%B2.vf\t%0,%3,%4,v0.t"
> @@ -7384,11 +7384,11 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "equality_operator"
> - [(vec_duplicate:VF
> + [(vec_duplicate:V_VLSF
> (match_operand:<VEL> 5 "register_operand" " f,
> f"))
> - (match_operand:VF 4 "register_operand" " vr,
> vr")])
> + (match_operand:V_VLSF 4 "register_operand" " vr,
> vr")])
> (match_operand:<VM> 2 "vector_merge_operand" " vu,
> 0")))]
> - "TARGET_VECTOR && known_le (GET_MODE_SIZE (<MODE>mode),
> BYTES_PER_RISCV_VECTOR)"
> + "TARGET_VECTOR && riscv_vector::cmp_lmul_le_one (<MODE>mode)"
> "vmf%B3.vf\t%0,%4,%5%p1"
> [(set_attr "type" "vfcmp")
> (set_attr "mode" "<MODE>")])
> @@ -7405,11 +7405,11 @@
> (reg:SI VL_REGNUM)
> (reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
> (match_operator:<VM> 3 "equality_operator"
> - [(vec_duplicate:VF
> + [(vec_duplicate:V_VLSF
> (match_operand:<VEL> 5 "register_operand" " f,
> f, f, f, f"))
> - (match_operand:VF 4 "register_operand" " vr,
> 0, 0, vr, vr")])
> + (match_operand:V_VLSF 4 "register_operand" " vr,
> 0, 0, vr, vr")])
> (match_operand:<VM> 2 "vector_merge_operand" " vu,
> vu, 0, vu, 0")))]
> - "TARGET_VECTOR && known_gt (GET_MODE_SIZE (<MODE>mode),
> BYTES_PER_RISCV_VECTOR)"
> + "TARGET_VECTOR && riscv_vector::cmp_lmul_gt_one (<MODE>mode)"
> "vmf%B3.vf\t%0,%4,%5%p1"
> [(set_attr "type" "vfcmp")
> (set_attr "mode" "<MODE>")])
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-1.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-1.c
> new file mode 100644
> index 00000000000..05742b90fd5
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-1.c
> @@ -0,0 +1,106 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_OP_VV (eq, 1, int8_t, ==)
> +DEF_OP_VV (eq, 2, int8_t, ==)
> +DEF_OP_VV (eq, 4, int8_t, ==)
> +DEF_OP_VV (eq, 8, int8_t, ==)
> +DEF_OP_VV (eq, 16, int8_t, ==)
> +DEF_OP_VV (eq, 32, int8_t, ==)
> +DEF_OP_VV (eq, 64, int8_t, ==)
> +DEF_OP_VV (eq, 128, int8_t, ==)
> +DEF_OP_VV (eq, 256, int8_t, ==)
> +DEF_OP_VV (eq, 512, int8_t, ==)
> +DEF_OP_VV (eq, 1024, int8_t, ==)
> +DEF_OP_VV (eq, 2048, int8_t, ==)
> +DEF_OP_VV (eq, 4096, int8_t, ==)
> +
> +DEF_OP_VV (eq, 1, int16_t, ==)
> +DEF_OP_VV (eq, 2, int16_t, ==)
> +DEF_OP_VV (eq, 4, int16_t, ==)
> +DEF_OP_VV (eq, 8, int16_t, ==)
> +DEF_OP_VV (eq, 16, int16_t, ==)
> +DEF_OP_VV (eq, 32, int16_t, ==)
> +DEF_OP_VV (eq, 64, int16_t, ==)
> +DEF_OP_VV (eq, 128, int16_t, ==)
> +DEF_OP_VV (eq, 256, int16_t, ==)
> +DEF_OP_VV (eq, 512, int16_t, ==)
> +DEF_OP_VV (eq, 1024, int16_t, ==)
> +DEF_OP_VV (eq, 2048, int16_t, ==)
> +
> +DEF_OP_VV (eq, 1, int32_t, ==)
> +DEF_OP_VV (eq, 2, int32_t, ==)
> +DEF_OP_VV (eq, 4, int32_t, ==)
> +DEF_OP_VV (eq, 8, int32_t, ==)
> +DEF_OP_VV (eq, 16, int32_t, ==)
> +DEF_OP_VV (eq, 32, int32_t, ==)
> +DEF_OP_VV (eq, 64, int32_t, ==)
> +DEF_OP_VV (eq, 128, int32_t, ==)
> +DEF_OP_VV (eq, 256, int32_t, ==)
> +DEF_OP_VV (eq, 512, int32_t, ==)
> +DEF_OP_VV (eq, 1024, int32_t, ==)
> +
> +DEF_OP_VV (eq, 1, int64_t, ==)
> +DEF_OP_VV (eq, 2, int64_t, ==)
> +DEF_OP_VV (eq, 4, int64_t, ==)
> +DEF_OP_VV (eq, 8, int64_t, ==)
> +DEF_OP_VV (eq, 16, int64_t, ==)
> +DEF_OP_VV (eq, 32, int64_t, ==)
> +DEF_OP_VV (eq, 64, int64_t, ==)
> +DEF_OP_VV (eq, 128, int64_t, ==)
> +DEF_OP_VV (eq, 256, int64_t, ==)
> +DEF_OP_VV (eq, 512, int64_t, ==)
> +
> +DEF_OP_VV (eq, 1, _Float16, ==)
> +DEF_OP_VV (eq, 2, _Float16, ==)
> +DEF_OP_VV (eq, 4, _Float16, ==)
> +DEF_OP_VV (eq, 8, _Float16, ==)
> +DEF_OP_VV (eq, 16, _Float16, ==)
> +DEF_OP_VV (eq, 32, _Float16, ==)
> +DEF_OP_VV (eq, 64, _Float16, ==)
> +DEF_OP_VV (eq, 128, _Float16, ==)
> +DEF_OP_VV (eq, 256, _Float16, ==)
> +DEF_OP_VV (eq, 512, _Float16, ==)
> +DEF_OP_VV (eq, 1024, _Float16, ==)
> +DEF_OP_VV (eq, 2048, _Float16, ==)
> +
> +DEF_OP_VV (eq, 1, float, ==)
> +DEF_OP_VV (eq, 2, float, ==)
> +DEF_OP_VV (eq, 4, float, ==)
> +DEF_OP_VV (eq, 8, float, ==)
> +DEF_OP_VV (eq, 16, float, ==)
> +DEF_OP_VV (eq, 32, float, ==)
> +DEF_OP_VV (eq, 64, float, ==)
> +DEF_OP_VV (eq, 128, float, ==)
> +DEF_OP_VV (eq, 256, float, ==)
> +DEF_OP_VV (eq, 512, float, ==)
> +DEF_OP_VV (eq, 1024, float, ==)
> +
> +DEF_OP_VV (eq, 1, double, ==)
> +DEF_OP_VV (eq, 2, double, ==)
> +DEF_OP_VV (eq, 4, double, ==)
> +DEF_OP_VV (eq, 8, double, ==)
> +DEF_OP_VV (eq, 16, double, ==)
> +DEF_OP_VV (eq, 32, double, ==)
> +DEF_OP_VV (eq, 64, double, ==)
> +DEF_OP_VV (eq, 128, double, ==)
> +DEF_OP_VV (eq, 256, double, ==)
> +DEF_OP_VV (eq, 512, double, ==)
> +
> +/* { dg-final { scan-assembler-times {vmseq\.vv} 42 } } */
> +/* { dg-final { scan-assembler-times {vmfeq\.vv} 27 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-2.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-2.c
> new file mode 100644
> index 00000000000..39a56025818
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-2.c
> @@ -0,0 +1,106 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_OP_VV (ne, 1, int8_t, !=)
> +DEF_OP_VV (ne, 2, int8_t, !=)
> +DEF_OP_VV (ne, 4, int8_t, !=)
> +DEF_OP_VV (ne, 8, int8_t, !=)
> +DEF_OP_VV (ne, 16, int8_t, !=)
> +DEF_OP_VV (ne, 32, int8_t, !=)
> +DEF_OP_VV (ne, 64, int8_t, !=)
> +DEF_OP_VV (ne, 128, int8_t, !=)
> +DEF_OP_VV (ne, 256, int8_t, !=)
> +DEF_OP_VV (ne, 512, int8_t, !=)
> +DEF_OP_VV (ne, 1024, int8_t, !=)
> +DEF_OP_VV (ne, 2048, int8_t, !=)
> +DEF_OP_VV (ne, 4096, int8_t, !=)
> +
> +DEF_OP_VV (ne, 1, int16_t, !=)
> +DEF_OP_VV (ne, 2, int16_t, !=)
> +DEF_OP_VV (ne, 4, int16_t, !=)
> +DEF_OP_VV (ne, 8, int16_t, !=)
> +DEF_OP_VV (ne, 16, int16_t, !=)
> +DEF_OP_VV (ne, 32, int16_t, !=)
> +DEF_OP_VV (ne, 64, int16_t, !=)
> +DEF_OP_VV (ne, 128, int16_t, !=)
> +DEF_OP_VV (ne, 256, int16_t, !=)
> +DEF_OP_VV (ne, 512, int16_t, !=)
> +DEF_OP_VV (ne, 1024, int16_t, !=)
> +DEF_OP_VV (ne, 2048, int16_t, !=)
> +
> +DEF_OP_VV (ne, 1, int32_t, !=)
> +DEF_OP_VV (ne, 2, int32_t, !=)
> +DEF_OP_VV (ne, 4, int32_t, !=)
> +DEF_OP_VV (ne, 8, int32_t, !=)
> +DEF_OP_VV (ne, 16, int32_t, !=)
> +DEF_OP_VV (ne, 32, int32_t, !=)
> +DEF_OP_VV (ne, 64, int32_t, !=)
> +DEF_OP_VV (ne, 128, int32_t, !=)
> +DEF_OP_VV (ne, 256, int32_t, !=)
> +DEF_OP_VV (ne, 512, int32_t, !=)
> +DEF_OP_VV (ne, 1024, int32_t, !=)
> +
> +DEF_OP_VV (ne, 1, int64_t, !=)
> +DEF_OP_VV (ne, 2, int64_t, !=)
> +DEF_OP_VV (ne, 4, int64_t, !=)
> +DEF_OP_VV (ne, 8, int64_t, !=)
> +DEF_OP_VV (ne, 16, int64_t, !=)
> +DEF_OP_VV (ne, 32, int64_t, !=)
> +DEF_OP_VV (ne, 64, int64_t, !=)
> +DEF_OP_VV (ne, 128, int64_t, !=)
> +DEF_OP_VV (ne, 256, int64_t, !=)
> +DEF_OP_VV (ne, 512, int64_t, !=)
> +
> +DEF_OP_VV (ne, 1, _Float16, !=)
> +DEF_OP_VV (ne, 2, _Float16, !=)
> +DEF_OP_VV (ne, 4, _Float16, !=)
> +DEF_OP_VV (ne, 8, _Float16, !=)
> +DEF_OP_VV (ne, 16, _Float16, !=)
> +DEF_OP_VV (ne, 32, _Float16, !=)
> +DEF_OP_VV (ne, 64, _Float16, !=)
> +DEF_OP_VV (ne, 128, _Float16, !=)
> +DEF_OP_VV (ne, 256, _Float16, !=)
> +DEF_OP_VV (ne, 512, _Float16, !=)
> +DEF_OP_VV (ne, 1024, _Float16, !=)
> +DEF_OP_VV (ne, 2048, _Float16, !=)
> +
> +DEF_OP_VV (ne, 1, float, !=)
> +DEF_OP_VV (ne, 2, float, !=)
> +DEF_OP_VV (ne, 4, float, !=)
> +DEF_OP_VV (ne, 8, float, !=)
> +DEF_OP_VV (ne, 16, float, !=)
> +DEF_OP_VV (ne, 32, float, !=)
> +DEF_OP_VV (ne, 64, float, !=)
> +DEF_OP_VV (ne, 128, float, !=)
> +DEF_OP_VV (ne, 256, float, !=)
> +DEF_OP_VV (ne, 512, float, !=)
> +DEF_OP_VV (ne, 1024, float, !=)
> +
> +DEF_OP_VV (ne, 1, double, !=)
> +DEF_OP_VV (ne, 2, double, !=)
> +DEF_OP_VV (ne, 4, double, !=)
> +DEF_OP_VV (ne, 8, double, !=)
> +DEF_OP_VV (ne, 16, double, !=)
> +DEF_OP_VV (ne, 32, double, !=)
> +DEF_OP_VV (ne, 64, double, !=)
> +DEF_OP_VV (ne, 128, double, !=)
> +DEF_OP_VV (ne, 256, double, !=)
> +DEF_OP_VV (ne, 512, double, !=)
> +
> +/* { dg-final { scan-assembler-times {vmsne\.vv} 42 } } */
> +/* { dg-final { scan-assembler-times {vmfne\.vv} 27 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-3.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-3.c
> new file mode 100644
> index 00000000000..387157d9be6
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-3.c
> @@ -0,0 +1,106 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_OP_VV (gt, 1, int8_t, >)
