d10b76552f
The data-layout independent constant folding currently has some rather gnarly code for canonicalizing GEP indices to reduce "notional overindexing", and then infers inbounds based on that canonicalization. Now that we canonicalize to i8 GEPs, this canonicalization is essentially useless, as we'll discard it as soon as the GEP hits the data-layout aware constant folder anyway. As such, I'd like to remove this code entirely. This shouldn't have any impact on optimization capabilities.
31 lines
1.6 KiB
LLVM
31 lines
1.6 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
|
|
; RUN: opt -S -mtriple=nvptx64-nvidia-cuda -passes=infer-address-spaces %s | FileCheck %s
|
|
|
|
target datalayout = "e-i64:64-v16:16-v32:32-n16:32:64"
|
|
|
|
%struct.bar = type { float, ptr }
|
|
|
|
@var1 = local_unnamed_addr addrspace(3) externally_initialized global %struct.bar undef, align 8
|
|
|
|
define void @bug31948(float %a, ptr nocapture readnone %x, ptr nocapture readnone %y) local_unnamed_addr #0 {
|
|
; CHECK-LABEL: define void @bug31948(
|
|
; CHECK-SAME: float [[A:%.*]], ptr nocapture readnone [[X:%.*]], ptr nocapture readnone [[Y:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] {
|
|
; CHECK-NEXT: [[ENTRY:.*:]]
|
|
; CHECK-NEXT: [[TMP:%.*]] = load ptr, ptr addrspace(3) getelementptr ([[STRUCT_BAR:%.*]], ptr addrspace(3) @var1, i64 0, i32 1), align 8
|
|
; CHECK-NEXT: [[TMP1:%.*]] = load float, ptr [[TMP]], align 4
|
|
; CHECK-NEXT: [[CONV1:%.*]] = fadd float [[TMP1]], 1.000000e+00
|
|
; CHECK-NEXT: store float [[CONV1]], ptr [[TMP]], align 4
|
|
; CHECK-NEXT: store i32 32, ptr addrspace(3) getelementptr ([[STRUCT_BAR]], ptr addrspace(3) @var1, i64 0, i32 1), align 4
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
entry:
|
|
%tmp = load ptr, ptr getelementptr (%struct.bar, ptr addrspacecast (ptr addrspace(3) @var1 to ptr), i64 0, i32 1), align 8
|
|
%tmp1 = load float, ptr %tmp, align 4
|
|
%conv1 = fadd float %tmp1, 1.000000e+00
|
|
store float %conv1, ptr %tmp, align 4
|
|
store i32 32, ptr getelementptr (%struct.bar, ptr addrspacecast (ptr addrspace(3) @var1 to ptr), i64 0, i32 1), align 4
|
|
ret void
|
|
}
|
|
|
|
attributes #0 = { norecurse nounwind }
|