shylie/llvm-project
1
0
Fork 0
Code Issues Pull Requests Actions 6 Packages Projects Releases Wiki Activity
llvm-project/clang/test/OpenMP/align_clause_codegen.cpp
Dhruva Chakrabarti 1c9ec74e3f [Clang][OpenMP] Insert alloca for kernel args at function entry block instead of the launch point. 
If an inlined kernel is called in a loop, the launch point alloca would
lead to increasing stack usage every time the kernel is invoked. This
could make the application run out of stack space and crash. This problem
is fixed by using the alloca insertion point while creating the alloca instruction.

Fixes https://github.com/llvm/llvm-project/issues/60602

Reviewed By: jdoerfert

Differential Revision: https://reviews.llvm.org/D145820
2023-03-17 16:36:12 -04:00

232 lines
15 KiB
C++
Raw Blame History

// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --function-signature --include-generated-funcs
// RUN: %clang_cc1 -emit-llvm -o - -fopenmp \
// RUN: -triple i386-unknown-unknown -fopenmp-version=51 %s | \
// RUN: FileCheck %s --check-prefix=CHECK-32
// RUN: %clang_cc1 -emit-llvm -o - -fopenmp \
// RUN: -triple x86_64-unknown-linux-gnu -fopenmp-version=51 %s | FileCheck %s
// RUN: %clang_cc1 -fopenmp \
// RUN: -triple x86_64-unknown-linux-gnu -fopenmp-version=51 \
// RUN: -emit-pch %s -o %t
// RUN: %clang_cc1 -fopenmp \
// RUN: -triple x86_64-unknown-linux-gnu -fopenmp-version=51 \
// RUN: -include-pch %t -emit-llvm %s -o - | FileCheck %s
// expected-no-diagnostics
#ifndef HEADER
#define HEADER
typedef enum omp_allocator_handle_t {
omp_null_allocator = 0,
omp_default_mem_alloc = 1,
omp_large_cap_mem_alloc = 2,
omp_const_mem_alloc = 3,
omp_high_bw_mem_alloc = 4,
omp_low_lat_mem_alloc = 5,
omp_cgroup_mem_alloc = 6,
omp_pteam_mem_alloc = 7,
omp_thread_mem_alloc = 8,
KMP_ALLOCATOR_MAX_HANDLE = __UINTPTR_MAX__
} omp_allocator_handle_t;
int main() {
int foo0[5];
int foo1[10];
int foo2[20];
int foo3[30];
int foo4[40];
int foo5[50];
int foo6[60];
int foo7[70];
int foo8[80];
omp_allocator_handle_t MyAlloc = omp_large_cap_mem_alloc;
#pragma omp allocate(foo0) align(1)
#pragma omp allocate(foo1) allocator(omp_pteam_mem_alloc) align(2)
#pragma omp allocate(foo2) align(4) allocator(omp_cgroup_mem_alloc)
#pragma omp allocate(foo3) align(8) allocator(omp_low_lat_mem_alloc)
#pragma omp allocate(foo4) align(16) allocator(omp_high_bw_mem_alloc)
#pragma omp allocate(foo5) align(32) allocator(omp_const_mem_alloc)
#pragma omp allocate(foo6) align(64) allocator(omp_large_cap_mem_alloc)
#pragma omp allocate(foo7) align(32) allocator(omp_thread_mem_alloc)
#pragma omp allocate(foo8) align(16) allocator(omp_null_allocator)
{
double foo9[80];
double foo10[90];
#pragma omp allocate(foo9) align(8) allocator(omp_thread_mem_alloc)
#pragma omp allocate(foo10) align(128)
}
{
int bar1;
int bar2[10];
int bar3[20];
int *bar4;
float bar5;
double bar6[30];
#pragma omp allocate(bar1, bar2, bar3) align(2) allocator(MyAlloc)
#pragma omp allocate(bar4, bar5, bar6) align(16)
}
}
// Verify align clause in template with non-type template parameter.
