llvm-for-llvmta/test/CodeGen/RISCV/rvv/vmadc.carry.in-rv32.ll

1137 lines
36 KiB
LLVM
Raw Permalink Normal View History

2022-04-25 10:02:23 +02:00
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=riscv32 -mattr=+experimental-v -verify-machineinstrs \
; RUN: --riscv-no-aliases < %s | FileCheck %s
declare <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i8.nxv1i8(
<vscale x 1 x i8>,
<vscale x 1 x i8>,
<vscale x 1 x i1>,
i32);
define <vscale x 1 x i1> @intrinsic_vmadc.carry.in_vvm_nxv1i1_nxv1i8_nxv1i8(<vscale x 1 x i8> %0, <vscale x 1 x i8> %1, <vscale x 1 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv1i1_nxv1i8_nxv1i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e8,mf8,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i8.nxv1i8(
<vscale x 1 x i8> %0,
<vscale x 1 x i8> %1,
<vscale x 1 x i1> %2,
i32 %3)
ret <vscale x 1 x i1> %a
}
declare <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i8.nxv2i8(
<vscale x 2 x i8>,
<vscale x 2 x i8>,
<vscale x 2 x i1>,
i32);
define <vscale x 2 x i1> @intrinsic_vmadc.carry.in_vvm_nxv2i1_nxv2i8_nxv2i8(<vscale x 2 x i8> %0, <vscale x 2 x i8> %1, <vscale x 2 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv2i1_nxv2i8_nxv2i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e8,mf4,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i8.nxv2i8(
<vscale x 2 x i8> %0,
<vscale x 2 x i8> %1,
<vscale x 2 x i1> %2,
i32 %3)
ret <vscale x 2 x i1> %a
}
declare <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i8.nxv4i8(
<vscale x 4 x i8>,
<vscale x 4 x i8>,
<vscale x 4 x i1>,
i32);
define <vscale x 4 x i1> @intrinsic_vmadc.carry.in_vvm_nxv4i1_nxv4i8_nxv4i8(<vscale x 4 x i8> %0, <vscale x 4 x i8> %1, <vscale x 4 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv4i1_nxv4i8_nxv4i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e8,mf2,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i8.nxv4i8(
<vscale x 4 x i8> %0,
<vscale x 4 x i8> %1,
<vscale x 4 x i1> %2,
i32 %3)
ret <vscale x 4 x i1> %a
}
declare <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i8.nxv8i8(
<vscale x 8 x i8>,
<vscale x 8 x i8>,
<vscale x 8 x i1>,
i32);
define <vscale x 8 x i1> @intrinsic_vmadc.carry.in_vvm_nxv8i1_nxv8i8_nxv8i8(<vscale x 8 x i8> %0, <vscale x 8 x i8> %1, <vscale x 8 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv8i1_nxv8i8_nxv8i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e8,m1,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i8.nxv8i8(
<vscale x 8 x i8> %0,
<vscale x 8 x i8> %1,
<vscale x 8 x i1> %2,
i32 %3)
ret <vscale x 8 x i1> %a
}
declare <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i8.nxv16i8(
<vscale x 16 x i8>,
<vscale x 16 x i8>,
<vscale x 16 x i1>,
i32);
define <vscale x 16 x i1> @intrinsic_vmadc.carry.in_vvm_nxv16i1_nxv16i8_nxv16i8(<vscale x 16 x i8> %0, <vscale x 16 x i8> %1, <vscale x 16 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv16i1_nxv16i8_nxv16i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e8,m2,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v10, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i8.nxv16i8(
<vscale x 16 x i8> %0,
<vscale x 16 x i8> %1,
<vscale x 16 x i1> %2,
i32 %3)
ret <vscale x 16 x i1> %a
}
declare <vscale x 32 x i1> @llvm.riscv.vmadc.carry.in.nxv32i8.nxv32i8(
<vscale x 32 x i8>,
<vscale x 32 x i8>,
<vscale x 32 x i1>,
i32);
