diff options
author | Bob Wilson <bob.wilson@apple.com> | 2009-06-22 23:27:02 +0000 |
---|---|---|
committer | Bob Wilson <bob.wilson@apple.com> | 2009-06-22 23:27:02 +0000 |
commit | 5bafff36c798608a189c517d37527e4a38863071 (patch) | |
tree | 79bd2abbc5253e6f00db07023cf7d829cbcdee5a /test/CodeGen/ARM/vbsl.ll | |
parent | 5de83afcdc3f4f0edf8caacba523f5d05ee48048 (diff) | |
download | external_llvm-5bafff36c798608a189c517d37527e4a38863071.zip external_llvm-5bafff36c798608a189c517d37527e4a38863071.tar.gz external_llvm-5bafff36c798608a189c517d37527e4a38863071.tar.bz2 |
Add support for ARM's Advanced SIMD (NEON) instruction set.
This is still a work in progress but most of the NEON instruction set
is supported.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73919 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test/CodeGen/ARM/vbsl.ll')
-rw-r--r-- | test/CodeGen/ARM/vbsl.ll | 91 |
1 files changed, 91 insertions, 0 deletions
diff --git a/test/CodeGen/ARM/vbsl.ll b/test/CodeGen/ARM/vbsl.ll new file mode 100644 index 0000000..37ddf4d --- /dev/null +++ b/test/CodeGen/ARM/vbsl.ll @@ -0,0 +1,91 @@ +; RUN: llvm-as < %s | llc -march=arm -mattr=+neon > %t +; RUN: grep vbsl %t | count 8 +; Note: function names do not include "vbsl" to allow simple grep for opcodes + +define <8 x i8> @v_bsli8(<8 x i8>* %A, <8 x i8>* %B, <8 x i8>* %C) nounwind { + %tmp1 = load <8 x i8>* %A + %tmp2 = load <8 x i8>* %B + %tmp3 = load <8 x i8>* %C + %tmp4 = and <8 x i8> %tmp1, %tmp2 + %tmp5 = xor <8 x i8> %tmp1, < i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1 > + %tmp6 = and <8 x i8> %tmp5, %tmp3 + %tmp7 = or <8 x i8> %tmp4, %tmp6 + ret <8 x i8> %tmp7 +} + +define <4 x i16> @v_bsli16(<4 x i16>* %A, <4 x i16>* %B, <4 x i16>* %C) nounwind { + %tmp1 = load <4 x i16>* %A + %tmp2 = load <4 x i16>* %B + %tmp3 = load <4 x i16>* %C + %tmp4 = and <4 x i16> %tmp1, %tmp2 + %tmp5 = xor <4 x i16> %tmp1, < i16 -1, i16 -1, i16 -1, i16 -1 > + %tmp6 = and <4 x i16> %tmp5, %tmp3 + %tmp7 = or <4 x i16> %tmp4, %tmp6 + ret <4 x i16> %tmp7 +} + +define <2 x i32> @v_bsli32(<2 x i32>* %A, <2 x i32>* %B, <2 x i32>* %C) nounwind { + %tmp1 = load <2 x i32>* %A + %tmp2 = load <2 x i32>* %B + %tmp3 = load <2 x i32>* %C + %tmp4 = and <2 x i32> %tmp1, %tmp2 + %tmp5 = xor <2 x i32> %tmp1, < i32 -1, i32 -1 > + %tmp6 = and <2 x i32> %tmp5, %tmp3 + %tmp7 = or <2 x i32> %tmp4, %tmp6 + ret <2 x i32> %tmp7 +} + +define <1 x i64> @v_bsli64(<1 x i64>* %A, <1 x i64>* %B, <1 x i64>* %C) nounwind { + %tmp1 = load <1 x i64>* %A + %tmp2 = load <1 x i64>* %B + %tmp3 = load <1 x i64>* %C + %tmp4 = and <1 x i64> %tmp1, %tmp2 + %tmp5 = xor <1 x i64> %tmp1, < i64 -1 > + %tmp6 = and <1 x i64> %tmp5, %tmp3 + %tmp7 = or <1 x i64> %tmp4, %tmp6 + ret <1 x i64> %tmp7 +} + +define <16 x i8> @v_bslQi8(<16 x i8>* %A, <16 x i8>* %B, <16 x i8>* %C) nounwind { + %tmp1 = load <16 x i8>* %A + %tmp2 = load <16 x i8>* %B + %tmp3 = load <16 x i8>* %C + %tmp4 = and <16 x i8> %tmp1, %tmp2 + %tmp5 = xor <16 x i8> %tmp1, < i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1 > + %tmp6 = and <16 x i8> %tmp5, %tmp3 + %tmp7 = or <16 x i8> %tmp4, %tmp6 + ret <16 x i8> %tmp7 +} + +define <8 x i16> @v_bslQi16(<8 x i16>* %A, <8 x i16>* %B, <8 x i16>* %C) nounwind { + %tmp1 = load <8 x i16>* %A + %tmp2 = load <8 x i16>* %B + %tmp3 = load <8 x i16>* %C + %tmp4 = and <8 x i16> %tmp1, %tmp2 + %tmp5 = xor <8 x i16> %tmp1, < i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1 > + %tmp6 = and <8 x i16> %tmp5, %tmp3 + %tmp7 = or <8 x i16> %tmp4, %tmp6 + ret <8 x i16> %tmp7 +} + +define <4 x i32> @v_bslQi32(<4 x i32>* %A, <4 x i32>* %B, <4 x i32>* %C) nounwind { + %tmp1 = load <4 x i32>* %A + %tmp2 = load <4 x i32>* %B + %tmp3 = load <4 x i32>* %C + %tmp4 = and <4 x i32> %tmp1, %tmp2 + %tmp5 = xor <4 x i32> %tmp1, < i32 -1, i32 -1, i32 -1, i32 -1 > + %tmp6 = and <4 x i32> %tmp5, %tmp3 + %tmp7 = or <4 x i32> %tmp4, %tmp6 + ret <4 x i32> %tmp7 +} + +define <2 x i64> @v_bslQi64(<2 x i64>* %A, <2 x i64>* %B, <2 x i64>* %C) nounwind { + %tmp1 = load <2 x i64>* %A + %tmp2 = load <2 x i64>* %B + %tmp3 = load <2 x i64>* %C + %tmp4 = and <2 x i64> %tmp1, %tmp2 + %tmp5 = xor <2 x i64> %tmp1, < i64 -1, i64 -1 > + %tmp6 = and <2 x i64> %tmp5, %tmp3 + %tmp7 = or <2 x i64> %tmp4, %tmp6 + ret <2 x i64> %tmp7 +} |