mirror of
https://github.com/c64scene-ar/llvm-6502.git
synced 2024-07-21 02:29:22 +00:00
7c9c6ed761
Essentially the same as the GEP change in r230786.
A similar migration script can be used to update test cases, though a few more
test case improvements/changes were required this time around: (r229269-r229278)
import fileinput
import sys
import re
pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)")
for line in sys.stdin:
sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line))
Reviewers: rafael, dexonsmith, grosser
Differential Revision: http://reviews.llvm.org/D7649
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@230794 91177308-0d34-0410-b5e6-96231b3b80d8
103 lines
3.0 KiB
LLVM
103 lines
3.0 KiB
LLVM
; RUN: llc < %s -mtriple=s390x-linux-gnu | FileCheck %s
|
|
|
|
declare float @llvm.fma.f32(float %f1, float %f2, float %f3)
|
|
|
|
define float @f1(float %f1, float %f2, float %acc) {
|
|
; CHECK-LABEL: f1:
|
|
; CHECK: maebr %f4, %f0, %f2
|
|
; CHECK: ler %f0, %f4
|
|
; CHECK: br %r14
|
|
%res = call float @llvm.fma.f32 (float %f1, float %f2, float %acc)
|
|
ret float %res
|
|
}
|
|
|
|
define float @f2(float %f1, float *%ptr, float %acc) {
|
|
; CHECK-LABEL: f2:
|
|
; CHECK: maeb %f2, %f0, 0(%r2)
|
|
; CHECK: ler %f0, %f2
|
|
; CHECK: br %r14
|
|
%f2 = load float , float *%ptr
|
|
%res = call float @llvm.fma.f32 (float %f1, float %f2, float %acc)
|
|
ret float %res
|
|
}
|
|
|
|
define float @f3(float %f1, float *%base, float %acc) {
|
|
; CHECK-LABEL: f3:
|
|
; CHECK: maeb %f2, %f0, 4092(%r2)
|
|
; CHECK: ler %f0, %f2
|
|
; CHECK: br %r14
|
|
%ptr = getelementptr float, float *%base, i64 1023
|
|
%f2 = load float , float *%ptr
|
|
%res = call float @llvm.fma.f32 (float %f1, float %f2, float %acc)
|
|
ret float %res
|
|
}
|
|
|
|
define float @f4(float %f1, float *%base, float %acc) {
|
|
; The important thing here is that we don't generate an out-of-range
|
|
; displacement. Other sequences besides this one would be OK.
|
|
;
|
|
; CHECK-LABEL: f4:
|
|
; CHECK: aghi %r2, 4096
|
|
; CHECK: maeb %f2, %f0, 0(%r2)
|
|
; CHECK: ler %f0, %f2
|
|
; CHECK: br %r14
|
|
%ptr = getelementptr float, float *%base, i64 1024
|
|
%f2 = load float , float *%ptr
|
|
%res = call float @llvm.fma.f32 (float %f1, float %f2, float %acc)
|
|
ret float %res
|
|
}
|
|
|
|
define float @f5(float %f1, float *%base, float %acc) {
|
|
; Here too the important thing is that we don't generate an out-of-range
|
|
; displacement. Other sequences besides this one would be OK.
|
|
;
|
|
; CHECK-LABEL: f5:
|
|
; CHECK: aghi %r2, -4
|
|
; CHECK: maeb %f2, %f0, 0(%r2)
|
|
; CHECK: ler %f0, %f2
|
|
; CHECK: br %r14
|
|
%ptr = getelementptr float, float *%base, i64 -1
|
|
%f2 = load float , float *%ptr
|
|
%res = call float @llvm.fma.f32 (float %f1, float %f2, float %acc)
|
|
ret float %res
|
|
}
|
|
|
|
define float @f6(float %f1, float *%base, i64 %index, float %acc) {
|
|
; CHECK-LABEL: f6:
|
|
; CHECK: sllg %r1, %r3, 2
|
|
; CHECK: maeb %f2, %f0, 0(%r1,%r2)
|
|
; CHECK: ler %f0, %f2
|
|
; CHECK: br %r14
|
|
%ptr = getelementptr float, float *%base, i64 %index
|
|
%f2 = load float , float *%ptr
|
|
%res = call float @llvm.fma.f32 (float %f1, float %f2, float %acc)
|
|
ret float %res
|
|
}
|
|
|
|
define float @f7(float %f1, float *%base, i64 %index, float %acc) {
|
|
; CHECK-LABEL: f7:
|
|
; CHECK: sllg %r1, %r3, 2
|
|
; CHECK: maeb %f2, %f0, 4092({{%r1,%r2|%r2,%r1}})
|
|
; CHECK: ler %f0, %f2
|
|
; CHECK: br %r14
|
|
%index2 = add i64 %index, 1023
|
|
%ptr = getelementptr float, float *%base, i64 %index2
|
|
%f2 = load float , float *%ptr
|
|
%res = call float @llvm.fma.f32 (float %f1, float %f2, float %acc)
|
|
ret float %res
|
|
}
|
|
|
|
define float @f8(float %f1, float *%base, i64 %index, float %acc) {
|
|
; CHECK-LABEL: f8:
|
|
; CHECK: sllg %r1, %r3, 2
|
|
; CHECK: lay %r1, 4096({{%r1,%r2|%r2,%r1}})
|
|
; CHECK: maeb %f2, %f0, 0(%r1)
|
|
; CHECK: ler %f0, %f2
|
|
; CHECK: br %r14
|
|
%index2 = add i64 %index, 1024
|
|
%ptr = getelementptr float, float *%base, i64 %index2
|
|
%f2 = load float , float *%ptr
|
|
%res = call float @llvm.fma.f32 (float %f1, float %f2, float %acc)
|
|
ret float %res
|
|
}
|