diff options
| author | Bob Wilson <bob.wilson@apple.com> | 2010-12-15 22:14:12 +0000 |
|---|---|---|
| committer | Bob Wilson <bob.wilson@apple.com> | 2010-12-15 22:14:12 +0000 |
| commit | 0406356cd4cb7b689e2472faa8dfb7d721f9d274 (patch) | |
| tree | 0c1e2f26438acad7f4053004cd38fe2d9282c52d /include/llvm/IntrinsicsARM.td | |
| parent | 8d1b7e57e56015576fd489a57d74c53b98c5a56f (diff) | |
| download | llvm-0406356cd4cb7b689e2472faa8dfb7d721f9d274.tar.gz llvm-0406356cd4cb7b689e2472faa8dfb7d721f9d274.tar.bz2 llvm-0406356cd4cb7b689e2472faa8dfb7d721f9d274.tar.xz | |
Add Neon VCVT instructions for f32 <-> f16 conversions.
Clang is now providing intrinsics for these and so we need to support them
in the backend. Radar 8068427.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@121902 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'include/llvm/IntrinsicsARM.td')
| -rw-r--r-- | include/llvm/IntrinsicsARM.td | 6 |
1 files changed, 6 insertions, 0 deletions
diff --git a/include/llvm/IntrinsicsARM.td b/include/llvm/IntrinsicsARM.td index 6c047718e6..546538a57a 100644 --- a/include/llvm/IntrinsicsARM.td +++ b/include/llvm/IntrinsicsARM.td @@ -286,6 +286,12 @@ def int_arm_neon_vcvtfp2fxu : Neon_CvtFPToFx_Intrinsic; def int_arm_neon_vcvtfxs2fp : Neon_CvtFxToFP_Intrinsic; def int_arm_neon_vcvtfxu2fp : Neon_CvtFxToFP_Intrinsic; +// Vector Conversions Between Half-Precision and Single-Precision. +def int_arm_neon_vcvtfp2hf + : Intrinsic<[llvm_v4i16_ty], [llvm_v4f32_ty], [IntrNoMem]>; +def int_arm_neon_vcvthf2fp + : Intrinsic<[llvm_v4f32_ty], [llvm_v4i16_ty], [IntrNoMem]>; + // Narrowing Saturating Vector Moves. def int_arm_neon_vqmovns : Neon_1Arg_Narrow_Intrinsic; def int_arm_neon_vqmovnu : Neon_1Arg_Narrow_Intrinsic; |