shylie/llvm-project
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
llvm-project/clang/test/CodeGen/AArch64/cpu-supports-target.c
Matthew Devereau 22576e2cce
[Clang][AArch64] Add pessimistic vscale_range for sve/sme (#137624) 
The "target-features" function attribute is not currently considered
when adding vscale_range to a function. When +sve/+sme are pushed onto
functions with "#pragma attribute push(+sve/+sme)", the function
potentially misses out on optimizations that rely on vscale_range being
present.
2025-05-16 09:39:07 +01:00

228 lines
11 KiB
C
Raw Permalink Blame History

// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --check-attributes --check-globals all --version 5
// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -emit-llvm -o - %s | FileCheck %s
//.
// CHECK: @__aarch64_cpu_features = external dso_local global { i64 }
//.
// CHECK: Function Attrs: noinline nounwind optnone
// CHECK-LABEL: define dso_local i32 @check_all_features(
// CHECK-SAME: ) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[RETVAL:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr @__aarch64_cpu_features, align 8
// CHECK-NEXT: [[TMP1:%.*]] = and i64 [[TMP0]], 66367
// CHECK-NEXT: [[TMP2:%.*]] = icmp eq i64 [[TMP1]], 66367
// CHECK-NEXT: [[TMP3:%.*]] = and i1 true, [[TMP2]]
// CHECK-NEXT: br i1 [[TMP3]], label %[[IF_THEN:.*]], label %[[IF_ELSE:.*]]
// CHECK: [[IF_THEN]]:
// CHECK-NEXT: store i32 1, ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN:.*]]
// CHECK: [[IF_ELSE]]:
// CHECK-NEXT: [[TMP4:%.*]] = load i64, ptr @__aarch64_cpu_features, align 8
// CHECK-NEXT: [[TMP5:%.*]] = and i64 [[TMP4]], 14272
// CHECK-NEXT: [[TMP6:%.*]] = icmp eq i64 [[TMP5]], 14272
// CHECK-NEXT: [[TMP7:%.*]] = and i1 true, [[TMP6]]
// CHECK-NEXT: br i1 [[TMP7]], label %[[IF_THEN1:.*]], label %[[IF_ELSE2:.*]]
// CHECK: [[IF_THEN1]]:
// CHECK-NEXT: store i32 2, ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN]]
// CHECK: [[IF_ELSE2]]:
// CHECK-NEXT: [[TMP8:%.*]] = load i64, ptr @__aarch64_cpu_features, align 8
// CHECK-NEXT: [[TMP9:%.*]] = and i64 [[TMP8]], 2065152
// CHECK-NEXT: [[TMP10:%.*]] = icmp eq i64 [[TMP9]], 2065152
// CHECK-NEXT: [[TMP11:%.*]] = and i1 true, [[TMP10]]
// CHECK-NEXT: br i1 [[TMP11]], label %[[IF_THEN3:.*]], label %[[IF_ELSE4:.*]]
// CHECK: [[IF_THEN3]]:
// CHECK-NEXT: store i32 3, ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN]]
// CHECK: [[IF_ELSE4]]:
// CHECK-NEXT: [[TMP12:%.*]] = load i64, ptr @__aarch64_cpu_features, align 8
// CHECK-NEXT: [[TMP13:%.*]] = and i64 [[TMP12]], 288230376183169792
// CHECK-NEXT: [[TMP14:%.*]] = icmp eq i64 [[TMP13]], 288230376183169792
// CHECK-NEXT: [[TMP15:%.*]] = and i1 true, [[TMP14]]
// CHECK-NEXT: br i1 [[TMP15]], label %[[IF_THEN5:.*]], label %[[IF_ELSE6:.*]]
// CHECK: [[IF_THEN5]]:
// CHECK-NEXT: store i32 4, ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN]]
// CHECK: [[IF_ELSE6]]:
// CHECK-NEXT: [[TMP16:%.*]] = load i64, ptr @__aarch64_cpu_features, align 8
// CHECK-NEXT: [[TMP17:%.*]] = and i64 [[TMP16]], 1275134720
// CHECK-NEXT: [[TMP18:%.*]] = icmp eq i64 [[TMP17]], 1275134720
// CHECK-NEXT: [[TMP19:%.*]] = and i1 true, [[TMP18]]
// CHECK-NEXT: br i1 [[TMP19]], label %[[IF_THEN7:.*]], label %[[IF_ELSE8:.*]]
// CHECK: [[IF_THEN7]]:
// CHECK-NEXT: store i32 5, ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN]]
// CHECK: [[IF_ELSE8]]:
// CHECK-NEXT: [[TMP20:%.*]] = load i64, ptr @__aarch64_cpu_features, align 8
// CHECK-NEXT: [[TMP21:%.*]] = and i64 [[TMP20]], 52814742272
// CHECK-NEXT: [[TMP22:%.*]] = icmp eq i64 [[TMP21]], 52814742272