> +DEF_OP_VV (gt, 2, int8_t, >)
> +DEF_OP_VV (gt, 4, int8_t, >)
> +DEF_OP_VV (gt, 8, int8_t, >)
> +DEF_OP_VV (gt, 16, int8_t, >)
> +DEF_OP_VV (gt, 32, int8_t, >)
> +DEF_OP_VV (gt, 64, int8_t, >)
> +DEF_OP_VV (gt, 128, int8_t, >)
> +DEF_OP_VV (gt, 256, int8_t, >)
> +DEF_OP_VV (gt, 512, int8_t, >)
> +DEF_OP_VV (gt, 1024, int8_t, >)
> +DEF_OP_VV (gt, 2048, int8_t, >)
> +DEF_OP_VV (gt, 4096, int8_t, >)
> +
> +DEF_OP_VV (gt, 1, int16_t, >)
> +DEF_OP_VV (gt, 2, int16_t, >)
> +DEF_OP_VV (gt, 4, int16_t, >)
> +DEF_OP_VV (gt, 8, int16_t, >)
> +DEF_OP_VV (gt, 16, int16_t, >)
> +DEF_OP_VV (gt, 32, int16_t, >)
> +DEF_OP_VV (gt, 64, int16_t, >)
> +DEF_OP_VV (gt, 128, int16_t, >)
> +DEF_OP_VV (gt, 256, int16_t, >)
> +DEF_OP_VV (gt, 512, int16_t, >)
> +DEF_OP_VV (gt, 1024, int16_t, >)
> +DEF_OP_VV (gt, 2048, int16_t, >)
> +
> +DEF_OP_VV (gt, 1, int32_t, >)
> +DEF_OP_VV (gt, 2, int32_t, >)
> +DEF_OP_VV (gt, 4, int32_t, >)
> +DEF_OP_VV (gt, 8, int32_t, >)
> +DEF_OP_VV (gt, 16, int32_t, >)
> +DEF_OP_VV (gt, 32, int32_t, >)
> +DEF_OP_VV (gt, 64, int32_t, >)
> +DEF_OP_VV (gt, 128, int32_t, >)
> +DEF_OP_VV (gt, 256, int32_t, >)
> +DEF_OP_VV (gt, 512, int32_t, >)
> +DEF_OP_VV (gt, 1024, int32_t, >)
> +
> +DEF_OP_VV (gt, 1, int64_t, >)
> +DEF_OP_VV (gt, 2, int64_t, >)
> +DEF_OP_VV (gt, 4, int64_t, >)
> +DEF_OP_VV (gt, 8, int64_t, >)
> +DEF_OP_VV (gt, 16, int64_t, >)
> +DEF_OP_VV (gt, 32, int64_t, >)
> +DEF_OP_VV (gt, 64, int64_t, >)
> +DEF_OP_VV (gt, 128, int64_t, >)
> +DEF_OP_VV (gt, 256, int64_t, >)
> +DEF_OP_VV (gt, 512, int64_t, >)
> +
> +DEF_OP_VV (gt, 1, _Float16, >)
> +DEF_OP_VV (gt, 2, _Float16, >)
> +DEF_OP_VV (gt, 4, _Float16, >)
> +DEF_OP_VV (gt, 8, _Float16, >)
> +DEF_OP_VV (gt, 16, _Float16, >)
> +DEF_OP_VV (gt, 32, _Float16, >)
> +DEF_OP_VV (gt, 64, _Float16, >)
> +DEF_OP_VV (gt, 128, _Float16, >)
> +DEF_OP_VV (gt, 256, _Float16, >)
> +DEF_OP_VV (gt, 512, _Float16, >)
> +DEF_OP_VV (gt, 1024, _Float16, >)
> +DEF_OP_VV (gt, 2048, _Float16, >)
> +
> +DEF_OP_VV (gt, 1, float, >)
> +DEF_OP_VV (gt, 2, float, >)
> +DEF_OP_VV (gt, 4, float, >)
> +DEF_OP_VV (gt, 8, float, >)
> +DEF_OP_VV (gt, 16, float, >)
> +DEF_OP_VV (gt, 32, float, >)
> +DEF_OP_VV (gt, 64, float, >)
> +DEF_OP_VV (gt, 128, float, >)
> +DEF_OP_VV (gt, 256, float, >)
> +DEF_OP_VV (gt, 512, float, >)
> +DEF_OP_VV (gt, 1024, float, >)
> +
> +DEF_OP_VV (gt, 1, double, >)
> +DEF_OP_VV (gt, 2, double, >)
> +DEF_OP_VV (gt, 4, double, >)
> +DEF_OP_VV (gt, 8, double, >)
> +DEF_OP_VV (gt, 16, double, >)
> +DEF_OP_VV (gt, 32, double, >)
> +DEF_OP_VV (gt, 64, double, >)
> +DEF_OP_VV (gt, 128, double, >)
> +DEF_OP_VV (gt, 256, double, >)
> +DEF_OP_VV (gt, 512, double, >)
> +
> +/* { dg-final { scan-assembler-times {vmsgt\.vv} 40 } } */
> +/* { dg-final { scan-assembler-times {vmfgt\.vv} 27 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-4.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-4.c
> new file mode 100644
> index 00000000000..40b8871ea3a
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-4.c
> @@ -0,0 +1,106 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_OP_VV (ge, 1, int8_t, >=)
> +DEF_OP_VV (ge, 2, int8_t, >=)
> +DEF_OP_VV (ge, 4, int8_t, >=)
> +DEF_OP_VV (ge, 8, int8_t, >=)
> +DEF_OP_VV (ge, 16, int8_t, >=)
> +DEF_OP_VV (ge, 32, int8_t, >=)
> +DEF_OP_VV (ge, 64, int8_t, >=)
> +DEF_OP_VV (ge, 128, int8_t, >=)
> +DEF_OP_VV (ge, 256, int8_t, >=)
> +DEF_OP_VV (ge, 512, int8_t, >=)
> +DEF_OP_VV (ge, 1024, int8_t, >=)
> +DEF_OP_VV (ge, 2048, int8_t, >=)
> +DEF_OP_VV (ge, 4096, int8_t, >=)
> +
> +DEF_OP_VV (ge, 1, int16_t, >=)
> +DEF_OP_VV (ge, 2, int16_t, >=)
> +DEF_OP_VV (ge, 4, int16_t, >=)
> +DEF_OP_VV (ge, 8, int16_t, >=)
> +DEF_OP_VV (ge, 16, int16_t, >=)
> +DEF_OP_VV (ge, 32, int16_t, >=)
> +DEF_OP_VV (ge, 64, int16_t, >=)
> +DEF_OP_VV (ge, 128, int16_t, >=)
> +DEF_OP_VV (ge, 256, int16_t, >=)
> +DEF_OP_VV (ge, 512, int16_t, >=)
> +DEF_OP_VV (ge, 1024, int16_t, >=)
> +DEF_OP_VV (ge, 2048, int16_t, >=)
> +
> +DEF_OP_VV (ge, 1, int32_t, >=)
> +DEF_OP_VV (ge, 2, int32_t, >=)
> +DEF_OP_VV (ge, 4, int32_t, >=)
> +DEF_OP_VV (ge, 8, int32_t, >=)
> +DEF_OP_VV (ge, 16, int32_t, >=)
> +DEF_OP_VV (ge, 32, int32_t, >=)
> +DEF_OP_VV (ge, 64, int32_t, >=)
> +DEF_OP_VV (ge, 128, int32_t, >=)
> +DEF_OP_VV (ge, 256, int32_t, >=)
> +DEF_OP_VV (ge, 512, int32_t, >=)
> +DEF_OP_VV (ge, 1024, int32_t, >=)
> +
> +DEF_OP_VV (ge, 1, int64_t, >=)
> +DEF_OP_VV (ge, 2, int64_t, >=)
> +DEF_OP_VV (ge, 4, int64_t, >=)
> +DEF_OP_VV (ge, 8, int64_t, >=)
> +DEF_OP_VV (ge, 16, int64_t, >=)
> +DEF_OP_VV (ge, 32, int64_t, >=)
> +DEF_OP_VV (ge, 64, int64_t, >=)
> +DEF_OP_VV (ge, 128, int64_t, >=)
> +DEF_OP_VV (ge, 256, int64_t, >=)
> +DEF_OP_VV (ge, 512, int64_t, >=)
> +
> +DEF_OP_VV (ge, 1, _Float16, >=)
> +DEF_OP_VV (ge, 2, _Float16, >=)
> +DEF_OP_VV (ge, 4, _Float16, >=)
> +DEF_OP_VV (ge, 8, _Float16, >=)
> +DEF_OP_VV (ge, 16, _Float16, >=)
> +DEF_OP_VV (ge, 32, _Float16, >=)
> +DEF_OP_VV (ge, 64, _Float16, >=)
> +DEF_OP_VV (ge, 128, _Float16, >=)
> +DEF_OP_VV (ge, 256, _Float16, >=)
> +DEF_OP_VV (ge, 512, _Float16, >=)
> +DEF_OP_VV (ge, 1024, _Float16, >=)
> +DEF_OP_VV (ge, 2048, _Float16, >=)
> +
> +DEF_OP_VV (ge, 1, float, >=)
> +DEF_OP_VV (ge, 2, float, >=)
> +DEF_OP_VV (ge, 4, float, >=)
> +DEF_OP_VV (ge, 8, float, >=)
> +DEF_OP_VV (ge, 16, float, >=)
> +DEF_OP_VV (ge, 32, float, >=)
> +DEF_OP_VV (ge, 64, float, >=)
> +DEF_OP_VV (ge, 128, float, >=)
> +DEF_OP_VV (ge, 256, float, >=)
> +DEF_OP_VV (ge, 512, float, >=)
> +DEF_OP_VV (ge, 1024, float, >=)
> +
> +DEF_OP_VV (ge, 1, double, >=)
> +DEF_OP_VV (ge, 2, double, >=)
> +DEF_OP_VV (ge, 4, double, >=)
> +DEF_OP_VV (ge, 8, double, >=)
> +DEF_OP_VV (ge, 16, double, >=)
> +DEF_OP_VV (ge, 32, double, >=)
> +DEF_OP_VV (ge, 64, double, >=)
> +DEF_OP_VV (ge, 128, double, >=)
> +DEF_OP_VV (ge, 256, double, >=)
> +DEF_OP_VV (ge, 512, double, >=)
> +
> +/* { dg-final { scan-assembler-times {vmsge\.vv} 40 } } */
> +/* { dg-final { scan-assembler-times {vmfge\.vv} 27 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-5.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-5.c
> new file mode 100644
> index 00000000000..378b704d360
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-5.c
> @@ -0,0 +1,106 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_OP_VV (lt, 1, int8_t, <)
> +DEF_OP_VV (lt, 2, int8_t, <)
> +DEF_OP_VV (lt, 4, int8_t, <)
> +DEF_OP_VV (lt, 8, int8_t, <)
> +DEF_OP_VV (lt, 16, int8_t, <)
> +DEF_OP_VV (lt, 32, int8_t, <)
> +DEF_OP_VV (lt, 64, int8_t, <)
> +DEF_OP_VV (lt, 128, int8_t, <)
> +DEF_OP_VV (lt, 256, int8_t, <)
> +DEF_OP_VV (lt, 512, int8_t, <)
> +DEF_OP_VV (lt, 1024, int8_t, <)
> +DEF_OP_VV (lt, 2048, int8_t, <)
> +DEF_OP_VV (lt, 4096, int8_t, <)
> +
> +DEF_OP_VV (lt, 1, int16_t, <)
> +DEF_OP_VV (lt, 2, int16_t, <)
> +DEF_OP_VV (lt, 4, int16_t, <)
> +DEF_OP_VV (lt, 8, int16_t, <)
> +DEF_OP_VV (lt, 16, int16_t, <)
> +DEF_OP_VV (lt, 32, int16_t, <)
> +DEF_OP_VV (lt, 64, int16_t, <)
> +DEF_OP_VV (lt, 128, int16_t, <)
> +DEF_OP_VV (lt, 256, int16_t, <)
> +DEF_OP_VV (lt, 512, int16_t, <)
> +DEF_OP_VV (lt, 1024, int16_t, <)
> +DEF_OP_VV (lt, 2048, int16_t, <)
> +
> +DEF_OP_VV (lt, 1, int32_t, <)
> +DEF_OP_VV (lt, 2, int32_t, <)
> +DEF_OP_VV (lt, 4, int32_t, <)
> +DEF_OP_VV (lt, 8, int32_t, <)
> +DEF_OP_VV (lt, 16, int32_t, <)
> +DEF_OP_VV (lt, 32, int32_t, <)
> +DEF_OP_VV (lt, 64, int32_t, <)
> +DEF_OP_VV (lt, 128, int32_t, <)
> +DEF_OP_VV (lt, 256, int32_t, <)
> +DEF_OP_VV (lt, 512, int32_t, <)
> +DEF_OP_VV (lt, 1024, int32_t, <)
> +
> +DEF_OP_VV (lt, 1, int64_t, <)
> +DEF_OP_VV (lt, 2, int64_t, <)
> +DEF_OP_VV (lt, 4, int64_t, <)
> +DEF_OP_VV (lt, 8, int64_t, <)
> +DEF_OP_VV (lt, 16, int64_t, <)
> +DEF_OP_VV (lt, 32, int64_t, <)
> +DEF_OP_VV (lt, 64, int64_t, <)
> +DEF_OP_VV (lt, 128, int64_t, <)
> +DEF_OP_VV (lt, 256, int64_t, <)
> +DEF_OP_VV (lt, 512, int64_t, <)
> +
> +DEF_OP_VV (lt, 1, _Float16, <)
> +DEF_OP_VV (lt, 2, _Float16, <)
> +DEF_OP_VV (lt, 4, _Float16, <)
> +DEF_OP_VV (lt, 8, _Float16, <)
