Skip to content

[NVPTX] Disable relative lookup tables #159748

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Sep 22, 2025
Merged

[NVPTX] Disable relative lookup tables #159748

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
4651eb7
Add test
nikic Sep 19, 2025
0053783
Disable relative lookup tables on nvptx
nikic Sep 19, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -190,6 +190,11 @@ class NVPTXTTIImpl final : public BasicTTIImplBase<NVPTXTTIImpl> {
void collectKernelLaunchBounds(
const Function &F,
SmallVectorImpl<std::pair<StringRef, int64_t>> &LB) const override;

bool shouldBuildRelLookupTables() const override {
// Self-referential globals are not supported.
return false;
}
};

} // end namespace llvm
Expand Down
32 changes: 32 additions & 0 deletions llvm/test/Transforms/RelLookupTableConverter/nvptx.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,32 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals all --version 5
; RUN: opt < %s -passes=rel-lookup-table-converter -relocation-model=pic -S | FileCheck %s
; REQUIRES: nvptx-registered-target
target triple = "nvptx64-nvidia-cuda"

; Do not produce relative lookup table for nvptx target.

@a1 = internal constant i32 0, align 4
@b1 = internal constant i32 0, align 4
@c1 = internal constant i32 0, align 4
@d1 = internal constant i32 0, align 4

@switch.table = private unnamed_addr constant [3 x ptr] [ptr @a1, ptr @b1, ptr @c1], align 8

;.
; CHECK: @a1 = internal constant i32 0, align 4
; CHECK: @b1 = internal constant i32 0, align 4
; CHECK: @c1 = internal constant i32 0, align 4
; CHECK: @d1 = internal constant i32 0, align 4
; CHECK: @switch.table = private unnamed_addr constant [3 x ptr] [ptr @a1, ptr @b1, ptr @c1], align 8
;.
define ptr @internal_linkage(i32 %cond) {
; CHECK-LABEL: define ptr @internal_linkage(
; CHECK-SAME: i32 [[COND:%.*]]) {
; CHECK-NEXT: [[SWITCH_GEP:%.*]] = getelementptr inbounds [3 x ptr], ptr @switch.table, i32 0, i32 [[COND]]
; CHECK-NEXT: [[RELTABLE_INTRINSIC:%.*]] = load ptr, ptr [[SWITCH_GEP]], align 8
; CHECK-NEXT: ret ptr [[RELTABLE_INTRINSIC]]
;
%switch.gep = getelementptr inbounds [3 x ptr], ptr @switch.table, i32 0, i32 %cond
%switch.load = load ptr, ptr %switch.gep, align 8
ret ptr %switch.load
}