mirror of
https://github.com/c64scene-ar/llvm-6502.git
synced 2024-12-15 04:30:12 +00:00
8b2b8a1835
This update was done with the following bash script:
find test/CodeGen -name "*.ll" | \
while read NAME; do
echo "$NAME"
if ! grep -q "^; *RUN: *llc.*debug" $NAME; then
TEMP=`mktemp -t temp`
cp $NAME $TEMP
sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \
while read FUNC; do
sed -i '' "s/;\(.*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC: *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP
done
sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP
sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP
sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP
sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP
mv $TEMP $NAME
fi
done
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@186280 91177308-0d34-0410-b5e6-96231b3b80d8
78 lines
2.5 KiB
LLVM
78 lines
2.5 KiB
LLVM
; RUN: llc -mcpu=pwr6 -mattr=+altivec -code-model=small < %s | FileCheck %s
|
|
|
|
; Check vector extend load expansion with altivec enabled.
|
|
|
|
target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128-n32:64"
|
|
target triple = "powerpc64-unknown-linux-gnu"
|
|
|
|
; Altivec does not provides an sext intruction, so it expands
|
|
; a set of vector stores (stvx), bytes load/sign expand/store
|
|
; (lbz/stb), and a final vector load (lvx) to load the result
|
|
; extended vector.
|
|
define <16 x i8> @v16si8_sext_in_reg(<16 x i8> %a) {
|
|
%b = trunc <16 x i8> %a to <16 x i4>
|
|
%c = sext <16 x i4> %b to <16 x i8>
|
|
ret <16 x i8> %c
|
|
}
|
|
; CHECK-LABEL: v16si8_sext_in_reg:
|
|
; CHECK: vslb
|
|
; CHECK: vsrab
|
|
; CHECK: blr
|
|
|
|
; The zero extend uses a more clever logic: a vector splat
|
|
; and a logic and to set higher bits to 0.
|
|
define <16 x i8> @v16si8_zext_in_reg(<16 x i8> %a) {
|
|
%b = trunc <16 x i8> %a to <16 x i4>
|
|
%c = zext <16 x i4> %b to <16 x i8>
|
|
ret <16 x i8> %c
|
|
}
|
|
; CHECK-LABEL: v16si8_zext_in_reg:
|
|
; CHECK: vspltisb [[VMASK:[0-9]+]], 15
|
|
; CHECK-NEXT: vand 2, 2, [[VMASK]]
|
|
|
|
; Same as v16si8_sext_in_reg, expands to load/store halfwords (lhz/sth).
|
|
define <8 x i16> @v8si16_sext_in_reg(<8 x i16> %a) {
|
|
%b = trunc <8 x i16> %a to <8 x i8>
|
|
%c = sext <8 x i8> %b to <8 x i16>
|
|
ret <8 x i16> %c
|
|
}
|
|
; CHECK-LABEL: v8si16_sext_in_reg:
|
|
; CHECK: vslh
|
|
; CHECK: vsrah
|
|
; CHECK: blr
|
|
|
|
; Same as v8si16_sext_in_reg, but instead of creating the mask
|
|
; with a splat, loads it from memory.
|
|
define <8 x i16> @v8si16_zext_in_reg(<8 x i16> %a) {
|
|
%b = trunc <8 x i16> %a to <8 x i8>
|
|
%c = zext <8 x i8> %b to <8 x i16>
|
|
ret <8 x i16> %c
|
|
}
|
|
; CHECK-LABEL: v8si16_zext_in_reg:
|
|
; CHECK: ld [[RMASKTOC:[0-9]+]], .LC{{[0-9]+}}@toc(2)
|
|
; CHECK-NEXT: lvx [[VMASK:[0-9]+]], {{[0-9]+}}, [[RMASKTOC]]
|
|
; CHECK-NEXT: vand 2, 2, [[VMASK]]
|
|
|
|
; Same as v16si8_sext_in_reg, expands to load halfword (lha) and
|
|
; store words (stw).
|
|
define <4 x i32> @v4si32_sext_in_reg(<4 x i32> %a) {
|
|
%b = trunc <4 x i32> %a to <4 x i16>
|
|
%c = sext <4 x i16> %b to <4 x i32>
|
|
ret <4 x i32> %c
|
|
}
|
|
; CHECK-LABEL: v4si32_sext_in_reg:
|
|
; CHECK: vslw
|
|
; CHECK: vsraw
|
|
; CHECK: blr
|
|
|
|
; Same as v8si16_sext_in_reg.
|
|
define <4 x i32> @v4si32_zext_in_reg(<4 x i32> %a) {
|
|
%b = trunc <4 x i32> %a to <4 x i16>
|
|
%c = zext <4 x i16> %b to <4 x i32>
|
|
ret <4 x i32> %c
|
|
}
|
|
; CHECK-LABEL: v4si32_zext_in_reg:
|
|
; CHECK: vspltisw [[VMASK:[0-9]+]], -16
|
|
; CHECK-NEXT: vsrw [[VMASK]], [[VMASK]], [[VMASK]]
|
|
; CHECK-NEXT: vand 2, 2, [[VMASK]]
|