> +DEF_OP_VV (lt, 16, _Float16, <)
> +DEF_OP_VV (lt, 32, _Float16, <)
> +DEF_OP_VV (lt, 64, _Float16, <)
> +DEF_OP_VV (lt, 128, _Float16, <)
> +DEF_OP_VV (lt, 256, _Float16, <)
> +DEF_OP_VV (lt, 512, _Float16, <)
> +DEF_OP_VV (lt, 1024, _Float16, <)
> +DEF_OP_VV (lt, 2048, _Float16, <)
> +
> +DEF_OP_VV (lt, 1, float, <)
> +DEF_OP_VV (lt, 2, float, <)
> +DEF_OP_VV (lt, 4, float, <)
> +DEF_OP_VV (lt, 8, float, <)
> +DEF_OP_VV (lt, 16, float, <)
> +DEF_OP_VV (lt, 32, float, <)
> +DEF_OP_VV (lt, 64, float, <)
> +DEF_OP_VV (lt, 128, float, <)
> +DEF_OP_VV (lt, 256, float, <)
> +DEF_OP_VV (lt, 512, float, <)
> +DEF_OP_VV (lt, 1024, float, <)
> +
> +DEF_OP_VV (lt, 1, double, <)
> +DEF_OP_VV (lt, 2, double, <)
> +DEF_OP_VV (lt, 4, double, <)
> +DEF_OP_VV (lt, 8, double, <)
> +DEF_OP_VV (lt, 16, double, <)
> +DEF_OP_VV (lt, 32, double, <)
> +DEF_OP_VV (lt, 64, double, <)
> +DEF_OP_VV (lt, 128, double, <)
> +DEF_OP_VV (lt, 256, double, <)
> +DEF_OP_VV (lt, 512, double, <)
> +
> +/* { dg-final { scan-assembler-times {vmslt\.vv} 40 } } */
> +/* { dg-final { scan-assembler-times {vmflt\.vv} 27 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-6.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-6.c
> new file mode 100644
> index 00000000000..f0351e0baf8
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/cmp-6.c
> @@ -0,0 +1,106 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_OP_VV (le, 1, int8_t, <=)
> +DEF_OP_VV (le, 2, int8_t, <=)
> +DEF_OP_VV (le, 4, int8_t, <=)
> +DEF_OP_VV (le, 8, int8_t, <=)
> +DEF_OP_VV (le, 16, int8_t, <=)
> +DEF_OP_VV (le, 32, int8_t, <=)
> +DEF_OP_VV (le, 64, int8_t, <=)
> +DEF_OP_VV (le, 128, int8_t, <=)
> +DEF_OP_VV (le, 256, int8_t, <=)
> +DEF_OP_VV (le, 512, int8_t, <=)
> +DEF_OP_VV (le, 1024, int8_t, <=)
> +DEF_OP_VV (le, 2048, int8_t, <=)
> +DEF_OP_VV (le, 4096, int8_t, <=)
> +
> +DEF_OP_VV (le, 1, int16_t, <=)
> +DEF_OP_VV (le, 2, int16_t, <=)
> +DEF_OP_VV (le, 4, int16_t, <=)
> +DEF_OP_VV (le, 8, int16_t, <=)
> +DEF_OP_VV (le, 16, int16_t, <=)
> +DEF_OP_VV (le, 32, int16_t, <=)
> +DEF_OP_VV (le, 64, int16_t, <=)
> +DEF_OP_VV (le, 128, int16_t, <=)
> +DEF_OP_VV (le, 256, int16_t, <=)
> +DEF_OP_VV (le, 512, int16_t, <=)
> +DEF_OP_VV (le, 1024, int16_t, <=)
> +DEF_OP_VV (le, 2048, int16_t, <=)
> +
> +DEF_OP_VV (le, 1, int32_t, <=)
> +DEF_OP_VV (le, 2, int32_t, <=)
> +DEF_OP_VV (le, 4, int32_t, <=)
> +DEF_OP_VV (le, 8, int32_t, <=)
> +DEF_OP_VV (le, 16, int32_t, <=)
> +DEF_OP_VV (le, 32, int32_t, <=)
> +DEF_OP_VV (le, 64, int32_t, <=)
> +DEF_OP_VV (le, 128, int32_t, <=)
> +DEF_OP_VV (le, 256, int32_t, <=)
> +DEF_OP_VV (le, 512, int32_t, <=)
> +DEF_OP_VV (le, 1024, int32_t, <=)
> +
> +DEF_OP_VV (le, 1, int64_t, <=)
> +DEF_OP_VV (le, 2, int64_t, <=)
> +DEF_OP_VV (le, 4, int64_t, <=)
> +DEF_OP_VV (le, 8, int64_t, <=)
> +DEF_OP_VV (le, 16, int64_t, <=)
> +DEF_OP_VV (le, 32, int64_t, <=)
> +DEF_OP_VV (le, 64, int64_t, <=)
> +DEF_OP_VV (le, 128, int64_t, <=)
> +DEF_OP_VV (le, 256, int64_t, <=)
> +DEF_OP_VV (le, 512, int64_t, <=)
> +
> +DEF_OP_VV (le, 1, _Float16, <=)
> +DEF_OP_VV (le, 2, _Float16, <=)
> +DEF_OP_VV (le, 4, _Float16, <=)
> +DEF_OP_VV (le, 8, _Float16, <=)
> +DEF_OP_VV (le, 16, _Float16, <=)
> +DEF_OP_VV (le, 32, _Float16, <=)
> +DEF_OP_VV (le, 64, _Float16, <=)
> +DEF_OP_VV (le, 128, _Float16, <=)
> +DEF_OP_VV (le, 256, _Float16, <=)
> +DEF_OP_VV (le, 512, _Float16, <=)
> +DEF_OP_VV (le, 1024, _Float16, <=)
> +DEF_OP_VV (le, 2048, _Float16, <=)
> +
> +DEF_OP_VV (le, 1, float, <=)
> +DEF_OP_VV (le, 2, float, <=)
> +DEF_OP_VV (le, 4, float, <=)
> +DEF_OP_VV (le, 8, float, <=)
> +DEF_OP_VV (le, 16, float, <=)
> +DEF_OP_VV (le, 32, float, <=)
> +DEF_OP_VV (le, 64, float, <=)
> +DEF_OP_VV (le, 128, float, <=)
> +DEF_OP_VV (le, 256, float, <=)
> +DEF_OP_VV (le, 512, float, <=)
> +DEF_OP_VV (le, 1024, float, <=)
> +
> +DEF_OP_VV (le, 1, double, <=)
> +DEF_OP_VV (le, 2, double, <=)
> +DEF_OP_VV (le, 4, double, <=)
> +DEF_OP_VV (le, 8, double, <=)
> +DEF_OP_VV (le, 16, double, <=)
> +DEF_OP_VV (le, 32, double, <=)
> +DEF_OP_VV (le, 64, double, <=)
> +DEF_OP_VV (le, 128, double, <=)
> +DEF_OP_VV (le, 256, double, <=)
> +DEF_OP_VV (le, 512, double, <=)
> +
> +/* { dg-final { scan-assembler-times {vmsle\.vv} 40 } } */
> +/* { dg-final { scan-assembler-times {vmfle\.vv} 27 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/def.h
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/def.h
> index 79b4fbc6d93..1edc1910920 100644
> --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/def.h
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/def.h
> @@ -249,3 +249,12 @@ typedef double v512df __attribute__ ((vector_size
> (4096)));
> {
> \
> return v[INDEX];
> \
> }
> +
> +#define DEF_MASK_LOGIC(PREFIX, NUM, TYPE, OP)
> \
> + void __attribute__ ((noinline, noclone))
> \
> + PREFIX##_##TYPE##NUM (TYPE *restrict a, TYPE *restrict b, TYPE
> *restrict c, \
> + TYPE *restrict d, TYPE *restrict e)
> \
> + {
> \
> + for (int i = 0; i < NUM; ++i)
> \
> + a[i] = (b[i] > c[i]) OP (d[i] < e[i]);
> \
> + }
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-1.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-1.c
> new file mode 100644
> index 00000000000..d5b65ff5500
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-1.c
> @@ -0,0 +1,69 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_MASK_LOGIC (and, 1, int8_t, &)
> +DEF_MASK_LOGIC (and, 2, int8_t, &)
> +DEF_MASK_LOGIC (and, 4, int8_t, &)
> +DEF_MASK_LOGIC (and, 8, int8_t, &)
> +DEF_MASK_LOGIC (and, 16, int8_t, &)
> +DEF_MASK_LOGIC (and, 32, int8_t, &)
> +DEF_MASK_LOGIC (and, 64, int8_t, &)
> +DEF_MASK_LOGIC (and, 128, int8_t, &)
> +DEF_MASK_LOGIC (and, 256, int8_t, &)
> +DEF_MASK_LOGIC (and, 512, int8_t, &)
> +DEF_MASK_LOGIC (and, 1024, int8_t, &)
> +DEF_MASK_LOGIC (and, 2048, int8_t, &)
> +DEF_MASK_LOGIC (and, 4096, int8_t, &)
> +
> +DEF_MASK_LOGIC (and, 1, int16_t, &)
> +DEF_MASK_LOGIC (and, 2, int16_t, &)
> +DEF_MASK_LOGIC (and, 4, int16_t, &)
> +DEF_MASK_LOGIC (and, 8, int16_t, &)
> +DEF_MASK_LOGIC (and, 16, int16_t, &)
> +DEF_MASK_LOGIC (and, 32, int16_t, &)
> +DEF_MASK_LOGIC (and, 64, int16_t, &)
> +DEF_MASK_LOGIC (and, 128, int16_t, &)
> +DEF_MASK_LOGIC (and, 256, int16_t, &)
> +DEF_MASK_LOGIC (and, 512, int16_t, &)
> +DEF_MASK_LOGIC (and, 1024, int16_t, &)
> +DEF_MASK_LOGIC (and, 2048, int16_t, &)
> +
> +DEF_MASK_LOGIC (and, 1, int32_t, &)
> +DEF_MASK_LOGIC (and, 2, int32_t, &)
> +DEF_MASK_LOGIC (and, 4, int32_t, &)
> +DEF_MASK_LOGIC (and, 8, int32_t, &)
> +DEF_MASK_LOGIC (and, 16, int32_t, &)
> +DEF_MASK_LOGIC (and, 32, int32_t, &)
> +DEF_MASK_LOGIC (and, 64, int32_t, &)
> +DEF_MASK_LOGIC (and, 128, int32_t, &)
> +DEF_MASK_LOGIC (and, 256, int32_t, &)
> +DEF_MASK_LOGIC (and, 512, int32_t, &)
> +DEF_MASK_LOGIC (and, 1024, int32_t, &)
> +
> +DEF_MASK_LOGIC (and, 1, int64_t, &)
> +DEF_MASK_LOGIC (and, 2, int64_t, &)
> +DEF_MASK_LOGIC (and, 4, int64_t, &)
> +DEF_MASK_LOGIC (and, 8, int64_t, &)
> +DEF_MASK_LOGIC (and, 16, int64_t, &)
> +DEF_MASK_LOGIC (and, 32, int64_t, &)
> +DEF_MASK_LOGIC (and, 64, int64_t, &)
> +DEF_MASK_LOGIC (and, 128, int64_t, &)
> +DEF_MASK_LOGIC (and, 256, int64_t, &)
> +DEF_MASK_LOGIC (and, 512, int64_t, &)
> +
> +/* { dg-final { scan-assembler-times {vmand\.mm} 42 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-2.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-2.c
> new file mode 100644
> index 00000000000..7fe4ec96dc8
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-2.c
> @@ -0,0 +1,69 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_MASK_LOGIC (and, 1, int8_t, |)
> +DEF_MASK_LOGIC (and, 2, int8_t, |)
> +DEF_MASK_LOGIC (and, 4, int8_t, |)
> +DEF_MASK_LOGIC (and, 8, int8_t, |)
> +DEF_MASK_LOGIC (and, 16, int8_t, |)
> +DEF_MASK_LOGIC (and, 32, int8_t, |)
> +DEF_MASK_LOGIC (and, 64, int8_t, |)
> +DEF_MASK_LOGIC (and, 128, int8_t, |)
> +DEF_MASK_LOGIC (and, 256, int8_t, |)
> +DEF_MASK_LOGIC (and, 512, int8_t, |)
> +DEF_MASK_LOGIC (and, 1024, int8_t, |)
> +DEF_MASK_LOGIC (and, 2048, int8_t, |)
> +DEF_MASK_LOGIC (and, 4096, int8_t, |)