define <vscale x 32 x i1> @intrinsic_vmadc.carry.in_vvm_nxv32i1_nxv32i8_nxv32i8(<vscale x 32 x i8> %0, <vscale x 32 x i8> %1, <vscale x 32 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv32i1_nxv32i8_nxv32i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e8,m4,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v12, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 32 x i1> @llvm.riscv.vmadc.carry.in.nxv32i8.nxv32i8(
<vscale x 32 x i8> %0,
<vscale x 32 x i8> %1,
<vscale x 32 x i1> %2,
i32 %3)
ret <vscale x 32 x i1> %a
}
declare <vscale x 64 x i1> @llvm.riscv.vmadc.carry.in.nxv64i8.nxv64i8(
<vscale x 64 x i8>,
<vscale x 64 x i8>,
<vscale x 64 x i1>,
i32);
define <vscale x 64 x i1> @intrinsic_vmadc.carry.in_vvm_nxv64i1_nxv64i8_nxv64i8(<vscale x 64 x i8> %0, <vscale x 64 x i8> %1, <vscale x 64 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv64i1_nxv64i8_nxv64i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e8,m8,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v16, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 64 x i1> @llvm.riscv.vmadc.carry.in.nxv64i8.nxv64i8(
<vscale x 64 x i8> %0,
<vscale x 64 x i8> %1,
<vscale x 64 x i1> %2,
i32 %3)
ret <vscale x 64 x i1> %a
}
declare <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i16.nxv1i16(
<vscale x 1 x i16>,
<vscale x 1 x i16>,
<vscale x 1 x i1>,
i32);
define <vscale x 1 x i1> @intrinsic_vmadc.carry.in_vvm_nxv1i1_nxv1i16_nxv1i16(<vscale x 1 x i16> %0, <vscale x 1 x i16> %1, <vscale x 1 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv1i1_nxv1i16_nxv1i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e16,mf4,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i16.nxv1i16(
<vscale x 1 x i16> %0,
<vscale x 1 x i16> %1,
<vscale x 1 x i1> %2,
i32 %3)
ret <vscale x 1 x i1> %a
}
declare <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i16.nxv2i16(
<vscale x 2 x i16>,
<vscale x 2 x i16>,
<vscale x 2 x i1>,
i32);
define <vscale x 2 x i1> @intrinsic_vmadc.carry.in_vvm_nxv2i1_nxv2i16_nxv2i16(<vscale x 2 x i16> %0, <vscale x 2 x i16> %1, <vscale x 2 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv2i1_nxv2i16_nxv2i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e16,mf2,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i16.nxv2i16(
<vscale x 2 x i16> %0,
<vscale x 2 x i16> %1,
<vscale x 2 x i1> %2,
i32 %3)
ret <vscale x 2 x i1> %a
}
declare <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i16.nxv4i16(
<vscale x 4 x i16>,
<vscale x 4 x i16>,
<vscale x 4 x i1>,
i32);
define <vscale x 4 x i1> @intrinsic_vmadc.carry.in_vvm_nxv4i1_nxv4i16_nxv4i16(<vscale x 4 x i16> %0, <vscale x 4 x i16> %1, <vscale x 4 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv4i1_nxv4i16_nxv4i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e16,m1,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i16.nxv4i16(
<vscale x 4 x i16> %0,
<vscale x 4 x i16> %1,
<vscale x 4 x i1> %2,
i32 %3)
ret <vscale x 4 x i1> %a
}
declare <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i16.nxv8i16(
<vscale x 8 x i16>,
<vscale x 8 x i16>,
<vscale x 8 x i1>,
i32);
define <vscale x 8 x i1> @intrinsic_vmadc.carry.in_vvm_nxv8i1_nxv8i16_nxv8i16(<vscale x 8 x i16> %0, <vscale x 8 x i16> %1, <vscale x 8 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv8i1_nxv8i16_nxv8i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e16,m2,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v10, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i16.nxv8i16(
<vscale x 8 x i16> %0,
<vscale x 8 x i16> %1,
<vscale x 8 x i1> %2,
i32 %3)
ret <vscale x 8 x i1> %a
}
declare <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i16.nxv16i16(
<vscale x 16 x i16>,
<vscale x 16 x i16>,
