llvm-for-llvmta/test/CodeGen/SystemZ/vec-bswap-01.ll

98 lines
2.8 KiB
LLVM

; Test loads of byte-swapped vector elements.
;
; RUN: llc < %s -mtriple=s390x-linux-gnu -mcpu=z15 | FileCheck %s
declare <8 x i16> @llvm.bswap.v8i16(<8 x i16>)
declare <4 x i32> @llvm.bswap.v4i32(<4 x i32>)
declare <2 x i64> @llvm.bswap.v2i64(<2 x i64>)
; Test v8i16 loads.
define <8 x i16> @f1(<8 x i16> *%ptr) {
; CHECK-LABEL: f1:
; CHECK: vlbrh %v24, 0(%r2)
; CHECK: br %r14
%load = load <8 x i16>, <8 x i16> *%ptr
%ret = call <8 x i16> @llvm.bswap.v8i16(<8 x i16> %load)
ret <8 x i16> %ret
}
; Test v4i32 loads.
define <4 x i32> @f2(<4 x i32> *%ptr) {
; CHECK-LABEL: f2:
; CHECK: vlbrf %v24, 0(%r2)
; CHECK: br %r14
%load = load <4 x i32>, <4 x i32> *%ptr
%ret = call <4 x i32> @llvm.bswap.v4i32(<4 x i32> %load)
ret <4 x i32> %ret
}
; Test v2i64 loads.
define <2 x i64> @f3(<2 x i64> *%ptr) {
; CHECK-LABEL: f3:
; CHECK: vlbrg %v24, 0(%r2)
; CHECK: br %r14
%load = load <2 x i64>, <2 x i64> *%ptr
%ret = call <2 x i64> @llvm.bswap.v2i64(<2 x i64> %load)
ret <2 x i64> %ret
}
; Test the highest aligned in-range offset.
define <4 x i32> @f4(<4 x i32> *%base) {
; CHECK-LABEL: f4:
; CHECK: vlbrf %v24, 4080(%r2)
; CHECK: br %r14
%ptr = getelementptr <4 x i32>, <4 x i32> *%base, i64 255
%load = load <4 x i32>, <4 x i32> *%ptr
%ret = call <4 x i32> @llvm.bswap.v4i32(<4 x i32> %load)
ret <4 x i32> %ret
}
; Test the highest unaligned in-range offset.
define <4 x i32> @f5(i8 *%base) {
; CHECK-LABEL: f5:
; CHECK: vlbrf %v24, 4095(%r2)
; CHECK: br %r14
%addr = getelementptr i8, i8 *%base, i64 4095
%ptr = bitcast i8 *%addr to <4 x i32> *
%load = load <4 x i32>, <4 x i32> *%ptr
%ret = call <4 x i32> @llvm.bswap.v4i32(<4 x i32> %load)
ret <4 x i32> %ret
}
; Test the next offset up, which requires separate address logic,
define <4 x i32> @f6(<4 x i32> *%base) {
; CHECK-LABEL: f6:
; CHECK: aghi %r2, 4096
; CHECK: vlbrf %v24, 0(%r2)
; CHECK: br %r14
%ptr = getelementptr <4 x i32>, <4 x i32> *%base, i64 256
%load = load <4 x i32>, <4 x i32> *%ptr
%ret = call <4 x i32> @llvm.bswap.v4i32(<4 x i32> %load)
ret <4 x i32> %ret
}
; Test negative offsets, which also require separate address logic,
define <4 x i32> @f7(<4 x i32> *%base) {
; CHECK-LABEL: f7:
; CHECK: aghi %r2, -16
; CHECK: vlbrf %v24, 0(%r2)
; CHECK: br %r14
%ptr = getelementptr <4 x i32>, <4 x i32> *%base, i64 -1
%load = load <4 x i32>, <4 x i32> *%ptr
%ret = call <4 x i32> @llvm.bswap.v4i32(<4 x i32> %load)
ret <4 x i32> %ret
}
; Check that indexes are allowed.
define <4 x i32> @f8(i8 *%base, i64 %index) {
; CHECK-LABEL: f8:
; CHECK: vlbrf %v24, 0(%r3,%r2)
; CHECK: br %r14
%addr = getelementptr i8, i8 *%base, i64 %index
%ptr = bitcast i8 *%addr to <4 x i32> *
%load = load <4 x i32>, <4 x i32> *%ptr
%ret = call <4 x i32> @llvm.bswap.v4i32(<4 x i32> %load)
ret <4 x i32> %ret
}