> +
> +DEF_MASK_LOGIC (and, 1, int16_t, |)
> +DEF_MASK_LOGIC (and, 2, int16_t, |)
> +DEF_MASK_LOGIC (and, 4, int16_t, |)
> +DEF_MASK_LOGIC (and, 8, int16_t, |)
> +DEF_MASK_LOGIC (and, 16, int16_t, |)
> +DEF_MASK_LOGIC (and, 32, int16_t, |)
> +DEF_MASK_LOGIC (and, 64, int16_t, |)
> +DEF_MASK_LOGIC (and, 128, int16_t, |)
> +DEF_MASK_LOGIC (and, 256, int16_t, |)
> +DEF_MASK_LOGIC (and, 512, int16_t, |)
> +DEF_MASK_LOGIC (and, 1024, int16_t, |)
> +DEF_MASK_LOGIC (and, 2048, int16_t, |)
> +
> +DEF_MASK_LOGIC (and, 1, int32_t, |)
> +DEF_MASK_LOGIC (and, 2, int32_t, |)
> +DEF_MASK_LOGIC (and, 4, int32_t, |)
> +DEF_MASK_LOGIC (and, 8, int32_t, |)
> +DEF_MASK_LOGIC (and, 16, int32_t, |)
> +DEF_MASK_LOGIC (and, 32, int32_t, |)
> +DEF_MASK_LOGIC (and, 64, int32_t, |)
> +DEF_MASK_LOGIC (and, 128, int32_t, |)
> +DEF_MASK_LOGIC (and, 256, int32_t, |)
> +DEF_MASK_LOGIC (and, 512, int32_t, |)
> +DEF_MASK_LOGIC (and, 1024, int32_t, |)
> +
> +DEF_MASK_LOGIC (and, 1, int64_t, |)
> +DEF_MASK_LOGIC (and, 2, int64_t, |)
> +DEF_MASK_LOGIC (and, 4, int64_t, |)
> +DEF_MASK_LOGIC (and, 8, int64_t, |)
> +DEF_MASK_LOGIC (and, 16, int64_t, |)
> +DEF_MASK_LOGIC (and, 32, int64_t, |)
> +DEF_MASK_LOGIC (and, 64, int64_t, |)
> +DEF_MASK_LOGIC (and, 128, int64_t, |)
> +DEF_MASK_LOGIC (and, 256, int64_t, |)
> +DEF_MASK_LOGIC (and, 512, int64_t, |)
> +
> +/* { dg-final { scan-assembler-times {vmor\.mm} 42 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-3.c
> b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-3.c
> new file mode 100644
> index 00000000000..7d2254b326d
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/vls/mask-3.c
> @@ -0,0 +1,69 @@
> +/* { dg-do compile } */
> +/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3
> -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8
> -fdump-tree-optimized" } */
> +
> +#include "def.h"
> +
> +DEF_MASK_LOGIC (and, 1, int8_t, ^)
> +DEF_MASK_LOGIC (and, 2, int8_t, ^)
> +DEF_MASK_LOGIC (and, 4, int8_t, ^)
> +DEF_MASK_LOGIC (and, 8, int8_t, ^)
> +DEF_MASK_LOGIC (and, 16, int8_t, ^)
> +DEF_MASK_LOGIC (and, 32, int8_t, ^)
> +DEF_MASK_LOGIC (and, 64, int8_t, ^)
> +DEF_MASK_LOGIC (and, 128, int8_t, ^)
> +DEF_MASK_LOGIC (and, 256, int8_t, ^)
> +DEF_MASK_LOGIC (and, 512, int8_t, ^)
> +DEF_MASK_LOGIC (and, 1024, int8_t, ^)
> +DEF_MASK_LOGIC (and, 2048, int8_t, ^)
> +DEF_MASK_LOGIC (and, 4096, int8_t, ^)
> +
> +DEF_MASK_LOGIC (and, 1, int16_t, ^)
> +DEF_MASK_LOGIC (and, 2, int16_t, ^)
> +DEF_MASK_LOGIC (and, 4, int16_t, ^)
> +DEF_MASK_LOGIC (and, 8, int16_t, ^)
> +DEF_MASK_LOGIC (and, 16, int16_t, ^)
> +DEF_MASK_LOGIC (and, 32, int16_t, ^)
> +DEF_MASK_LOGIC (and, 64, int16_t, ^)
> +DEF_MASK_LOGIC (and, 128, int16_t, ^)
> +DEF_MASK_LOGIC (and, 256, int16_t, ^)
> +DEF_MASK_LOGIC (and, 512, int16_t, ^)
> +DEF_MASK_LOGIC (and, 1024, int16_t, ^)
> +DEF_MASK_LOGIC (and, 2048, int16_t, ^)
> +
> +DEF_MASK_LOGIC (and, 1, int32_t, ^)
> +DEF_MASK_LOGIC (and, 2, int32_t, ^)
> +DEF_MASK_LOGIC (and, 4, int32_t, ^)
> +DEF_MASK_LOGIC (and, 8, int32_t, ^)
> +DEF_MASK_LOGIC (and, 16, int32_t, ^)
> +DEF_MASK_LOGIC (and, 32, int32_t, ^)
> +DEF_MASK_LOGIC (and, 64, int32_t, ^)
> +DEF_MASK_LOGIC (and, 128, int32_t, ^)
> +DEF_MASK_LOGIC (and, 256, int32_t, ^)
> +DEF_MASK_LOGIC (and, 512, int32_t, ^)
> +DEF_MASK_LOGIC (and, 1024, int32_t, ^)
> +
> +DEF_MASK_LOGIC (and, 1, int64_t, ^)
> +DEF_MASK_LOGIC (and, 2, int64_t, ^)
> +DEF_MASK_LOGIC (and, 4, int64_t, ^)
> +DEF_MASK_LOGIC (and, 8, int64_t, ^)
> +DEF_MASK_LOGIC (and, 16, int64_t, ^)
> +DEF_MASK_LOGIC (and, 32, int64_t, ^)
> +DEF_MASK_LOGIC (and, 64, int64_t, ^)
> +DEF_MASK_LOGIC (and, 128, int64_t, ^)
> +DEF_MASK_LOGIC (and, 256, int64_t, ^)
> +DEF_MASK_LOGIC (and, 512, int64_t, ^)
> +
> +/* { dg-final { scan-assembler-times {vmxor\.mm} 42 } } */
> +/* { dg-final { scan-assembler-not {csrr} } } */
> +/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
> --
> 2.36.3
>
>
@@ -67,10 +67,10 @@
;; -------------------------------------------------------------------------
(define_insn_and_split "*<optab>not<mode>"
- [(set (match_operand:VB 0 "register_operand" "=vr")
- (bitmanip_bitwise:VB
- (not:VB (match_operand:VB 2 "register_operand" " vr"))
- (match_operand:VB 1 "register_operand" " vr")))]
+ [(set (match_operand:VB_VLS 0 "register_operand" "=vr")
+ (bitmanip_bitwise:VB_VLS
+ (not:VB_VLS (match_operand:VB_VLS 2 "register_operand" " vr"))
+ (match_operand:VB_VLS 1 "register_operand" " vr")))]
"TARGET_VECTOR && can_create_pseudo_p ()"
"#"
"&& 1"
@@ -93,11 +93,11 @@
;; -------------------------------------------------------------------------
(define_insn_and_split "*n<optab><mode>"
- [(set (match_operand:VB 0 "register_operand" "=vr")
- (not:VB
- (any_bitwise:VB
- (match_operand:VB 1 "register_operand" " vr")
- (match_operand:VB 2 "register_operand" " vr"))))]
+ [(set (match_operand:VB_VLS 0 "register_operand" "=vr")
+ (not:VB_VLS
+ (any_bitwise:VB_VLS
+ (match_operand:VB_VLS 1 "register_operand" " vr")
+ (match_operand:VB_VLS 2 "register_operand" " vr"))))]
"TARGET_VECTOR && can_create_pseudo_p ()"
"#"
"&& 1"
@@ -495,9 +495,9 @@
;; -------------------------------------------------------------------------
(define_insn_and_split "<optab><mode>3"
- [(set (match_operand:VB 0 "register_operand" "=vr")
- (any_bitwise:VB (match_operand:VB 1 "register_operand" " vr")
- (match_operand:VB 2 "register_operand" " vr")))]
+ [(set (match_operand:VB_VLS 0 "register_operand" "=vr")
+ (any_bitwise:VB_VLS (match_operand:VB_VLS 1 "register_operand" " vr")
+ (match_operand:VB_VLS 2 "register_operand" " vr")))]
"TARGET_VECTOR && can_create_pseudo_p ()"
"#"
"&& 1"
@@ -518,8 +518,8 @@
;; -------------------------------------------------------------------------
(define_insn_and_split "one_cmpl<mode>2"
- [(set (match_operand:VB 0 "register_operand" "=vr")
- (not:VB (match_operand:VB 1 "register_operand" " vr")))]
+ [(set (match_operand:VB_VLS 0 "register_operand" "=vr")
+ (not:VB_VLS (match_operand:VB_VLS 1 "register_operand" " vr")))]
"TARGET_VECTOR && can_create_pseudo_p ()"
"#"
"&& 1"
@@ -545,12 +545,12 @@
;; - vfmerge.vf
;; -------------------------------------------------------------------------
-(define_insn_and_split "@vcond_mask_<mode><vm>"
- [(set (match_operand:V 0 "register_operand")
- (if_then_else:V
+(define_insn_and_split "vcond_mask_<mode><vm>"
+ [(set (match_operand:V_VLS 0 "register_operand")
+ (if_then_else:V_VLS
(match_operand:<VM> 3 "register_operand")
- (match_operand:V 1 "nonmemory_operand")
- (match_operand:V 2 "register_operand")))]
+ (match_operand:V_VLS 1 "nonmemory_operand")
+ (match_operand:V_VLS 2 "register_operand")))]
"TARGET_VECTOR && can_create_pseudo_p ()"
"#"
"&& 1"
@@ -609,8 +609,8 @@
(define_expand "vec_cmp<mode><vm>"
[(set (match_operand:<VM> 0 "register_operand")
(match_operator:<VM> 1 "comparison_operator"
- [(match_operand:VI 2 "register_operand")
- (match_operand:VI 3 "register_operand")]))]
+ [(match_operand:V_VLSI 2 "register_operand")
+ (match_operand:V_VLSI 3 "register_operand")]))]
"TARGET_VECTOR"
{
riscv_vector::expand_vec_cmp (operands[0], GET_CODE (operands[1]),
@@ -622,8 +622,8 @@
(define_expand "vec_cmpu<mode><vm>"
[(set (match_operand:<VM> 0 "register_operand")
(match_operator:<VM> 1 "comparison_operator"
- [(match_operand:VI 2 "register_operand")
- (match_operand:VI 3 "register_operand")]))]
+ [(match_operand:V_VLSI 2 "register_operand")
+ (match_operand:V_VLSI 3 "register_operand")]))]
"TARGET_VECTOR"
{
riscv_vector::expand_vec_cmp (operands[0], GET_CODE (operands[1]),
@@ -635,8 +635,8 @@
(define_expand "vec_cmp<mode><vm>"
[(set (match_operand:<VM> 0 "register_operand")
(match_operator:<VM> 1 "comparison_operator"
- [(match_operand:VF 2 "register_operand")
- (match_operand:VF 3 "register_operand")]))]
+ [(match_operand:V_VLSF 2 "register_operand")
+ (match_operand:V_VLSF 3 "register_operand")]))]
"TARGET_VECTOR"
{
riscv_vector::expand_vec_cmp_float (operands[0], GET_CODE (operands[1]),
@@ -7177,8 +7177,8 @@
(reg:SI VL_REGNUM)
(reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