<vscale x 16 x i1>,
i32);
define <vscale x 16 x i1> @intrinsic_vmadc.carry.in_vvm_nxv16i1_nxv16i16_nxv16i16(<vscale x 16 x i16> %0, <vscale x 16 x i16> %1, <vscale x 16 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv16i1_nxv16i16_nxv16i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e16,m4,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v12, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i16.nxv16i16(
<vscale x 16 x i16> %0,
<vscale x 16 x i16> %1,
<vscale x 16 x i1> %2,
i32 %3)
ret <vscale x 16 x i1> %a
}
declare <vscale x 32 x i1> @llvm.riscv.vmadc.carry.in.nxv32i16.nxv32i16(
<vscale x 32 x i16>,
<vscale x 32 x i16>,
<vscale x 32 x i1>,
i32);
define <vscale x 32 x i1> @intrinsic_vmadc.carry.in_vvm_nxv32i1_nxv32i16_nxv32i16(<vscale x 32 x i16> %0, <vscale x 32 x i16> %1, <vscale x 32 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv32i1_nxv32i16_nxv32i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e16,m8,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v16, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 32 x i1> @llvm.riscv.vmadc.carry.in.nxv32i16.nxv32i16(
<vscale x 32 x i16> %0,
<vscale x 32 x i16> %1,
<vscale x 32 x i1> %2,
i32 %3)
ret <vscale x 32 x i1> %a
}
declare <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i32.nxv1i32(
<vscale x 1 x i32>,
<vscale x 1 x i32>,
<vscale x 1 x i1>,
i32);
define <vscale x 1 x i1> @intrinsic_vmadc.carry.in_vvm_nxv1i1_nxv1i32_nxv1i32(<vscale x 1 x i32> %0, <vscale x 1 x i32> %1, <vscale x 1 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv1i1_nxv1i32_nxv1i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e32,mf2,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i32.nxv1i32(
<vscale x 1 x i32> %0,
<vscale x 1 x i32> %1,
<vscale x 1 x i1> %2,
i32 %3)
ret <vscale x 1 x i1> %a
}
declare <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i32.nxv2i32(
<vscale x 2 x i32>,
<vscale x 2 x i32>,
<vscale x 2 x i1>,
i32);
define <vscale x 2 x i1> @intrinsic_vmadc.carry.in_vvm_nxv2i1_nxv2i32_nxv2i32(<vscale x 2 x i32> %0, <vscale x 2 x i32> %1, <vscale x 2 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv2i1_nxv2i32_nxv2i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e32,m1,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i32.nxv2i32(
<vscale x 2 x i32> %0,
<vscale x 2 x i32> %1,
<vscale x 2 x i1> %2,
i32 %3)
ret <vscale x 2 x i1> %a
}
declare <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i32.nxv4i32(
<vscale x 4 x i32>,
<vscale x 4 x i32>,
<vscale x 4 x i1>,
i32);
define <vscale x 4 x i1> @intrinsic_vmadc.carry.in_vvm_nxv4i1_nxv4i32_nxv4i32(<vscale x 4 x i32> %0, <vscale x 4 x i32> %1, <vscale x 4 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv4i1_nxv4i32_nxv4i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e32,m2,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v10, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i32.nxv4i32(
<vscale x 4 x i32> %0,
<vscale x 4 x i32> %1,
<vscale x 4 x i1> %2,
i32 %3)
ret <vscale x 4 x i1> %a
}
declare <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i32.nxv8i32(
<vscale x 8 x i32>,
<vscale x 8 x i32>,
<vscale x 8 x i1>,
i32);
define <vscale x 8 x i1> @intrinsic_vmadc.carry.in_vvm_nxv8i1_nxv8i32_nxv8i32(<vscale x 8 x i32> %0, <vscale x 8 x i32> %1, <vscale x 8 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv8i1_nxv8i32_nxv8i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e32,m4,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v12, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i32.nxv8i32(