template <typename T, unsigned size, unsigned align>
T run() {
T foo[size];
#pragma omp allocate(foo) align(align) allocator(omp_cgroup_mem_alloc)
return foo[0];
}
int template_test() {
double result;
result = run<double, 1000, 16>();
return 0;
}
#endif
// CHECK-32-LABEL: define {{[^@]+}}@main
// CHECK-32-SAME: () #[[ATTR0:[0-9]+]] {
// CHECK-32-NEXT: entry:
// CHECK-32-NEXT: [[MYALLOC:%.*]] = alloca i32, align 4
// CHECK-32-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1:[0-9]+]])
// CHECK-32-NEXT: [[DOTFOO0__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 4, i32 20, ptr null)
// CHECK-32-NEXT: [[DOTFOO1__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 4, i32 40, ptr inttoptr (i32 7 to ptr))
// CHECK-32-NEXT: [[DOTFOO2__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 4, i32 80, ptr inttoptr (i32 6 to ptr))
// CHECK-32-NEXT: [[DOTFOO3__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 8, i32 120, ptr inttoptr (i32 5 to ptr))
// CHECK-32-NEXT: [[DOTFOO4__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 16, i32 160, ptr inttoptr (i32 4 to ptr))
// CHECK-32-NEXT: [[DOTFOO5__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 32, i32 200, ptr inttoptr (i32 3 to ptr))
// CHECK-32-NEXT: [[DOTFOO6__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 64, i32 240, ptr inttoptr (i32 2 to ptr))
// CHECK-32-NEXT: [[DOTFOO7__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 32, i32 280, ptr inttoptr (i32 8 to ptr))
// CHECK-32-NEXT: [[DOTFOO8__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 16, i32 320, ptr null)
// CHECK-32-NEXT: store i32 2, ptr [[MYALLOC]], align 4
// CHECK-32-NEXT: [[DOTFOO9__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 8, i32 640, ptr inttoptr (i32 8 to ptr))
// CHECK-32-NEXT: [[DOTFOO10__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 128, i32 720, ptr null)
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO10__VOID_ADDR]], ptr null)
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO9__VOID_ADDR]], ptr inttoptr (i32 8 to ptr))
// CHECK-32-NEXT: [[TMP1:%.*]] = load i32, ptr [[MYALLOC]], align 4
// CHECK-32-NEXT: [[CONV:%.*]] = inttoptr i32 [[TMP1]] to ptr
// CHECK-32-NEXT: [[DOTBAR1__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 4, i32 4, ptr [[CONV]])
// CHECK-32-NEXT: [[TMP2:%.*]] = load i32, ptr [[MYALLOC]], align 4
// CHECK-32-NEXT: [[CONV1:%.*]] = inttoptr i32 [[TMP2]] to ptr
// CHECK-32-NEXT: [[DOTBAR2__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 4, i32 40, ptr [[CONV1]])
// CHECK-32-NEXT: [[TMP3:%.*]] = load i32, ptr [[MYALLOC]], align 4
// CHECK-32-NEXT: [[CONV2:%.*]] = inttoptr i32 [[TMP3]] to ptr
// CHECK-32-NEXT: [[DOTBAR3__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 4, i32 80, ptr [[CONV2]])
// CHECK-32-NEXT: [[DOTBAR4__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 16, i32 4, ptr null)
// CHECK-32-NEXT: [[DOTBAR5__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 16, i32 4, ptr null)
// CHECK-32-NEXT: [[DOTBAR6__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 16, i32 240, ptr null)
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTBAR6__VOID_ADDR]], ptr null)
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTBAR5__VOID_ADDR]], ptr null)
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTBAR4__VOID_ADDR]], ptr null)
// CHECK-32-NEXT: [[TMP4:%.*]] = load i32, ptr [[MYALLOC]], align 4
// CHECK-32-NEXT: [[CONV3:%.*]] = inttoptr i32 [[TMP4]] to ptr
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTBAR3__VOID_ADDR]], ptr [[CONV3]])
// CHECK-32-NEXT: [[TMP5:%.*]] = load i32, ptr [[MYALLOC]], align 4
// CHECK-32-NEXT: [[CONV4:%.*]] = inttoptr i32 [[TMP5]] to ptr
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTBAR2__VOID_ADDR]], ptr [[CONV4]])
// CHECK-32-NEXT: [[TMP6:%.*]] = load i32, ptr [[MYALLOC]], align 4