(match_operator:<VM> 3 "signed_order_operator"
- [(match_operand:VF 4 "register_operand")
- (match_operand:VF 5 "register_operand")])
+ [(match_operand:V_VLSF 4 "register_operand")
+ (match_operand:V_VLSF 5 "register_operand")])
(match_operand:<VM> 2 "vector_merge_operand")))]
"TARGET_VECTOR"
{})
@@ -7195,10 +7195,10 @@
(reg:SI VL_REGNUM)
(reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
(match_operator:<VM> 3 "signed_order_operator"
- [(match_operand:VF 4 "register_operand" " vr, vr")
- (match_operand:VF 5 "register_operand" " vr, vr")])
+ [(match_operand:V_VLSF 4 "register_operand" " vr, vr")
+ (match_operand:V_VLSF 5 "register_operand" " vr, vr")])
(match_operand:<VM> 2 "vector_merge_operand" " vu, 0")))]
- "TARGET_VECTOR && known_le (GET_MODE_SIZE (<MODE>mode), BYTES_PER_RISCV_VECTOR)"
+ "TARGET_VECTOR && riscv_vector::cmp_lmul_le_one (<MODE>mode)"
"vmf%B3.vv\t%0,%4,%5%p1"
[(set_attr "type" "vfcmp")
(set_attr "mode" "<MODE>")])
@@ -7214,8 +7214,8 @@
(reg:SI VL_REGNUM)
(reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
(match_operator:<VM> 2 "signed_order_operator"
- [(match_operand:VF 3 "register_operand" " vr")
- (match_operand:VF 4 "register_operand" " vr")])
+ [(match_operand:V_VLSF 3 "register_operand" " vr")
+ (match_operand:V_VLSF 4 "register_operand" " vr")])
(match_dup 1)))]
"TARGET_VECTOR"
"vmf%B2.vv\t%0,%3,%4,v0.t"
@@ -7238,10 +7238,10 @@
(reg:SI VL_REGNUM)
(reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
(match_operator:<VM> 3 "signed_order_operator"
- [(match_operand:VF 4 "register_operand" " vr, 0, vr, 0, 0, vr, 0, vr, vr")
- (match_operand:VF 5 "register_operand" " vr, vr, 0, 0, vr, 0, 0, vr, vr")])
+ [(match_operand:V_VLSF 4 "register_operand" " vr, 0, vr, 0, 0, vr, 0, vr, vr")
+ (match_operand:V_VLSF 5 "register_operand" " vr, vr, 0, 0, vr, 0, 0, vr, vr")])
(match_operand:<VM> 2 "vector_merge_operand" " vu, vu, vu, vu, 0, 0, 0, vu, 0")))]
- "TARGET_VECTOR && known_gt (GET_MODE_SIZE (<MODE>mode), BYTES_PER_RISCV_VECTOR)"
+ "TARGET_VECTOR && riscv_vector::cmp_lmul_gt_one (<MODE>mode)"
"vmf%B3.vv\t%0,%4,%5%p1"
[(set_attr "type" "vfcmp")
(set_attr "mode" "<MODE>")])
@@ -7257,8 +7257,8 @@
(reg:SI VL_REGNUM)
(reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
(match_operator:<VM> 3 "signed_order_operator"
- [(match_operand:VF 4 "register_operand")
- (vec_duplicate:VF
+ [(match_operand:V_VLSF 4 "register_operand")
+ (vec_duplicate:V_VLSF
(match_operand:<VEL> 5 "register_operand"))])
(match_operand:<VM> 2 "vector_merge_operand")))]
"TARGET_VECTOR"
@@ -7275,8 +7275,8 @@
(reg:SI VL_REGNUM)
(reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
(match_operator:<VM> 2 "signed_order_operator"
- [(match_operand:VF 3 "register_operand" " vr")
- (vec_duplicate:VF
+ [(match_operand:V_VLSF 3 "register_operand" " vr")
+ (vec_duplicate:V_VLSF
(match_operand:<VEL> 4 "register_operand" " f"))])
(match_dup 1)))]
"TARGET_VECTOR"
@@ -7300,11 +7300,11 @@
(reg:SI VL_REGNUM)
(reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
(match_operator:<VM> 3 "signed_order_operator"
- [(match_operand:VF 4 "register_operand" " vr, vr")
- (vec_duplicate:VF
+ [(match_operand:V_VLSF 4 "register_operand" " vr, vr")
+ (vec_duplicate:V_VLSF
(match_operand:<VEL> 5 "register_operand" " f, f"))])
(match_operand:<VM> 2 "vector_merge_operand" " vu, 0")))]
- "TARGET_VECTOR && known_le (GET_MODE_SIZE (<MODE>mode), BYTES_PER_RISCV_VECTOR)"
+ "TARGET_VECTOR && riscv_vector::cmp_lmul_le_one (<MODE>mode)"
"vmf%B3.vf\t%0,%4,%5%p1"
[(set_attr "type" "vfcmp")
(set_attr "mode" "<MODE>")])
@@ -7321,11 +7321,11 @@
(reg:SI VL_REGNUM)
(reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
(match_operator:<VM> 3 "signed_order_operator"
- [(match_operand:VF 4 "register_operand" " vr, 0, 0, vr, vr")
- (vec_duplicate:VF
+ [(match_operand:V_VLSF 4 "register_operand" " vr, 0, 0, vr, vr")
+ (vec_duplicate:V_VLSF
(match_operand:<VEL> 5 "register_operand" " f, f, f, f, f"))])
(match_operand:<VM> 2 "vector_merge_operand" " vu, vu, 0, vu, 0")))]
- "TARGET_VECTOR && known_gt (GET_MODE_SIZE (<MODE>mode), BYTES_PER_RISCV_VECTOR)"
+ "TARGET_VECTOR && riscv_vector::cmp_lmul_gt_one (<MODE>mode)"
"vmf%B3.vf\t%0,%4,%5%p1"
[(set_attr "type" "vfcmp")
(set_attr "mode" "<MODE>")])
@@ -7341,9 +7341,9 @@
(reg:SI VL_REGNUM)
(reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
(match_operator:<VM> 3 "equality_operator"
- [(vec_duplicate:VF
+ [(vec_duplicate:V_VLSF
(match_operand:<VEL> 5 "register_operand"))
- (match_operand:VF 4 "register_operand")])
+ (match_operand:V_VLSF 4 "register_operand")])
(match_operand:<VM> 2 "vector_merge_operand")))]
"TARGET_VECTOR"
{})
@@ -7359,9 +7359,9 @@
(reg:SI VL_REGNUM)
(reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
(match_operator:<VM> 2 "equality_operator"
- [(vec_duplicate:VF
+ [(vec_duplicate:V_VLSF
(match_operand:<VEL> 4 "register_operand" " f"))
- (match_operand:VF 3 "register_operand" " vr")])
+ (match_operand:V_VLSF 3 "register_operand" " vr")])
(match_dup 1)))]
"TARGET_VECTOR"
"vmf%B2.vf\t%0,%3,%4,v0.t"
@@ -7384,11 +7384,11 @@
(reg:SI VL_REGNUM)
(reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
(match_operator:<VM> 3 "equality_operator"
- [(vec_duplicate:VF
+ [(vec_duplicate:V_VLSF
(match_operand:<VEL> 5 "register_operand" " f, f"))
- (match_operand:VF 4 "register_operand" " vr, vr")])
+ (match_operand:V_VLSF 4 "register_operand" " vr, vr")])
(match_operand:<VM> 2 "vector_merge_operand" " vu, 0")))]
- "TARGET_VECTOR && known_le (GET_MODE_SIZE (<MODE>mode), BYTES_PER_RISCV_VECTOR)"
+ "TARGET_VECTOR && riscv_vector::cmp_lmul_le_one (<MODE>mode)"
"vmf%B3.vf\t%0,%4,%5%p1"
[(set_attr "type" "vfcmp")
(set_attr "mode" "<MODE>")])
@@ -7405,11 +7405,11 @@
(reg:SI VL_REGNUM)
(reg:SI VTYPE_REGNUM)] UNSPEC_VPREDICATE)
(match_operator:<VM> 3 "equality_operator"
- [(vec_duplicate:VF
+ [(vec_duplicate:V_VLSF
(match_operand:<VEL> 5 "register_operand" " f, f, f, f, f"))
- (match_operand:VF 4 "register_operand" " vr, 0, 0, vr, vr")])
+ (match_operand:V_VLSF 4 "register_operand" " vr, 0, 0, vr, vr")])
(match_operand:<VM> 2 "vector_merge_operand" " vu, vu, 0, vu, 0")))]
- "TARGET_VECTOR && known_gt (GET_MODE_SIZE (<MODE>mode), BYTES_PER_RISCV_VECTOR)"
+ "TARGET_VECTOR && riscv_vector::cmp_lmul_gt_one (<MODE>mode)"
"vmf%B3.vf\t%0,%4,%5%p1"
[(set_attr "type" "vfcmp")
(set_attr "mode" "<MODE>")])
new file mode 100644
@@ -0,0 +1,106 @@
+/* { dg-do compile } */
+/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3 -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8 -fdump-tree-optimized" } */
+
+#include "def.h"
+
+DEF_OP_VV (eq, 1, int8_t, ==)
+DEF_OP_VV (eq, 2, int8_t, ==)
+DEF_OP_VV (eq, 4, int8_t, ==)
+DEF_OP_VV (eq, 8, int8_t, ==)
+DEF_OP_VV (eq, 16, int8_t, ==)
+DEF_OP_VV (eq, 32, int8_t, ==)
+DEF_OP_VV (eq, 64, int8_t, ==)
+DEF_OP_VV (eq, 128, int8_t, ==)
+DEF_OP_VV (eq, 256, int8_t, ==)
+DEF_OP_VV (eq, 512, int8_t, ==)
+DEF_OP_VV (eq, 1024, int8_t, ==)
+DEF_OP_VV (eq, 2048, int8_t, ==)
+DEF_OP_VV (eq, 4096, int8_t, ==)
+
+DEF_OP_VV (eq, 1, int16_t, ==)
+DEF_OP_VV (eq, 2, int16_t, ==)
+DEF_OP_VV (eq, 4, int16_t, ==)
+DEF_OP_VV (eq, 8, int16_t, ==)
+DEF_OP_VV (eq, 16, int16_t, ==)
+DEF_OP_VV (eq, 32, int16_t, ==)
+DEF_OP_VV (eq, 64, int16_t, ==)
+DEF_OP_VV (eq, 128, int16_t, ==)
+DEF_OP_VV (eq, 256, int16_t, ==)
+DEF_OP_VV (eq, 512, int16_t, ==)
+DEF_OP_VV (eq, 1024, int16_t, ==)
+DEF_OP_VV (eq, 2048, int16_t, ==)
+
+DEF_OP_VV (eq, 1, int32_t, ==)
+DEF_OP_VV (eq, 2, int32_t, ==)
+DEF_OP_VV (eq, 4, int32_t, ==)
+DEF_OP_VV (eq, 8, int32_t, ==)
+DEF_OP_VV (eq, 16, int32_t, ==)
+DEF_OP_VV (eq, 32, int32_t, ==)
+DEF_OP_VV (eq, 64, int32_t, ==)
+DEF_OP_VV (eq, 128, int32_t, ==)
+DEF_OP_VV (eq, 256, int32_t, ==)
+DEF_OP_VV (eq, 512, int32_t, ==)
+DEF_OP_VV (eq, 1024, int32_t, ==)
+
+DEF_OP_VV (eq, 1, int64_t, ==)
+DEF_OP_VV (eq, 2, int64_t, ==)
+DEF_OP_VV (eq, 4, int64_t, ==)
+DEF_OP_VV (eq, 8, int64_t, ==)
+DEF_OP_VV (eq, 16, int64_t, ==)
+DEF_OP_VV (eq, 32, int64_t, ==)
+DEF_OP_VV (eq, 64, int64_t, ==)
+DEF_OP_VV (eq, 128, int64_t, ==)