<vscale x 8 x i32> %0,
<vscale x 8 x i32> %1,
<vscale x 8 x i1> %2,
i32 %3)
ret <vscale x 8 x i1> %a
}
declare <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i32.nxv16i32(
<vscale x 16 x i32>,
<vscale x 16 x i32>,
<vscale x 16 x i1>,
i32);
define <vscale x 16 x i1> @intrinsic_vmadc.carry.in_vvm_nxv16i1_nxv16i32_nxv16i32(<vscale x 16 x i32> %0, <vscale x 16 x i32> %1, <vscale x 16 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vvm_nxv16i1_nxv16i32_nxv16i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e32,m8,ta,mu
; CHECK-NEXT: vmadc.vvm v25, v8, v16, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i32.nxv16i32(
<vscale x 16 x i32> %0,
<vscale x 16 x i32> %1,
<vscale x 16 x i1> %2,
i32 %3)
ret <vscale x 16 x i1> %a
}
declare <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i8.i8(
<vscale x 1 x i8>,
i8,
<vscale x 1 x i1>,
i32);
define <vscale x 1 x i1> @intrinsic_vmadc.carry.in_vxm_nxv1i1_nxv1i8_i8(<vscale x 1 x i8> %0, i8 %1, <vscale x 1 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv1i1_nxv1i8_i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e8,mf8,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i8.i8(
<vscale x 1 x i8> %0,
i8 %1,
<vscale x 1 x i1> %2,
i32 %3)
ret <vscale x 1 x i1> %a
}
declare <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i8.i8(
<vscale x 2 x i8>,
i8,
<vscale x 2 x i1>,
i32);
define <vscale x 2 x i1> @intrinsic_vmadc.carry.in_vxm_nxv2i1_nxv2i8_i8(<vscale x 2 x i8> %0, i8 %1, <vscale x 2 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv2i1_nxv2i8_i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e8,mf4,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i8.i8(
<vscale x 2 x i8> %0,
i8 %1,
<vscale x 2 x i1> %2,
i32 %3)
ret <vscale x 2 x i1> %a
}
declare <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i8.i8(
<vscale x 4 x i8>,
i8,
<vscale x 4 x i1>,
i32);
define <vscale x 4 x i1> @intrinsic_vmadc.carry.in_vxm_nxv4i1_nxv4i8_i8(<vscale x 4 x i8> %0, i8 %1, <vscale x 4 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv4i1_nxv4i8_i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e8,mf2,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i8.i8(
<vscale x 4 x i8> %0,
i8 %1,
<vscale x 4 x i1> %2,
i32 %3)
ret <vscale x 4 x i1> %a
}
declare <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i8.i8(
<vscale x 8 x i8>,
i8,
<vscale x 8 x i1>,
i32);
define <vscale x 8 x i1> @intrinsic_vmadc.carry.in_vxm_nxv8i1_nxv8i8_i8(<vscale x 8 x i8> %0, i8 %1, <vscale x 8 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv8i1_nxv8i8_i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e8,m1,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i8.i8(
<vscale x 8 x i8> %0,
i8 %1,
<vscale x 8 x i1> %2,
i32 %3)
ret <vscale x 8 x i1> %a
}
declare <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i8.i8(
<vscale x 16 x i8>,
i8,
<vscale x 16 x i1>,
i32);
define <vscale x 16 x i1> @intrinsic_vmadc.carry.in_vxm_nxv16i1_nxv16i8_i8(<vscale x 16 x i8> %0, i8 %1, <vscale x 16 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv16i1_nxv16i8_i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e8,m2,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i8.i8(
<vscale x 16 x i8> %0,
i8 %1,
<vscale x 16 x i1> %2,
i32 %3)
ret <vscale x 16 x i1> %a
}
declare <vscale x 32 x i1> @llvm.riscv.vmadc.carry.in.nxv32i8.i8(
<vscale x 32 x i8>,
i8,
<vscale x 32 x i1>,
i32);