// CHECK-32-NEXT: [[CONV5:%.*]] = inttoptr i32 [[TMP6]] to ptr
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTBAR1__VOID_ADDR]], ptr [[CONV5]])
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO8__VOID_ADDR]], ptr null)
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO7__VOID_ADDR]], ptr inttoptr (i32 8 to ptr))
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO6__VOID_ADDR]], ptr inttoptr (i32 2 to ptr))
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO5__VOID_ADDR]], ptr inttoptr (i32 3 to ptr))
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO4__VOID_ADDR]], ptr inttoptr (i32 4 to ptr))
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO3__VOID_ADDR]], ptr inttoptr (i32 5 to ptr))
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO2__VOID_ADDR]], ptr inttoptr (i32 6 to ptr))
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO1__VOID_ADDR]], ptr inttoptr (i32 7 to ptr))
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO0__VOID_ADDR]], ptr null)
// CHECK-32-NEXT: ret i32 0
//
//
// CHECK-32-LABEL: define {{[^@]+}}@_Z13template_testv
// CHECK-32-SAME: () #[[ATTR2:[0-9]+]] {
// CHECK-32-NEXT: entry:
// CHECK-32-NEXT: [[RESULT:%.*]] = alloca double, align 8
// CHECK-32-NEXT: [[CALL:%.*]] = call noundef double @_Z3runIdLj1000ELj16EET_v()
// CHECK-32-NEXT: store double [[CALL]], ptr [[RESULT]], align 8
// CHECK-32-NEXT: ret i32 0
//
//
// CHECK-32-LABEL: define {{[^@]+}}@_Z3runIdLj1000ELj16EET_v
// CHECK-32-SAME: () #[[ATTR2]] comdat {
// CHECK-32-NEXT: entry:
// CHECK-32-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]])
// CHECK-32-NEXT: [[DOTFOO__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i32 16, i32 8000, ptr inttoptr (i32 6 to ptr))
// CHECK-32-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [1000 x double], ptr [[DOTFOO__VOID_ADDR]], i32 0, i32 0
// CHECK-32-NEXT: [[TMP1:%.*]] = load double, ptr [[ARRAYIDX]], align 8
// CHECK-32-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO__VOID_ADDR]], ptr inttoptr (i32 6 to ptr))
// CHECK-32-NEXT: ret double [[TMP1]]
//
//
// CHECK-LABEL: define {{[^@]+}}@main
// CHECK-SAME: () #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: entry:
// CHECK-NEXT: [[MYALLOC:%.*]] = alloca i64, align 8
// CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1:[0-9]+]])
// CHECK-NEXT: [[DOTFOO0__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 4, i64 32, ptr null)
// CHECK-NEXT: [[DOTFOO1__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 4, i64 48, ptr inttoptr (i64 7 to ptr))
// CHECK-NEXT: [[DOTFOO2__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 4, i64 80, ptr inttoptr (i64 6 to ptr))
// CHECK-NEXT: [[DOTFOO3__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 8, i64 128, ptr inttoptr (i64 5 to ptr))
// CHECK-NEXT: [[DOTFOO4__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 16, i64 160, ptr inttoptr (i64 4 to ptr))
// CHECK-NEXT: [[DOTFOO5__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 32, i64 208, ptr inttoptr (i64 3 to ptr))
// CHECK-NEXT: [[DOTFOO6__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 64, i64 240, ptr inttoptr (i64 2 to ptr))
// CHECK-NEXT: [[DOTFOO7__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 32, i64 288, ptr inttoptr (i64 8 to ptr))
// CHECK-NEXT: [[DOTFOO8__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 16, i64 320, ptr null)
// CHECK-NEXT: store i64 2, ptr [[MYALLOC]], align 8
// CHECK-NEXT: [[DOTFOO9__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 8, i64 640, ptr inttoptr (i64 8 to ptr))
// CHECK-NEXT: [[DOTFOO10__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 128, i64 720, ptr null)
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO10__VOID_ADDR]], ptr null)
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO9__VOID_ADDR]], ptr inttoptr (i64 8 to ptr))
// CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr [[MYALLOC]], align 8
// CHECK-NEXT: [[CONV:%.*]] = inttoptr i64 [[TMP1]] to ptr
// CHECK-NEXT: [[DOTBAR1__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 4, i64 4, ptr [[CONV]])
// CHECK-NEXT: [[TMP2:%.*]] = load i64, ptr [[MYALLOC]], align 8
// CHECK-NEXT: [[CONV1:%.*]] = inttoptr i64 [[TMP2]] to ptr
// CHECK-NEXT: [[DOTBAR2__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 4, i64 48, ptr [[CONV1]])
// CHECK-NEXT: [[TMP3:%.*]] = load i64, ptr [[MYALLOC]], align 8
// CHECK-NEXT: [[CONV2:%.*]] = inttoptr i64 [[TMP3]] to ptr
// CHECK-NEXT: [[DOTBAR3__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 4, i64 80, ptr [[CONV2]])
// CHECK-NEXT: [[DOTBAR4__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 16, i64 8, ptr null)
// CHECK-NEXT: [[DOTBAR5__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 16, i64 4, ptr null)
// CHECK-NEXT: [[DOTBAR6__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 16, i64 240, ptr null)
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTBAR6__VOID_ADDR]], ptr null)
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTBAR5__VOID_ADDR]], ptr null)
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTBAR4__VOID_ADDR]], ptr null)
// CHECK-NEXT: [[TMP4:%.*]] = load i64, ptr [[MYALLOC]], align 8
// CHECK-NEXT: [[CONV3:%.*]] = inttoptr i64 [[TMP4]] to ptr
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTBAR3__VOID_ADDR]], ptr [[CONV3]])
// CHECK-NEXT: [[TMP5:%.*]] = load i64, ptr [[MYALLOC]], align 8
// CHECK-NEXT: [[CONV4:%.*]] = inttoptr i64 [[TMP5]] to ptr
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTBAR2__VOID_ADDR]], ptr [[CONV4]])
// CHECK-NEXT: [[TMP6:%.*]] = load i64, ptr [[MYALLOC]], align 8
// CHECK-NEXT: [[CONV5:%.*]] = inttoptr i64 [[TMP6]] to ptr
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTBAR1__VOID_ADDR]], ptr [[CONV5]])
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO8__VOID_ADDR]], ptr null)
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO7__VOID_ADDR]], ptr inttoptr (i64 8 to ptr))
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO6__VOID_ADDR]], ptr inttoptr (i64 2 to ptr))
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO5__VOID_ADDR]], ptr inttoptr (i64 3 to ptr))
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO4__VOID_ADDR]], ptr inttoptr (i64 4 to ptr))
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO3__VOID_ADDR]], ptr inttoptr (i64 5 to ptr))
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO2__VOID_ADDR]], ptr inttoptr (i64 6 to ptr))
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO1__VOID_ADDR]], ptr inttoptr (i64 7 to ptr))
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO0__VOID_ADDR]], ptr null)
// CHECK-NEXT: ret i32 0
//
//
// CHECK-LABEL: define {{[^@]+}}@_Z13template_testv
// CHECK-SAME: () #[[ATTR2:[0-9]+]] {
// CHECK-NEXT: entry:
// CHECK-NEXT: [[RESULT:%.*]] = alloca double, align 8
// CHECK-NEXT: [[CALL:%.*]] = call noundef double @_Z3runIdLj1000ELj16EET_v()
// CHECK-NEXT: store double [[CALL]], ptr [[RESULT]], align 8
// CHECK-NEXT: ret i32 0
//
//
// CHECK-LABEL: define {{[^@]+}}@_Z3runIdLj1000ELj16EET_v
// CHECK-SAME: () #[[ATTR2]] comdat {
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]])
// CHECK-NEXT: [[DOTFOO__VOID_ADDR:%.*]] = call ptr @__kmpc_aligned_alloc(i32 [[TMP0]], i64 16, i64 8000, ptr inttoptr (i64 6 to ptr))
// CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [1000 x double], ptr [[DOTFOO__VOID_ADDR]], i64 0, i64 0
// CHECK-NEXT: [[TMP1:%.*]] = load double, ptr [[ARRAYIDX]], align 16
// CHECK-NEXT: call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTFOO__VOID_ADDR]], ptr inttoptr (i64 6 to ptr))
// CHECK-NEXT: ret double [[TMP1]]
//
Reference in New Issue View Git Blame Copy Permalink