+DEF_OP_VV (eq, 256, int64_t, ==)
+DEF_OP_VV (eq, 512, int64_t, ==)
+
+DEF_OP_VV (eq, 1, _Float16, ==)
+DEF_OP_VV (eq, 2, _Float16, ==)
+DEF_OP_VV (eq, 4, _Float16, ==)
+DEF_OP_VV (eq, 8, _Float16, ==)
+DEF_OP_VV (eq, 16, _Float16, ==)
+DEF_OP_VV (eq, 32, _Float16, ==)
+DEF_OP_VV (eq, 64, _Float16, ==)
+DEF_OP_VV (eq, 128, _Float16, ==)
+DEF_OP_VV (eq, 256, _Float16, ==)
+DEF_OP_VV (eq, 512, _Float16, ==)
+DEF_OP_VV (eq, 1024, _Float16, ==)
+DEF_OP_VV (eq, 2048, _Float16, ==)
+
+DEF_OP_VV (eq, 1, float, ==)
+DEF_OP_VV (eq, 2, float, ==)
+DEF_OP_VV (eq, 4, float, ==)
+DEF_OP_VV (eq, 8, float, ==)
+DEF_OP_VV (eq, 16, float, ==)
+DEF_OP_VV (eq, 32, float, ==)
+DEF_OP_VV (eq, 64, float, ==)
+DEF_OP_VV (eq, 128, float, ==)
+DEF_OP_VV (eq, 256, float, ==)
+DEF_OP_VV (eq, 512, float, ==)
+DEF_OP_VV (eq, 1024, float, ==)
+
+DEF_OP_VV (eq, 1, double, ==)
+DEF_OP_VV (eq, 2, double, ==)
+DEF_OP_VV (eq, 4, double, ==)
+DEF_OP_VV (eq, 8, double, ==)
+DEF_OP_VV (eq, 16, double, ==)
+DEF_OP_VV (eq, 32, double, ==)
+DEF_OP_VV (eq, 64, double, ==)
+DEF_OP_VV (eq, 128, double, ==)
+DEF_OP_VV (eq, 256, double, ==)
+DEF_OP_VV (eq, 512, double, ==)
+
+/* { dg-final { scan-assembler-times {vmseq\.vv} 42 } } */
+/* { dg-final { scan-assembler-times {vmfeq\.vv} 27 } } */
+/* { dg-final { scan-assembler-not {csrr} } } */
+/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
new file mode 100644
@@ -0,0 +1,106 @@
+/* { dg-do compile } */
+/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3 -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8 -fdump-tree-optimized" } */
+
+#include "def.h"
+
+DEF_OP_VV (ne, 1, int8_t, !=)
+DEF_OP_VV (ne, 2, int8_t, !=)
+DEF_OP_VV (ne, 4, int8_t, !=)
+DEF_OP_VV (ne, 8, int8_t, !=)
+DEF_OP_VV (ne, 16, int8_t, !=)
+DEF_OP_VV (ne, 32, int8_t, !=)
+DEF_OP_VV (ne, 64, int8_t, !=)
+DEF_OP_VV (ne, 128, int8_t, !=)
+DEF_OP_VV (ne, 256, int8_t, !=)
+DEF_OP_VV (ne, 512, int8_t, !=)
+DEF_OP_VV (ne, 1024, int8_t, !=)
+DEF_OP_VV (ne, 2048, int8_t, !=)
+DEF_OP_VV (ne, 4096, int8_t, !=)
+
+DEF_OP_VV (ne, 1, int16_t, !=)
+DEF_OP_VV (ne, 2, int16_t, !=)
+DEF_OP_VV (ne, 4, int16_t, !=)
+DEF_OP_VV (ne, 8, int16_t, !=)
+DEF_OP_VV (ne, 16, int16_t, !=)
+DEF_OP_VV (ne, 32, int16_t, !=)
+DEF_OP_VV (ne, 64, int16_t, !=)
+DEF_OP_VV (ne, 128, int16_t, !=)
+DEF_OP_VV (ne, 256, int16_t, !=)
+DEF_OP_VV (ne, 512, int16_t, !=)
+DEF_OP_VV (ne, 1024, int16_t, !=)
+DEF_OP_VV (ne, 2048, int16_t, !=)
+
+DEF_OP_VV (ne, 1, int32_t, !=)
+DEF_OP_VV (ne, 2, int32_t, !=)
+DEF_OP_VV (ne, 4, int32_t, !=)
+DEF_OP_VV (ne, 8, int32_t, !=)
+DEF_OP_VV (ne, 16, int32_t, !=)
+DEF_OP_VV (ne, 32, int32_t, !=)
+DEF_OP_VV (ne, 64, int32_t, !=)
+DEF_OP_VV (ne, 128, int32_t, !=)
+DEF_OP_VV (ne, 256, int32_t, !=)
+DEF_OP_VV (ne, 512, int32_t, !=)
+DEF_OP_VV (ne, 1024, int32_t, !=)
+
+DEF_OP_VV (ne, 1, int64_t, !=)
+DEF_OP_VV (ne, 2, int64_t, !=)
+DEF_OP_VV (ne, 4, int64_t, !=)
+DEF_OP_VV (ne, 8, int64_t, !=)
+DEF_OP_VV (ne, 16, int64_t, !=)
+DEF_OP_VV (ne, 32, int64_t, !=)
+DEF_OP_VV (ne, 64, int64_t, !=)
+DEF_OP_VV (ne, 128, int64_t, !=)
+DEF_OP_VV (ne, 256, int64_t, !=)
+DEF_OP_VV (ne, 512, int64_t, !=)
+
+DEF_OP_VV (ne, 1, _Float16, !=)
+DEF_OP_VV (ne, 2, _Float16, !=)
+DEF_OP_VV (ne, 4, _Float16, !=)
+DEF_OP_VV (ne, 8, _Float16, !=)
+DEF_OP_VV (ne, 16, _Float16, !=)
+DEF_OP_VV (ne, 32, _Float16, !=)
+DEF_OP_VV (ne, 64, _Float16, !=)
+DEF_OP_VV (ne, 128, _Float16, !=)
+DEF_OP_VV (ne, 256, _Float16, !=)
+DEF_OP_VV (ne, 512, _Float16, !=)
+DEF_OP_VV (ne, 1024, _Float16, !=)
+DEF_OP_VV (ne, 2048, _Float16, !=)
+
+DEF_OP_VV (ne, 1, float, !=)
+DEF_OP_VV (ne, 2, float, !=)
+DEF_OP_VV (ne, 4, float, !=)
+DEF_OP_VV (ne, 8, float, !=)
+DEF_OP_VV (ne, 16, float, !=)
+DEF_OP_VV (ne, 32, float, !=)
+DEF_OP_VV (ne, 64, float, !=)
+DEF_OP_VV (ne, 128, float, !=)
+DEF_OP_VV (ne, 256, float, !=)
+DEF_OP_VV (ne, 512, float, !=)
+DEF_OP_VV (ne, 1024, float, !=)
+
+DEF_OP_VV (ne, 1, double, !=)
+DEF_OP_VV (ne, 2, double, !=)
+DEF_OP_VV (ne, 4, double, !=)
+DEF_OP_VV (ne, 8, double, !=)
+DEF_OP_VV (ne, 16, double, !=)
+DEF_OP_VV (ne, 32, double, !=)
+DEF_OP_VV (ne, 64, double, !=)
+DEF_OP_VV (ne, 128, double, !=)
+DEF_OP_VV (ne, 256, double, !=)
+DEF_OP_VV (ne, 512, double, !=)
+
+/* { dg-final { scan-assembler-times {vmsne\.vv} 42 } } */
+/* { dg-final { scan-assembler-times {vmfne\.vv} 27 } } */
+/* { dg-final { scan-assembler-not {csrr} } } */
+/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
new file mode 100644
@@ -0,0 +1,106 @@
+/* { dg-do compile } */
+/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3 -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8 -fdump-tree-optimized" } */
+
+#include "def.h"
+
+DEF_OP_VV (gt, 1, int8_t, >)
+DEF_OP_VV (gt, 2, int8_t, >)
+DEF_OP_VV (gt, 4, int8_t, >)
+DEF_OP_VV (gt, 8, int8_t, >)
+DEF_OP_VV (gt, 16, int8_t, >)
+DEF_OP_VV (gt, 32, int8_t, >)
+DEF_OP_VV (gt, 64, int8_t, >)
+DEF_OP_VV (gt, 128, int8_t, >)
+DEF_OP_VV (gt, 256, int8_t, >)
+DEF_OP_VV (gt, 512, int8_t, >)
+DEF_OP_VV (gt, 1024, int8_t, >)
+DEF_OP_VV (gt, 2048, int8_t, >)
+DEF_OP_VV (gt, 4096, int8_t, >)
+
+DEF_OP_VV (gt, 1, int16_t, >)
+DEF_OP_VV (gt, 2, int16_t, >)
+DEF_OP_VV (gt, 4, int16_t, >)
+DEF_OP_VV (gt, 8, int16_t, >)
+DEF_OP_VV (gt, 16, int16_t, >)
+DEF_OP_VV (gt, 32, int16_t, >)
+DEF_OP_VV (gt, 64, int16_t, >)
+DEF_OP_VV (gt, 128, int16_t, >)
+DEF_OP_VV (gt, 256, int16_t, >)
+DEF_OP_VV (gt, 512, int16_t, >)
+DEF_OP_VV (gt, 1024, int16_t, >)
+DEF_OP_VV (gt, 2048, int16_t, >)
+
+DEF_OP_VV (gt, 1, int32_t, >)
+DEF_OP_VV (gt, 2, int32_t, >)
+DEF_OP_VV (gt, 4, int32_t, >)
+DEF_OP_VV (gt, 8, int32_t, >)
+DEF_OP_VV (gt, 16, int32_t, >)
+DEF_OP_VV (gt, 32, int32_t, >)
+DEF_OP_VV (gt, 64, int32_t, >)
+DEF_OP_VV (gt, 128, int32_t, >)
+DEF_OP_VV (gt, 256, int32_t, >)
+DEF_OP_VV (gt, 512, int32_t, >)
+DEF_OP_VV (gt, 1024, int32_t, >)
+
+DEF_OP_VV (gt, 1, int64_t, >)
+DEF_OP_VV (gt, 2, int64_t, >)
+DEF_OP_VV (gt, 4, int64_t, >)
+DEF_OP_VV (gt, 8, int64_t, >)
+DEF_OP_VV (gt, 16, int64_t, >)
+DEF_OP_VV (gt, 32, int64_t, >)
+DEF_OP_VV (gt, 64, int64_t, >)
+DEF_OP_VV (gt, 128, int64_t, >)
+DEF_OP_VV (gt, 256, int64_t, >)
+DEF_OP_VV (gt, 512, int64_t, >)
+
+DEF_OP_VV (gt, 1, _Float16, >)
+DEF_OP_VV (gt, 2, _Float16, >)
+DEF_OP_VV (gt, 4, _Float16, >)
+DEF_OP_VV (gt, 8, _Float16, >)
+DEF_OP_VV (gt, 16, _Float16, >)
+DEF_OP_VV (gt, 32, _Float16, >)
+DEF_OP_VV (gt, 64, _Float16, >)
+DEF_OP_VV (gt, 128, _Float16, >)
+DEF_OP_VV (gt, 256, _Float16, >)
+DEF_OP_VV (gt, 512, _Float16, >)
+DEF_OP_VV (gt, 1024, _Float16, >)
+DEF_OP_VV (gt, 2048, _Float16, >)
+
+DEF_OP_VV (gt, 1, float, >)
+DEF_OP_VV (gt, 2, float, >)
+DEF_OP_VV (gt, 4, float, >)
+DEF_OP_VV (gt, 8, float, >)
+DEF_OP_VV (gt, 16, float, >)
+DEF_OP_VV (gt, 32, float, >)
+DEF_OP_VV (gt, 64, float, >)
+DEF_OP_VV (gt, 128, float, >)
+DEF_OP_VV (gt, 256, float, >)
+DEF_OP_VV (gt, 512, float, >)
+DEF_OP_VV (gt, 1024, float, >)
+
+DEF_OP_VV (gt, 1, double, >)
+DEF_OP_VV (gt, 2, double, >)
+DEF_OP_VV (gt, 4, double, >)
+DEF_OP_VV (gt, 8, double, >)
+DEF_OP_VV (gt, 16, double, >)
+DEF_OP_VV (gt, 32, double, >)
+DEF_OP_VV (gt, 64, double, >)
+DEF_OP_VV (gt, 128, double, >)
+DEF_OP_VV (gt, 256, double, >)
+DEF_OP_VV (gt, 512, double, >)
+
+/* { dg-final { scan-assembler-times {vmsgt\.vv} 40 } } */
+/* { dg-final { scan-assembler-times {vmfgt\.vv} 27 } } */
+/* { dg-final { scan-assembler-not {csrr} } } */
+/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
new file mode 100644
@@ -0,0 +1,106 @@
+/* { dg-do compile } */
+/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3 -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8 -fdump-tree-optimized" } */
+
+#include "def.h"
+
+DEF_OP_VV (ge, 1, int8_t, >=)
+DEF_OP_VV (ge, 2, int8_t, >=)
+DEF_OP_VV (ge, 4, int8_t, >=)
+DEF_OP_VV (ge, 8, int8_t, >=)
+DEF_OP_VV (ge, 16, int8_t, >=)
+DEF_OP_VV (ge, 32, int8_t, >=)
+DEF_OP_VV (ge, 64, int8_t, >=)
+DEF_OP_VV (ge, 128, int8_t, >=)
+DEF_OP_VV (ge, 256, int8_t, >=)
+DEF_OP_VV (ge, 512, int8_t, >=)
+DEF_OP_VV (ge, 1024, int8_t, >=)
+DEF_OP_VV (ge, 2048, int8_t, >=)