define <vscale x 32 x i1> @intrinsic_vmadc.carry.in_vxm_nxv32i1_nxv32i8_i8(<vscale x 32 x i8> %0, i8 %1, <vscale x 32 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv32i1_nxv32i8_i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e8,m4,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 32 x i1> @llvm.riscv.vmadc.carry.in.nxv32i8.i8(
<vscale x 32 x i8> %0,
i8 %1,
<vscale x 32 x i1> %2,
i32 %3)
ret <vscale x 32 x i1> %a
}
declare <vscale x 64 x i1> @llvm.riscv.vmadc.carry.in.nxv64i8.i8(
<vscale x 64 x i8>,
i8,
<vscale x 64 x i1>,
i32);
define <vscale x 64 x i1> @intrinsic_vmadc.carry.in_vxm_nxv64i1_nxv64i8_i8(<vscale x 64 x i8> %0, i8 %1, <vscale x 64 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv64i1_nxv64i8_i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e8,m8,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 64 x i1> @llvm.riscv.vmadc.carry.in.nxv64i8.i8(
<vscale x 64 x i8> %0,
i8 %1,
<vscale x 64 x i1> %2,
i32 %3)
ret <vscale x 64 x i1> %a
}
declare <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i16.i16(
<vscale x 1 x i16>,
i16,
<vscale x 1 x i1>,
i32);
define <vscale x 1 x i1> @intrinsic_vmadc.carry.in_vxm_nxv1i1_nxv1i16_i16(<vscale x 1 x i16> %0, i16 %1, <vscale x 1 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv1i1_nxv1i16_i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e16,mf4,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i16.i16(
<vscale x 1 x i16> %0,
i16 %1,
<vscale x 1 x i1> %2,
i32 %3)
ret <vscale x 1 x i1> %a
}
declare <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i16.i16(
<vscale x 2 x i16>,
i16,
<vscale x 2 x i1>,
i32);
define <vscale x 2 x i1> @intrinsic_vmadc.carry.in_vxm_nxv2i1_nxv2i16_i16(<vscale x 2 x i16> %0, i16 %1, <vscale x 2 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv2i1_nxv2i16_i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e16,mf2,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i16.i16(
<vscale x 2 x i16> %0,
i16 %1,
<vscale x 2 x i1> %2,
i32 %3)
ret <vscale x 2 x i1> %a
}
declare <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i16.i16(
<vscale x 4 x i16>,
i16,
<vscale x 4 x i1>,
i32);
define <vscale x 4 x i1> @intrinsic_vmadc.carry.in_vxm_nxv4i1_nxv4i16_i16(<vscale x 4 x i16> %0, i16 %1, <vscale x 4 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv4i1_nxv4i16_i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e16,m1,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i16.i16(
<vscale x 4 x i16> %0,
i16 %1,
<vscale x 4 x i1> %2,
i32 %3)
ret <vscale x 4 x i1> %a
}
declare <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i16.i16(
<vscale x 8 x i16>,
i16,
<vscale x 8 x i1>,
i32);
define <vscale x 8 x i1> @intrinsic_vmadc.carry.in_vxm_nxv8i1_nxv8i16_i16(<vscale x 8 x i16> %0, i16 %1, <vscale x 8 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv8i1_nxv8i16_i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e16,m2,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i16.i16(
<vscale x 8 x i16> %0,
i16 %1,
<vscale x 8 x i1> %2,
i32 %3)
ret <vscale x 8 x i1> %a
}
declare <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i16.i16(
<vscale x 16 x i16>,
i16,
<vscale x 16 x i1>,
i32);
define <vscale x 16 x i1> @intrinsic_vmadc.carry.in_vxm_nxv16i1_nxv16i16_i16(<vscale x 16 x i16> %0, i16 %1, <vscale x 16 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv16i1_nxv16i16_i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e16,m4,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i16.i16(
<vscale x 16 x i16> %0,
i16 %1,
<vscale x 16 x i1> %2,
i32 %3)
ret <vscale x 16 x i1> %a
}
declare <vscale x 32 x i1> @llvm.riscv.vmadc.carry.in.nxv32i16.i16(