// CHECK-NEXT: [[TMP23:%.*]] = and i1 true, [[TMP22]]
// CHECK-NEXT: br i1 [[TMP23]], label %[[IF_THEN9:.*]], label %[[IF_ELSE10:.*]]
// CHECK: [[IF_THEN9]]:
// CHECK-NEXT: store i32 6, ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN]]
// CHECK: [[IF_ELSE10]]:
// CHECK-NEXT: [[TMP24:%.*]] = load i64, ptr @__aarch64_cpu_features, align 8
// CHECK-NEXT: [[TMP25:%.*]] = and i64 [[TMP24]], 344671224576
// CHECK-NEXT: [[TMP26:%.*]] = icmp eq i64 [[TMP25]], 344671224576
// CHECK-NEXT: [[TMP27:%.*]] = and i1 true, [[TMP26]]
// CHECK-NEXT: br i1 [[TMP27]], label %[[IF_THEN11:.*]], label %[[IF_ELSE12:.*]]
// CHECK: [[IF_THEN11]]:
// CHECK-NEXT: store i32 7, ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN]]
// CHECK: [[IF_ELSE12]]:
// CHECK-NEXT: [[TMP28:%.*]] = load i64, ptr @__aarch64_cpu_features, align 8
// CHECK-NEXT: [[TMP29:%.*]] = and i64 [[TMP28]], 3918083994400
// CHECK-NEXT: [[TMP30:%.*]] = icmp eq i64 [[TMP29]], 3918083994400
// CHECK-NEXT: [[TMP31:%.*]] = and i1 true, [[TMP30]]
// CHECK-NEXT: br i1 [[TMP31]], label %[[IF_THEN13:.*]], label %[[IF_ELSE14:.*]]
// CHECK: [[IF_THEN13]]:
// CHECK-NEXT: store i32 8, ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN]]
// CHECK: [[IF_ELSE14]]:
// CHECK-NEXT: [[TMP32:%.*]] = load i64, ptr @__aarch64_cpu_features, align 8
// CHECK-NEXT: [[TMP33:%.*]] = and i64 [[TMP32]], 92359111017216
// CHECK-NEXT: [[TMP34:%.*]] = icmp eq i64 [[TMP33]], 92359111017216
// CHECK-NEXT: [[TMP35:%.*]] = and i1 true, [[TMP34]]
// CHECK-NEXT: br i1 [[TMP35]], label %[[IF_THEN15:.*]], label %[[IF_ELSE16:.*]]
// CHECK: [[IF_THEN15]]:
// CHECK-NEXT: store i32 9, ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN]]
// CHECK: [[IF_ELSE16]]:
// CHECK-NEXT: [[TMP36:%.*]] = load i64, ptr @__aarch64_cpu_features, align 8
// CHECK-NEXT: [[TMP37:%.*]] = and i64 [[TMP36]], 1688849860263936
// CHECK-NEXT: [[TMP38:%.*]] = icmp eq i64 [[TMP37]], 1688849860263936
// CHECK-NEXT: [[TMP39:%.*]] = and i1 true, [[TMP38]]
// CHECK-NEXT: br i1 [[TMP39]], label %[[IF_THEN17:.*]], label %[[IF_ELSE18:.*]]
// CHECK: [[IF_THEN17]]:
// CHECK-NEXT: store i32 10, ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN]]
// CHECK: [[IF_ELSE18]]:
// CHECK-NEXT: [[TMP40:%.*]] = load i64, ptr @__aarch64_cpu_features, align 8
// CHECK-NEXT: [[TMP41:%.*]] = and i64 [[TMP40]], 54047593709241088
// CHECK-NEXT: [[TMP42:%.*]] = icmp eq i64 [[TMP41]], 54047593709241088
// CHECK-NEXT: [[TMP43:%.*]] = and i1 true, [[TMP42]]
// CHECK-NEXT: br i1 [[TMP43]], label %[[IF_THEN19:.*]], label %[[IF_ELSE20:.*]]
// CHECK: [[IF_THEN19]]:
// CHECK-NEXT: store i32 11, ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN]]
// CHECK: [[IF_ELSE20]]:
// CHECK-NEXT: [[TMP44:%.*]] = load i64, ptr @__aarch64_cpu_features, align 8
// CHECK-NEXT: [[TMP45:%.*]] = and i64 [[TMP44]], 216177180294578944
// CHECK-NEXT: [[TMP46:%.*]] = icmp eq i64 [[TMP45]], 216177180294578944
// CHECK-NEXT: [[TMP47:%.*]] = and i1 true, [[TMP46]]
// CHECK-NEXT: br i1 [[TMP47]], label %[[IF_THEN21:.*]], label %[[IF_ELSE22:.*]]
// CHECK: [[IF_THEN21]]:
// CHECK-NEXT: store i32 12, ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN]]
// CHECK: [[IF_ELSE22]]:
// CHECK-NEXT: store i32 0, ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN]]
// CHECK: [[RETURN]]:
// CHECK-NEXT: [[TMP48:%.*]] = load i32, ptr [[RETVAL]], align 4
// CHECK-NEXT: ret i32 [[TMP48]]
//
int check_all_features() {
if (__builtin_cpu_supports("rng+flagm+flagm2+fp16fml+dotprod+sm4"))
return 1;
else if (__builtin_cpu_supports("rdm+lse+fp+simd+crc+sha2+sha3"))
return 2;
else if (__builtin_cpu_supports("aes+fp16+dit+dpb+dpb2+jscvt"))