+DEF_OP_VV (ge, 4096, int8_t, >=)
+
+DEF_OP_VV (ge, 1, int16_t, >=)
+DEF_OP_VV (ge, 2, int16_t, >=)
+DEF_OP_VV (ge, 4, int16_t, >=)
+DEF_OP_VV (ge, 8, int16_t, >=)
+DEF_OP_VV (ge, 16, int16_t, >=)
+DEF_OP_VV (ge, 32, int16_t, >=)
+DEF_OP_VV (ge, 64, int16_t, >=)
+DEF_OP_VV (ge, 128, int16_t, >=)
+DEF_OP_VV (ge, 256, int16_t, >=)
+DEF_OP_VV (ge, 512, int16_t, >=)
+DEF_OP_VV (ge, 1024, int16_t, >=)
+DEF_OP_VV (ge, 2048, int16_t, >=)
+
+DEF_OP_VV (ge, 1, int32_t, >=)
+DEF_OP_VV (ge, 2, int32_t, >=)
+DEF_OP_VV (ge, 4, int32_t, >=)
+DEF_OP_VV (ge, 8, int32_t, >=)
+DEF_OP_VV (ge, 16, int32_t, >=)
+DEF_OP_VV (ge, 32, int32_t, >=)
+DEF_OP_VV (ge, 64, int32_t, >=)
+DEF_OP_VV (ge, 128, int32_t, >=)
+DEF_OP_VV (ge, 256, int32_t, >=)
+DEF_OP_VV (ge, 512, int32_t, >=)
+DEF_OP_VV (ge, 1024, int32_t, >=)
+
+DEF_OP_VV (ge, 1, int64_t, >=)
+DEF_OP_VV (ge, 2, int64_t, >=)
+DEF_OP_VV (ge, 4, int64_t, >=)
+DEF_OP_VV (ge, 8, int64_t, >=)
+DEF_OP_VV (ge, 16, int64_t, >=)
+DEF_OP_VV (ge, 32, int64_t, >=)
+DEF_OP_VV (ge, 64, int64_t, >=)
+DEF_OP_VV (ge, 128, int64_t, >=)
+DEF_OP_VV (ge, 256, int64_t, >=)
+DEF_OP_VV (ge, 512, int64_t, >=)
+
+DEF_OP_VV (ge, 1, _Float16, >=)
+DEF_OP_VV (ge, 2, _Float16, >=)
+DEF_OP_VV (ge, 4, _Float16, >=)
+DEF_OP_VV (ge, 8, _Float16, >=)
+DEF_OP_VV (ge, 16, _Float16, >=)
+DEF_OP_VV (ge, 32, _Float16, >=)
+DEF_OP_VV (ge, 64, _Float16, >=)
+DEF_OP_VV (ge, 128, _Float16, >=)
+DEF_OP_VV (ge, 256, _Float16, >=)
+DEF_OP_VV (ge, 512, _Float16, >=)
+DEF_OP_VV (ge, 1024, _Float16, >=)
+DEF_OP_VV (ge, 2048, _Float16, >=)
+
+DEF_OP_VV (ge, 1, float, >=)
+DEF_OP_VV (ge, 2, float, >=)
+DEF_OP_VV (ge, 4, float, >=)
+DEF_OP_VV (ge, 8, float, >=)
+DEF_OP_VV (ge, 16, float, >=)
+DEF_OP_VV (ge, 32, float, >=)
+DEF_OP_VV (ge, 64, float, >=)
+DEF_OP_VV (ge, 128, float, >=)
+DEF_OP_VV (ge, 256, float, >=)
+DEF_OP_VV (ge, 512, float, >=)
+DEF_OP_VV (ge, 1024, float, >=)
+
+DEF_OP_VV (ge, 1, double, >=)
+DEF_OP_VV (ge, 2, double, >=)
+DEF_OP_VV (ge, 4, double, >=)
+DEF_OP_VV (ge, 8, double, >=)
+DEF_OP_VV (ge, 16, double, >=)
+DEF_OP_VV (ge, 32, double, >=)
+DEF_OP_VV (ge, 64, double, >=)
+DEF_OP_VV (ge, 128, double, >=)
+DEF_OP_VV (ge, 256, double, >=)
+DEF_OP_VV (ge, 512, double, >=)
+
+/* { dg-final { scan-assembler-times {vmsge\.vv} 40 } } */
+/* { dg-final { scan-assembler-times {vmfge\.vv} 27 } } */
+/* { dg-final { scan-assembler-not {csrr} } } */
+/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
new file mode 100644
@@ -0,0 +1,106 @@
+/* { dg-do compile } */
+/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3 -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8 -fdump-tree-optimized" } */
+
+#include "def.h"
+
+DEF_OP_VV (lt, 1, int8_t, <)
+DEF_OP_VV (lt, 2, int8_t, <)
+DEF_OP_VV (lt, 4, int8_t, <)
+DEF_OP_VV (lt, 8, int8_t, <)
+DEF_OP_VV (lt, 16, int8_t, <)
+DEF_OP_VV (lt, 32, int8_t, <)
+DEF_OP_VV (lt, 64, int8_t, <)
+DEF_OP_VV (lt, 128, int8_t, <)
+DEF_OP_VV (lt, 256, int8_t, <)
+DEF_OP_VV (lt, 512, int8_t, <)
+DEF_OP_VV (lt, 1024, int8_t, <)
+DEF_OP_VV (lt, 2048, int8_t, <)
+DEF_OP_VV (lt, 4096, int8_t, <)
+
+DEF_OP_VV (lt, 1, int16_t, <)
+DEF_OP_VV (lt, 2, int16_t, <)
+DEF_OP_VV (lt, 4, int16_t, <)
+DEF_OP_VV (lt, 8, int16_t, <)
+DEF_OP_VV (lt, 16, int16_t, <)
+DEF_OP_VV (lt, 32, int16_t, <)
+DEF_OP_VV (lt, 64, int16_t, <)
+DEF_OP_VV (lt, 128, int16_t, <)
+DEF_OP_VV (lt, 256, int16_t, <)
+DEF_OP_VV (lt, 512, int16_t, <)
+DEF_OP_VV (lt, 1024, int16_t, <)
+DEF_OP_VV (lt, 2048, int16_t, <)
+
+DEF_OP_VV (lt, 1, int32_t, <)
+DEF_OP_VV (lt, 2, int32_t, <)
+DEF_OP_VV (lt, 4, int32_t, <)
+DEF_OP_VV (lt, 8, int32_t, <)
+DEF_OP_VV (lt, 16, int32_t, <)
+DEF_OP_VV (lt, 32, int32_t, <)
+DEF_OP_VV (lt, 64, int32_t, <)
+DEF_OP_VV (lt, 128, int32_t, <)
+DEF_OP_VV (lt, 256, int32_t, <)
+DEF_OP_VV (lt, 512, int32_t, <)
+DEF_OP_VV (lt, 1024, int32_t, <)
+
+DEF_OP_VV (lt, 1, int64_t, <)
+DEF_OP_VV (lt, 2, int64_t, <)
+DEF_OP_VV (lt, 4, int64_t, <)
+DEF_OP_VV (lt, 8, int64_t, <)
+DEF_OP_VV (lt, 16, int64_t, <)
+DEF_OP_VV (lt, 32, int64_t, <)
+DEF_OP_VV (lt, 64, int64_t, <)
+DEF_OP_VV (lt, 128, int64_t, <)
+DEF_OP_VV (lt, 256, int64_t, <)
+DEF_OP_VV (lt, 512, int64_t, <)
+
+DEF_OP_VV (lt, 1, _Float16, <)
+DEF_OP_VV (lt, 2, _Float16, <)
+DEF_OP_VV (lt, 4, _Float16, <)
+DEF_OP_VV (lt, 8, _Float16, <)
+DEF_OP_VV (lt, 16, _Float16, <)
+DEF_OP_VV (lt, 32, _Float16, <)
+DEF_OP_VV (lt, 64, _Float16, <)
+DEF_OP_VV (lt, 128, _Float16, <)
+DEF_OP_VV (lt, 256, _Float16, <)
+DEF_OP_VV (lt, 512, _Float16, <)
+DEF_OP_VV (lt, 1024, _Float16, <)
+DEF_OP_VV (lt, 2048, _Float16, <)
+
+DEF_OP_VV (lt, 1, float, <)
+DEF_OP_VV (lt, 2, float, <)
+DEF_OP_VV (lt, 4, float, <)
+DEF_OP_VV (lt, 8, float, <)
+DEF_OP_VV (lt, 16, float, <)
+DEF_OP_VV (lt, 32, float, <)
+DEF_OP_VV (lt, 64, float, <)
+DEF_OP_VV (lt, 128, float, <)
+DEF_OP_VV (lt, 256, float, <)
+DEF_OP_VV (lt, 512, float, <)
+DEF_OP_VV (lt, 1024, float, <)
+
+DEF_OP_VV (lt, 1, double, <)
+DEF_OP_VV (lt, 2, double, <)
+DEF_OP_VV (lt, 4, double, <)
+DEF_OP_VV (lt, 8, double, <)
+DEF_OP_VV (lt, 16, double, <)
+DEF_OP_VV (lt, 32, double, <)
+DEF_OP_VV (lt, 64, double, <)
+DEF_OP_VV (lt, 128, double, <)
+DEF_OP_VV (lt, 256, double, <)
+DEF_OP_VV (lt, 512, double, <)
+
+/* { dg-final { scan-assembler-times {vmslt\.vv} 40 } } */
+/* { dg-final { scan-assembler-times {vmflt\.vv} 27 } } */
+/* { dg-final { scan-assembler-not {csrr} } } */
+/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
new file mode 100644
@@ -0,0 +1,106 @@
+/* { dg-do compile } */
+/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3 -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8 -fdump-tree-optimized" } */
+
+#include "def.h"
+
+DEF_OP_VV (le, 1, int8_t, <=)
+DEF_OP_VV (le, 2, int8_t, <=)
+DEF_OP_VV (le, 4, int8_t, <=)
+DEF_OP_VV (le, 8, int8_t, <=)
+DEF_OP_VV (le, 16, int8_t, <=)
+DEF_OP_VV (le, 32, int8_t, <=)
+DEF_OP_VV (le, 64, int8_t, <=)
+DEF_OP_VV (le, 128, int8_t, <=)
+DEF_OP_VV (le, 256, int8_t, <=)
+DEF_OP_VV (le, 512, int8_t, <=)
+DEF_OP_VV (le, 1024, int8_t, <=)
+DEF_OP_VV (le, 2048, int8_t, <=)
+DEF_OP_VV (le, 4096, int8_t, <=)
+
+DEF_OP_VV (le, 1, int16_t, <=)
+DEF_OP_VV (le, 2, int16_t, <=)
+DEF_OP_VV (le, 4, int16_t, <=)
+DEF_OP_VV (le, 8, int16_t, <=)
+DEF_OP_VV (le, 16, int16_t, <=)
+DEF_OP_VV (le, 32, int16_t, <=)
+DEF_OP_VV (le, 64, int16_t, <=)
+DEF_OP_VV (le, 128, int16_t, <=)
+DEF_OP_VV (le, 256, int16_t, <=)
+DEF_OP_VV (le, 512, int16_t, <=)
+DEF_OP_VV (le, 1024, int16_t, <=)
+DEF_OP_VV (le, 2048, int16_t, <=)
+
+DEF_OP_VV (le, 1, int32_t, <=)
+DEF_OP_VV (le, 2, int32_t, <=)
+DEF_OP_VV (le, 4, int32_t, <=)
+DEF_OP_VV (le, 8, int32_t, <=)
+DEF_OP_VV (le, 16, int32_t, <=)
+DEF_OP_VV (le, 32, int32_t, <=)
+DEF_OP_VV (le, 64, int32_t, <=)
+DEF_OP_VV (le, 128, int32_t, <=)
+DEF_OP_VV (le, 256, int32_t, <=)
+DEF_OP_VV (le, 512, int32_t, <=)
+DEF_OP_VV (le, 1024, int32_t, <=)
+
+DEF_OP_VV (le, 1, int64_t, <=)
+DEF_OP_VV (le, 2, int64_t, <=)
+DEF_OP_VV (le, 4, int64_t, <=)
+DEF_OP_VV (le, 8, int64_t, <=)
+DEF_OP_VV (le, 16, int64_t, <=)
+DEF_OP_VV (le, 32, int64_t, <=)
+DEF_OP_VV (le, 64, int64_t, <=)
+DEF_OP_VV (le, 128, int64_t, <=)
+DEF_OP_VV (le, 256, int64_t, <=)
+DEF_OP_VV (le, 512, int64_t, <=)
+
+DEF_OP_VV (le, 1, _Float16, <=)
+DEF_OP_VV (le, 2, _Float16, <=)
+DEF_OP_VV (le, 4, _Float16, <=)
+DEF_OP_VV (le, 8, _Float16, <=)
+DEF_OP_VV (le, 16, _Float16, <=)
+DEF_OP_VV (le, 32, _Float16, <=)
+DEF_OP_VV (le, 64, _Float16, <=)
+DEF_OP_VV (le, 128, _Float16, <=)
+DEF_OP_VV (le, 256, _Float16, <=)
+DEF_OP_VV (le, 512, _Float16, <=)
+DEF_OP_VV (le, 1024, _Float16, <=)
+DEF_OP_VV (le, 2048, _Float16, <=)
+
+DEF_OP_VV (le, 1, float, <=)
+DEF_OP_VV (le, 2, float, <=)
+DEF_OP_VV (le, 4, float, <=)
+DEF_OP_VV (le, 8, float, <=)
+DEF_OP_VV (le, 16, float, <=)
+DEF_OP_VV (le, 32, float, <=)
+DEF_OP_VV (le, 64, float, <=)
+DEF_OP_VV (le, 128, float, <=)
+DEF_OP_VV (le, 256, float, <=)
+DEF_OP_VV (le, 512, float, <=)
+DEF_OP_VV (le, 1024, float, <=)
+
+DEF_OP_VV (le, 1, double, <=)
+DEF_OP_VV (le, 2, double, <=)