<vscale x 32 x i16>,
i16,
<vscale x 32 x i1>,
i32);
define <vscale x 32 x i1> @intrinsic_vmadc.carry.in_vxm_nxv32i1_nxv32i16_i16(<vscale x 32 x i16> %0, i16 %1, <vscale x 32 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv32i1_nxv32i16_i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e16,m8,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 32 x i1> @llvm.riscv.vmadc.carry.in.nxv32i16.i16(
<vscale x 32 x i16> %0,
i16 %1,
<vscale x 32 x i1> %2,
i32 %3)
ret <vscale x 32 x i1> %a
}
declare <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i32.i32(
<vscale x 1 x i32>,
i32,
<vscale x 1 x i1>,
i32);
define <vscale x 1 x i1> @intrinsic_vmadc.carry.in_vxm_nxv1i1_nxv1i32_i32(<vscale x 1 x i32> %0, i32 %1, <vscale x 1 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv1i1_nxv1i32_i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e32,mf2,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i32.i32(
<vscale x 1 x i32> %0,
i32 %1,
<vscale x 1 x i1> %2,
i32 %3)
ret <vscale x 1 x i1> %a
}
declare <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i32.i32(
<vscale x 2 x i32>,
i32,
<vscale x 2 x i1>,
i32);
define <vscale x 2 x i1> @intrinsic_vmadc.carry.in_vxm_nxv2i1_nxv2i32_i32(<vscale x 2 x i32> %0, i32 %1, <vscale x 2 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv2i1_nxv2i32_i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e32,m1,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i32.i32(
<vscale x 2 x i32> %0,
i32 %1,
<vscale x 2 x i1> %2,
i32 %3)
ret <vscale x 2 x i1> %a
}
declare <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i32.i32(
<vscale x 4 x i32>,
i32,
<vscale x 4 x i1>,
i32);
define <vscale x 4 x i1> @intrinsic_vmadc.carry.in_vxm_nxv4i1_nxv4i32_i32(<vscale x 4 x i32> %0, i32 %1, <vscale x 4 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv4i1_nxv4i32_i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e32,m2,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i32.i32(
<vscale x 4 x i32> %0,
i32 %1,
<vscale x 4 x i1> %2,
i32 %3)
ret <vscale x 4 x i1> %a
}
declare <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i32.i32(
<vscale x 8 x i32>,
i32,
<vscale x 8 x i1>,
i32);
define <vscale x 8 x i1> @intrinsic_vmadc.carry.in_vxm_nxv8i1_nxv8i32_i32(<vscale x 8 x i32> %0, i32 %1, <vscale x 8 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv8i1_nxv8i32_i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e32,m4,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i32.i32(
<vscale x 8 x i32> %0,
i32 %1,
<vscale x 8 x i1> %2,
i32 %3)
ret <vscale x 8 x i1> %a
}
declare <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i32.i32(
<vscale x 16 x i32>,
i32,
<vscale x 16 x i1>,
i32);
define <vscale x 16 x i1> @intrinsic_vmadc.carry.in_vxm_nxv16i1_nxv16i32_i32(<vscale x 16 x i32> %0, i32 %1, <vscale x 16 x i1> %2, i32 %3) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vxm_nxv16i1_nxv16i32_i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a1, a1, e32,m8,ta,mu
; CHECK-NEXT: vmadc.vxm v25, v8, a0, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i32.i32(
<vscale x 16 x i32> %0,
i32 %1,
<vscale x 16 x i1> %2,
i32 %3)
ret <vscale x 16 x i1> %a
}
define <vscale x 1 x i1> @intrinsic_vmadc.carry.in_vim_nxv1i1_nxv1i8_i8(<vscale x 1 x i8> %0, <vscale x 1 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv1i1_nxv1i8_i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e8,mf8,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i8.i8(
<vscale x 1 x i8> %0,
i8 9,
<vscale x 1 x i1> %1,
i32 %2)