return 3;
else if (__builtin_cpu_supports("fcma+rcpc+rcpc2+rcpc3+frintts"))
return 4;
else if (__builtin_cpu_supports("i8mm+bf16+sve"))
return 5;
else if (__builtin_cpu_supports("sve+bf16+i8mm+f32mm+f64mm"))
return 6;
else if (__builtin_cpu_supports("sve2+sve2-aes"))
return 7;
else if (__builtin_cpu_supports("sve2-bitperm+sve2-sha3+sve2-sm4"))
return 8;
else if (__builtin_cpu_supports("sme+memtag+sb"))
return 9;
else if (__builtin_cpu_supports("ssbs+bti"))
return 10;
else if (__builtin_cpu_supports("wfxt+sme-f64f64"))
return 11;
else if (__builtin_cpu_supports("sme-i16i64+sme2"))
return 12;
else
return 0;
}
// CHECK: Function Attrs: noinline nounwind optnone
// CHECK-LABEL: define dso_local i32 @neon_code(
// CHECK-SAME: ) #[[ATTR1:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: ret i32 1
//
int __attribute__((target("simd"))) neon_code() { return 1; }
// CHECK: Function Attrs: noinline nounwind optnone
// CHECK-LABEL: define dso_local i32 @sve_code(
// CHECK-SAME: ) #[[ATTR2:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: ret i32 2
//
int __attribute__((target("sve"))) sve_code() { return 2; }
// CHECK: Function Attrs: noinline nounwind optnone
// CHECK-LABEL: define dso_local i32 @code(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: ret i32 3
//
int code() { return 3; }
// CHECK: Function Attrs: noinline nounwind optnone
// CHECK-LABEL: define dso_local i32 @test_versions(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[RETVAL:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[TMP0:%.*]] = load i64, ptr @__aarch64_cpu_features, align 8
// CHECK-NEXT: [[TMP1:%.*]] = and i64 [[TMP0]], 1073807616
// CHECK-NEXT: [[TMP2:%.*]] = icmp eq i64 [[TMP1]], 1073807616
// CHECK-NEXT: [[TMP3:%.*]] = and i1 true, [[TMP2]]
// CHECK-NEXT: br i1 [[TMP3]], label %[[IF_THEN:.*]], label %[[IF_ELSE:.*]]
// CHECK: [[IF_THEN]]:
// CHECK-NEXT: [[CALL:%.*]] = call i32 @sve_code()
// CHECK-NEXT: store i32 [[CALL]], ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN:.*]]
// CHECK: [[IF_ELSE]]:
// CHECK-NEXT: [[TMP4:%.*]] = load i64, ptr @__aarch64_cpu_features, align 8
// CHECK-NEXT: [[TMP5:%.*]] = and i64 [[TMP4]], 768
// CHECK-NEXT: [[TMP6:%.*]] = icmp eq i64 [[TMP5]], 768
// CHECK-NEXT: [[TMP7:%.*]] = and i1 true, [[TMP6]]
// CHECK-NEXT: br i1 [[TMP7]], label %[[IF_THEN1:.*]], label %[[IF_ELSE3:.*]]
// CHECK: [[IF_THEN1]]:
// CHECK-NEXT: [[CALL2:%.*]] = call i32 @neon_code()
// CHECK-NEXT: store i32 [[CALL2]], ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN]]
// CHECK: [[IF_ELSE3]]:
// CHECK-NEXT: [[CALL4:%.*]] = call i32 @code()
// CHECK-NEXT: store i32 [[CALL4]], ptr [[RETVAL]], align 4
// CHECK-NEXT: br label %[[RETURN]]
// CHECK: [[RETURN]]:
// CHECK-NEXT: [[TMP8:%.*]] = load i32, ptr [[RETVAL]], align 4
// CHECK-NEXT: ret i32 [[TMP8]]
//
int test_versions() {
if (__builtin_cpu_supports("sve"))
return sve_code();
else if (__builtin_cpu_supports("simd"))
return neon_code();
else
return code();
}
//.
// CHECK: attributes #[[ATTR0]] = { noinline nounwind optnone "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
// CHECK: attributes #[[ATTR1]] = { noinline nounwind optnone "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+fp-armv8,+neon" }
// CHECK: attributes #[[ATTR2]] = { noinline nounwind optnone vscale_range(1,16) "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+fp-armv8,+fullfp16,+sve" }
//.
// CHECK: [[META0:![0-9]+]] = !{i32 1, !"wchar_size", i32 4}
// CHECK: [[META1:![0-9]+]] = !{!"{{.*}}clang version {{.*}}"}
//.
Reference in New Issue View Git Blame Copy Permalink