+DEF_OP_VV (le, 4, double, <=)
+DEF_OP_VV (le, 8, double, <=)
+DEF_OP_VV (le, 16, double, <=)
+DEF_OP_VV (le, 32, double, <=)
+DEF_OP_VV (le, 64, double, <=)
+DEF_OP_VV (le, 128, double, <=)
+DEF_OP_VV (le, 256, double, <=)
+DEF_OP_VV (le, 512, double, <=)
+
+/* { dg-final { scan-assembler-times {vmsle\.vv} 40 } } */
+/* { dg-final { scan-assembler-times {vmfle\.vv} 27 } } */
+/* { dg-final { scan-assembler-not {csrr} } } */
+/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
@@ -249,3 +249,12 @@ typedef double v512df __attribute__ ((vector_size (4096)));
{ \
return v[INDEX]; \
}
+
+#define DEF_MASK_LOGIC(PREFIX, NUM, TYPE, OP) \
+ void __attribute__ ((noinline, noclone)) \
+ PREFIX##_##TYPE##NUM (TYPE *restrict a, TYPE *restrict b, TYPE *restrict c, \
+ TYPE *restrict d, TYPE *restrict e) \
+ { \
+ for (int i = 0; i < NUM; ++i) \
+ a[i] = (b[i] > c[i]) OP (d[i] < e[i]); \
+ }
new file mode 100644
@@ -0,0 +1,69 @@
+/* { dg-do compile } */
+/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3 -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8 -fdump-tree-optimized" } */
+
+#include "def.h"
+
+DEF_MASK_LOGIC (and, 1, int8_t, &)
+DEF_MASK_LOGIC (and, 2, int8_t, &)
+DEF_MASK_LOGIC (and, 4, int8_t, &)
+DEF_MASK_LOGIC (and, 8, int8_t, &)
+DEF_MASK_LOGIC (and, 16, int8_t, &)
+DEF_MASK_LOGIC (and, 32, int8_t, &)
+DEF_MASK_LOGIC (and, 64, int8_t, &)
+DEF_MASK_LOGIC (and, 128, int8_t, &)
+DEF_MASK_LOGIC (and, 256, int8_t, &)
+DEF_MASK_LOGIC (and, 512, int8_t, &)
+DEF_MASK_LOGIC (and, 1024, int8_t, &)
+DEF_MASK_LOGIC (and, 2048, int8_t, &)
+DEF_MASK_LOGIC (and, 4096, int8_t, &)
+
+DEF_MASK_LOGIC (and, 1, int16_t, &)
+DEF_MASK_LOGIC (and, 2, int16_t, &)
+DEF_MASK_LOGIC (and, 4, int16_t, &)
+DEF_MASK_LOGIC (and, 8, int16_t, &)
+DEF_MASK_LOGIC (and, 16, int16_t, &)
+DEF_MASK_LOGIC (and, 32, int16_t, &)
+DEF_MASK_LOGIC (and, 64, int16_t, &)
+DEF_MASK_LOGIC (and, 128, int16_t, &)
+DEF_MASK_LOGIC (and, 256, int16_t, &)
+DEF_MASK_LOGIC (and, 512, int16_t, &)
+DEF_MASK_LOGIC (and, 1024, int16_t, &)
+DEF_MASK_LOGIC (and, 2048, int16_t, &)
+
+DEF_MASK_LOGIC (and, 1, int32_t, &)
+DEF_MASK_LOGIC (and, 2, int32_t, &)
+DEF_MASK_LOGIC (and, 4, int32_t, &)
+DEF_MASK_LOGIC (and, 8, int32_t, &)
+DEF_MASK_LOGIC (and, 16, int32_t, &)
+DEF_MASK_LOGIC (and, 32, int32_t, &)
+DEF_MASK_LOGIC (and, 64, int32_t, &)
+DEF_MASK_LOGIC (and, 128, int32_t, &)
+DEF_MASK_LOGIC (and, 256, int32_t, &)
+DEF_MASK_LOGIC (and, 512, int32_t, &)
+DEF_MASK_LOGIC (and, 1024, int32_t, &)
+
+DEF_MASK_LOGIC (and, 1, int64_t, &)
+DEF_MASK_LOGIC (and, 2, int64_t, &)
+DEF_MASK_LOGIC (and, 4, int64_t, &)
+DEF_MASK_LOGIC (and, 8, int64_t, &)
+DEF_MASK_LOGIC (and, 16, int64_t, &)
+DEF_MASK_LOGIC (and, 32, int64_t, &)
+DEF_MASK_LOGIC (and, 64, int64_t, &)
+DEF_MASK_LOGIC (and, 128, int64_t, &)
+DEF_MASK_LOGIC (and, 256, int64_t, &)
+DEF_MASK_LOGIC (and, 512, int64_t, &)
+
+/* { dg-final { scan-assembler-times {vmand\.mm} 42 } } */
+/* { dg-final { scan-assembler-not {csrr} } } */
+/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
new file mode 100644
@@ -0,0 +1,69 @@
+/* { dg-do compile } */
+/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3 -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8 -fdump-tree-optimized" } */
+
+#include "def.h"
+
+DEF_MASK_LOGIC (and, 1, int8_t, |)
+DEF_MASK_LOGIC (and, 2, int8_t, |)
+DEF_MASK_LOGIC (and, 4, int8_t, |)
+DEF_MASK_LOGIC (and, 8, int8_t, |)
+DEF_MASK_LOGIC (and, 16, int8_t, |)
+DEF_MASK_LOGIC (and, 32, int8_t, |)
+DEF_MASK_LOGIC (and, 64, int8_t, |)
+DEF_MASK_LOGIC (and, 128, int8_t, |)
+DEF_MASK_LOGIC (and, 256, int8_t, |)
+DEF_MASK_LOGIC (and, 512, int8_t, |)
+DEF_MASK_LOGIC (and, 1024, int8_t, |)
+DEF_MASK_LOGIC (and, 2048, int8_t, |)
+DEF_MASK_LOGIC (and, 4096, int8_t, |)
+
+DEF_MASK_LOGIC (and, 1, int16_t, |)
+DEF_MASK_LOGIC (and, 2, int16_t, |)
+DEF_MASK_LOGIC (and, 4, int16_t, |)
+DEF_MASK_LOGIC (and, 8, int16_t, |)
+DEF_MASK_LOGIC (and, 16, int16_t, |)
+DEF_MASK_LOGIC (and, 32, int16_t, |)
+DEF_MASK_LOGIC (and, 64, int16_t, |)
+DEF_MASK_LOGIC (and, 128, int16_t, |)
+DEF_MASK_LOGIC (and, 256, int16_t, |)
+DEF_MASK_LOGIC (and, 512, int16_t, |)
+DEF_MASK_LOGIC (and, 1024, int16_t, |)
+DEF_MASK_LOGIC (and, 2048, int16_t, |)
+
+DEF_MASK_LOGIC (and, 1, int32_t, |)
+DEF_MASK_LOGIC (and, 2, int32_t, |)
+DEF_MASK_LOGIC (and, 4, int32_t, |)
+DEF_MASK_LOGIC (and, 8, int32_t, |)
+DEF_MASK_LOGIC (and, 16, int32_t, |)
+DEF_MASK_LOGIC (and, 32, int32_t, |)
+DEF_MASK_LOGIC (and, 64, int32_t, |)
+DEF_MASK_LOGIC (and, 128, int32_t, |)
+DEF_MASK_LOGIC (and, 256, int32_t, |)
+DEF_MASK_LOGIC (and, 512, int32_t, |)
+DEF_MASK_LOGIC (and, 1024, int32_t, |)
+
+DEF_MASK_LOGIC (and, 1, int64_t, |)
+DEF_MASK_LOGIC (and, 2, int64_t, |)
+DEF_MASK_LOGIC (and, 4, int64_t, |)
+DEF_MASK_LOGIC (and, 8, int64_t, |)
+DEF_MASK_LOGIC (and, 16, int64_t, |)
+DEF_MASK_LOGIC (and, 32, int64_t, |)
+DEF_MASK_LOGIC (and, 64, int64_t, |)
+DEF_MASK_LOGIC (and, 128, int64_t, |)
+DEF_MASK_LOGIC (and, 256, int64_t, |)
+DEF_MASK_LOGIC (and, 512, int64_t, |)
+
+/* { dg-final { scan-assembler-times {vmor\.mm} 42 } } */
+/* { dg-final { scan-assembler-not {csrr} } } */
+/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */
new file mode 100644
@@ -0,0 +1,69 @@
+/* { dg-do compile } */
+/* { dg-options "-march=rv64gcv_zvfh_zvl4096b -mabi=lp64d -O3 -fno-schedule-insns -fno-schedule-insns2 --param=riscv-autovec-lmul=m8 -fdump-tree-optimized" } */
+
+#include "def.h"
+
+DEF_MASK_LOGIC (and, 1, int8_t, ^)
+DEF_MASK_LOGIC (and, 2, int8_t, ^)
+DEF_MASK_LOGIC (and, 4, int8_t, ^)
+DEF_MASK_LOGIC (and, 8, int8_t, ^)
+DEF_MASK_LOGIC (and, 16, int8_t, ^)
+DEF_MASK_LOGIC (and, 32, int8_t, ^)
+DEF_MASK_LOGIC (and, 64, int8_t, ^)
+DEF_MASK_LOGIC (and, 128, int8_t, ^)
+DEF_MASK_LOGIC (and, 256, int8_t, ^)
+DEF_MASK_LOGIC (and, 512, int8_t, ^)
+DEF_MASK_LOGIC (and, 1024, int8_t, ^)
+DEF_MASK_LOGIC (and, 2048, int8_t, ^)
+DEF_MASK_LOGIC (and, 4096, int8_t, ^)
+
+DEF_MASK_LOGIC (and, 1, int16_t, ^)
+DEF_MASK_LOGIC (and, 2, int16_t, ^)
+DEF_MASK_LOGIC (and, 4, int16_t, ^)
+DEF_MASK_LOGIC (and, 8, int16_t, ^)
+DEF_MASK_LOGIC (and, 16, int16_t, ^)
+DEF_MASK_LOGIC (and, 32, int16_t, ^)
+DEF_MASK_LOGIC (and, 64, int16_t, ^)
+DEF_MASK_LOGIC (and, 128, int16_t, ^)
+DEF_MASK_LOGIC (and, 256, int16_t, ^)
+DEF_MASK_LOGIC (and, 512, int16_t, ^)
+DEF_MASK_LOGIC (and, 1024, int16_t, ^)
+DEF_MASK_LOGIC (and, 2048, int16_t, ^)
+
+DEF_MASK_LOGIC (and, 1, int32_t, ^)
+DEF_MASK_LOGIC (and, 2, int32_t, ^)
+DEF_MASK_LOGIC (and, 4, int32_t, ^)
+DEF_MASK_LOGIC (and, 8, int32_t, ^)
+DEF_MASK_LOGIC (and, 16, int32_t, ^)
+DEF_MASK_LOGIC (and, 32, int32_t, ^)
+DEF_MASK_LOGIC (and, 64, int32_t, ^)
+DEF_MASK_LOGIC (and, 128, int32_t, ^)
+DEF_MASK_LOGIC (and, 256, int32_t, ^)
+DEF_MASK_LOGIC (and, 512, int32_t, ^)
+DEF_MASK_LOGIC (and, 1024, int32_t, ^)
+
+DEF_MASK_LOGIC (and, 1, int64_t, ^)
+DEF_MASK_LOGIC (and, 2, int64_t, ^)
+DEF_MASK_LOGIC (and, 4, int64_t, ^)
+DEF_MASK_LOGIC (and, 8, int64_t, ^)
+DEF_MASK_LOGIC (and, 16, int64_t, ^)
+DEF_MASK_LOGIC (and, 32, int64_t, ^)
+DEF_MASK_LOGIC (and, 64, int64_t, ^)
+DEF_MASK_LOGIC (and, 128, int64_t, ^)
+DEF_MASK_LOGIC (and, 256, int64_t, ^)
+DEF_MASK_LOGIC (and, 512, int64_t, ^)
+
+/* { dg-final { scan-assembler-times {vmxor\.mm} 42 } } */
+/* { dg-final { scan-assembler-not {csrr} } } */
+/* { dg-final { scan-tree-dump-not "1,1" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2,2" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4,4" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "16,16" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "32,32" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "64,64" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "128,128" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "256,256" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "512,512" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "1024,1024" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "2048,2048" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "4096,4096" "optimized" } } */