ret <vscale x 1 x i1> %a
}
define <vscale x 2 x i1> @intrinsic_vmadc.carry.in_vim_nxv2i1_nxv2i8_i8(<vscale x 2 x i8> %0, <vscale x 2 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv2i1_nxv2i8_i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e8,mf4,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i8.i8(
<vscale x 2 x i8> %0,
i8 9,
<vscale x 2 x i1> %1,
i32 %2)
ret <vscale x 2 x i1> %a
}
define <vscale x 4 x i1> @intrinsic_vmadc.carry.in_vim_nxv4i1_nxv4i8_i8(<vscale x 4 x i8> %0, <vscale x 4 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv4i1_nxv4i8_i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e8,mf2,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i8.i8(
<vscale x 4 x i8> %0,
i8 9,
<vscale x 4 x i1> %1,
i32 %2)
ret <vscale x 4 x i1> %a
}
define <vscale x 8 x i1> @intrinsic_vmadc.carry.in_vim_nxv8i1_nxv8i8_i8(<vscale x 8 x i8> %0, <vscale x 8 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv8i1_nxv8i8_i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e8,m1,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i8.i8(
<vscale x 8 x i8> %0,
i8 9,
<vscale x 8 x i1> %1,
i32 %2)
ret <vscale x 8 x i1> %a
}
define <vscale x 16 x i1> @intrinsic_vmadc.carry.in_vim_nxv16i1_nxv16i8_i8(<vscale x 16 x i8> %0, <vscale x 16 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv16i1_nxv16i8_i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e8,m2,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i8.i8(
<vscale x 16 x i8> %0,
i8 9,
<vscale x 16 x i1> %1,
i32 %2)
ret <vscale x 16 x i1> %a
}
define <vscale x 32 x i1> @intrinsic_vmadc.carry.in_vim_nxv32i1_nxv32i8_i8(<vscale x 32 x i8> %0, <vscale x 32 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv32i1_nxv32i8_i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e8,m4,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 32 x i1> @llvm.riscv.vmadc.carry.in.nxv32i8.i8(
<vscale x 32 x i8> %0,
i8 9,
<vscale x 32 x i1> %1,
i32 %2)
ret <vscale x 32 x i1> %a
}
define <vscale x 64 x i1> @intrinsic_vmadc.carry.in_vim_nxv64i1_nxv64i8_i8(<vscale x 64 x i8> %0, <vscale x 64 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv64i1_nxv64i8_i8:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e8,m8,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 64 x i1> @llvm.riscv.vmadc.carry.in.nxv64i8.i8(
<vscale x 64 x i8> %0,
i8 9,
<vscale x 64 x i1> %1,
i32 %2)
ret <vscale x 64 x i1> %a
}
define <vscale x 1 x i1> @intrinsic_vmadc.carry.in_vim_nxv1i1_nxv1i16_i16(<vscale x 1 x i16> %0, <vscale x 1 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv1i1_nxv1i16_i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e16,mf4,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i16.i16(
<vscale x 1 x i16> %0,
i16 9,
<vscale x 1 x i1> %1,
i32 %2)
ret <vscale x 1 x i1> %a
}
define <vscale x 2 x i1> @intrinsic_vmadc.carry.in_vim_nxv2i1_nxv2i16_i16(<vscale x 2 x i16> %0, <vscale x 2 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv2i1_nxv2i16_i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e16,mf2,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i16.i16(
<vscale x 2 x i16> %0,
i16 9,
<vscale x 2 x i1> %1,
i32 %2)
ret <vscale x 2 x i1> %a
}
define <vscale x 4 x i1> @intrinsic_vmadc.carry.in_vim_nxv4i1_nxv4i16_i16(<vscale x 4 x i16> %0, <vscale x 4 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv4i1_nxv4i16_i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e16,m1,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i16.i16(
<vscale x 4 x i16> %0,
i16 9,
<vscale x 4 x i1> %1,
