llvm-project/llvm/test/Transforms/OpenMP/single_threaded_execution.ll
Johannes Doerfert 3de645efe3 [OpenMP][NFC] Split the reduction buffer size into two components
Before we tracked the size of the teams reduction buffer in order to
allocate it at runtime per kernel launch. This patch splits the number
into two parts, the size of the reduction data (=all reduction
variables) and the (maximal) length of the buffer. This will allow us to
allocate less if we need less, e.g., if we have less teams than the
maximal length. It also allows us to move code from clangs codegen into
the runtime as we now know how large the reduction data is.
2023-11-06 11:50:41 -08:00

132 lines
4.8 KiB
LLVM

; RUN: opt -passes=openmp-opt -debug-only=openmp-opt -disable-output < %s 2>&1 | FileCheck %s
; RUN: opt -passes=openmp-opt -pass-remarks-analysis=openmp-opt -disable-output < %s 2>&1 | FileCheck %s --check-prefix=REMARKS
; REQUIRES: asserts
; ModuleID = 'single_threaded_exeuction.c'
%struct.ident_t = type { i32, i32, i32, i32, ptr }
%struct.KernelEnvironmentTy = type { %struct.ConfigurationEnvironmentTy, ptr, ptr }
%struct.ConfigurationEnvironmentTy = type { i8, i8, i8, i32, i32, i32, i32, i32, i32 }
@0 = private unnamed_addr constant [1 x i8] c"\00", align 1
@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, ptr @0 }, align 8
@kernel_kernel_environment = local_unnamed_addr constant %struct.KernelEnvironmentTy { %struct.ConfigurationEnvironmentTy { i8 0, i8 0, i8 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0 }, ptr @1, ptr null }
; CHECK-NOT: [openmp-opt] Basic block @kernel entry is executed by a single thread.
; CHECK: [openmp-opt] Basic block @kernel if.then is executed by a single thread.
; CHECK-NOT: [openmp-opt] Basic block @kernel if.else is executed by a single thread.
; CHECK-NOT: [openmp-opt] Basic block @kernel if.end is executed by a single thread.
define void @kernel(ptr %dyn) "kernel" {
%call = call i32 @__kmpc_target_init(ptr @kernel_kernel_environment, ptr %dyn)
%cmp = icmp eq i32 %call, -1
br i1 %cmp, label %if.then, label %if.else
if.then:
br label %if.end
if.else:
br label %if.end
if.end:
call void @__kmpc_target_deinit()
ret void
}
; CHECK: [openmp-opt] Basic block @foo entry is executed by a single thread.
; Function Attrs: noinline
define internal void @foo() {
entry:
ret void
}
; CHECK: [openmp-opt] Basic block @bar.internalized entry is executed by a single thread.
; Function Attrs: noinline
define void @bar() {
entry:
ret void
}
; REMARKS: remark: single_threaded_execution.c:1:0: Could not internalize function. Some optimizations may not be possible.
; REMARKS-NOT: remark: single_threaded_execution.c:1:0: Could not internalize function. Some optimizations may not be possible.
; CHECK-NOT: [openmp-opt] Basic block @nvptx entry is executed by a single thread.
; CHECK-DAG: [openmp-opt] Basic block @nvptx if.then is executed by a single thread.
; CHECK-NOT: [openmp-opt] Basic block @nvptx if.end is executed by a single thread.
; Function Attrs: noinline
define void @nvptx() {
entry:
%call = call i32 @llvm.nvvm.read.ptx.sreg.tid.x()
%cmp = icmp eq i32 %call, 0
br i1 %cmp, label %if.then, label %if.end
if.then:
call void @foo()
call void @bar()
call void @baz()
call void @cold()
br label %if.end
if.end:
ret void
}
; CHECK-NOT: [openmp-opt] Basic block @amdgcn entry is executed by a single thread.
; CHECK-DAG: [openmp-opt] Basic block @amdgcn if.then is executed by a single thread.
; CHECK-NOT: [openmp-opt] Basic block @amdgcn if.end is executed by a single thread.
; Function Attrs: noinline
define void @amdgcn() {
entry:
%call = call i32 @llvm.amdgcn.workitem.id.x()
%cmp = icmp eq i32 %call, 0
br i1 %cmp, label %if.then, label %if.end
if.then:
call void @foo()
call void @bar()
call void @baz()
call void @cold()
br label %if.end
if.end:
ret void
}
; CHECK-NOT: [openmp-opt] Basic block @baz entry is executed by a single thread.
; Function Attrs: noinline
define weak void @baz() !dbg !8 {
entry:
ret void
}
; CHECK-NOT: [openmp-opt] Basic block @cold entry is executed by a single thread.
; Function Attrs: cold convergent noinline nounwind optnone mustprogress
define weak void @cold() #0 !dbg !9 {
entry:
ret void
}
declare i32 @llvm.nvvm.read.ptx.sreg.tid.x()
declare i32 @llvm.amdgcn.workitem.id.x()
declare void @__kmpc_kernel_prepare_parallel(ptr)
declare i32 @__kmpc_target_init(ptr, ptr)
declare void @__kmpc_target_deinit()
attributes #0 = { cold noinline }
!llvm.dbg.cu = !{!0}
!llvm.module.flags = !{!3, !4, !5, !6}
!nvvm.annotations = !{!7}
!0 = distinct !DICompileUnit(language: DW_LANG_C99, file: !1, producer: "clang version 12.0.0", isOptimized: false, runtimeVersion: 0, emissionKind: FullDebug, enums: !2, splitDebugInlining: false, nameTableKind: None)
!1 = !DIFile(filename: "single_threaded_execution.c", directory: "/tmp/single_threaded_execution.c")
!2 = !{}
!3 = !{i32 2, !"Debug Info Version", i32 3}
!4 = !{i32 1, !"wchar_size", i32 4}
!5 = !{i32 7, !"openmp", i32 50}
!6 = !{i32 7, !"openmp-device", i32 50}
!7 = !{ptr @kernel, !"kernel", i32 1}
!8 = distinct !DISubprogram(name: "bar", scope: !1, file: !1, line: 8, type: !10, scopeLine: 1, flags: DIFlagPrototyped, spFlags: DISPFlagDefinition | DISPFlagOptimized, unit: !0, retainedNodes: !2)
!9 = distinct !DISubprogram(name: "cold", scope: !1, file: !1, line: 8, type: !10, scopeLine: 2, flags: DIFlagPrototyped, spFlags: DISPFlagDefinition | DISPFlagOptimized, unit: !0, retainedNodes: !2)
!10 = !DISubroutineType(types: !2)