i32 %2)
ret <vscale x 4 x i1> %a
}
define <vscale x 8 x i1> @intrinsic_vmadc.carry.in_vim_nxv8i1_nxv8i16_i16(<vscale x 8 x i16> %0, <vscale x 8 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv8i1_nxv8i16_i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e16,m2,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i16.i16(
<vscale x 8 x i16> %0,
i16 9,
<vscale x 8 x i1> %1,
i32 %2)
ret <vscale x 8 x i1> %a
}
define <vscale x 16 x i1> @intrinsic_vmadc.carry.in_vim_nxv16i1_nxv16i16_i16(<vscale x 16 x i16> %0, <vscale x 16 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv16i1_nxv16i16_i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e16,m4,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i16.i16(
<vscale x 16 x i16> %0,
i16 9,
<vscale x 16 x i1> %1,
i32 %2)
ret <vscale x 16 x i1> %a
}
define <vscale x 32 x i1> @intrinsic_vmadc.carry.in_vim_nxv32i1_nxv32i16_i16(<vscale x 32 x i16> %0, <vscale x 32 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv32i1_nxv32i16_i16:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e16,m8,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 32 x i1> @llvm.riscv.vmadc.carry.in.nxv32i16.i16(
<vscale x 32 x i16> %0,
i16 9,
<vscale x 32 x i1> %1,
i32 %2)
ret <vscale x 32 x i1> %a
}
define <vscale x 1 x i1> @intrinsic_vmadc.carry.in_vim_nxv1i1_nxv1i32_i32(<vscale x 1 x i32> %0, <vscale x 1 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv1i1_nxv1i32_i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e32,mf2,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 1 x i1> @llvm.riscv.vmadc.carry.in.nxv1i32.i32(
<vscale x 1 x i32> %0,
i32 9,
<vscale x 1 x i1> %1,
i32 %2)
ret <vscale x 1 x i1> %a
}
define <vscale x 2 x i1> @intrinsic_vmadc.carry.in_vim_nxv2i1_nxv2i32_i32(<vscale x 2 x i32> %0, <vscale x 2 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv2i1_nxv2i32_i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e32,m1,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 2 x i1> @llvm.riscv.vmadc.carry.in.nxv2i32.i32(
<vscale x 2 x i32> %0,
i32 9,
<vscale x 2 x i1> %1,
i32 %2)
ret <vscale x 2 x i1> %a
}
define <vscale x 4 x i1> @intrinsic_vmadc.carry.in_vim_nxv4i1_nxv4i32_i32(<vscale x 4 x i32> %0, <vscale x 4 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv4i1_nxv4i32_i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e32,m2,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 4 x i1> @llvm.riscv.vmadc.carry.in.nxv4i32.i32(
<vscale x 4 x i32> %0,
i32 9,
<vscale x 4 x i1> %1,
i32 %2)
ret <vscale x 4 x i1> %a
}
define <vscale x 8 x i1> @intrinsic_vmadc.carry.in_vim_nxv8i1_nxv8i32_i32(<vscale x 8 x i32> %0, <vscale x 8 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv8i1_nxv8i32_i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e32,m4,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 8 x i1> @llvm.riscv.vmadc.carry.in.nxv8i32.i32(
<vscale x 8 x i32> %0,
i32 9,
<vscale x 8 x i1> %1,
i32 %2)
ret <vscale x 8 x i1> %a
}
define <vscale x 16 x i1> @intrinsic_vmadc.carry.in_vim_nxv16i1_nxv16i32_i32(<vscale x 16 x i32> %0, <vscale x 16 x i1> %1, i32 %2) nounwind {
; CHECK-LABEL: intrinsic_vmadc.carry.in_vim_nxv16i1_nxv16i32_i32:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: vsetvli a0, a0, e32,m8,ta,mu
; CHECK-NEXT: vmadc.vim v25, v8, 9, v0
; CHECK-NEXT: vmv1r.v v0, v25
; CHECK-NEXT: jalr zero, 0(ra)
entry:
%a = call <vscale x 16 x i1> @llvm.riscv.vmadc.carry.in.nxv16i32.i32(
<vscale x 16 x i32> %0,
i32 9,
<vscale x 16 x i1> %1,
i32 %2)
ret <vscale x 16 x i1> %a
}