[PGO][Offload] Update PGO GPU tests #132262
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
65aae01 to
10cb0f6
Compare
|
@llvm/pr-subscribers-offload Author: Ethan Luis McDonough (EthanLuisMcDonough) ChangesThis pull request aims to improve the reliability of pgo1.c and pgo2.c Full diff: https://github.com/llvm/llvm-project/pull/132262.diff 2 Files Affected:
diff --git a/offload/test/offloading/gpupgo/pgo1.c b/offload/test/offloading/gpupgo/pgo1.c
index c8011cbae83c0..4fb11bf702aba 100644
--- a/offload/test/offloading/gpupgo/pgo1.c
+++ b/offload/test/offloading/gpupgo/pgo1.c
@@ -1,5 +1,6 @@
// RUN: %libomptarget-compile-generic -fcreate-profile \
-// RUN: -Xarch_device -fprofile-generate
+// RUN: -Xarch_device -fprofile-generate \
+// RUN: -Xarch_device -fprofile-update=atomic
// RUN: env LLVM_PROFILE_FILE=%basename_t.llvm.profraw \
// RUN: %libomptarget-run-generic 2>&1
// RUN: llvm-profdata show --all-functions --counts \
@@ -7,7 +8,8 @@
// RUN: %fcheck-generic --check-prefix="LLVM-PGO"
// RUN: %libomptarget-compile-generic -fcreate-profile \
-// RUN: -Xarch_device -fprofile-instr-generate
+// RUN: -Xarch_device -fprofile-instr-generate \
+// RUN: -Xarch_device -fprofile-update=atomic
// RUN: env LLVM_PROFILE_FILE=%basename_t.clang.profraw \
// RUN: %libomptarget-run-generic 2>&1
// RUN: llvm-profdata show --all-functions --counts \
@@ -23,10 +25,12 @@ int test2(int a) { return a * 2; }
int main() {
int m = 2;
#pragma omp target
- for (int i = 0; i < 10; i++) {
- m = test1(m);
- for (int j = 0; j < 2; j++) {
- m = test2(m);
+ {
+ for (int i = 0; i < 10; i++) {
+ m = test1(m);
+ for (int j = 0; j < 2; j++) {
+ m = test2(m);
+ }
}
}
}
@@ -34,7 +38,7 @@ int main() {
// LLVM-PGO-LABEL: __omp_offloading_{{[_0-9a-zA-Z]*}}_main_{{[_0-9a-zA-Z]*}}:
// LLVM-PGO: Hash: {{0[xX][0-9a-fA-F]+}}
// LLVM-PGO: Counters: 4
-// LLVM-PGO: Block counts: [20, 10, 2, 1]
+// LLVM-PGO: Block counts: [20, 10, {{.*}}, 1]
// LLVM-PGO-LABEL: test1:
// LLVM-PGO: Hash: {{0[xX][0-9a-fA-F]+}}
@@ -53,14 +57,10 @@ int main() {
// LLVM-PGO-SAME: 3
// LLVM-PGO-LABEL: Maximum function count:
// LLVM-PGO-SAME: 20
-// LLVM-PGO-LABEL: Maximum internal block count:
-// LLVM-PGO-SAME: 10
// CLANG-PGO-LABEL: __omp_offloading_{{[_0-9a-zA-Z]*}}_main_{{[_0-9a-zA-Z]*}}:
// CLANG-PGO: Hash: {{0[xX][0-9a-fA-F]+}}
-// CLANG-PGO: Counters: 3
-// CLANG-PGO: Function count: 0
-// CLANG-PGO: Block counts: [11, 20]
+// CLANG-PGO: Block counts: [10, 20]
// CLANG-PGO-LABEL: test1:
// CLANG-PGO: Hash: {{0[xX][0-9a-fA-F]+}}
@@ -78,7 +78,5 @@ int main() {
// CLANG-PGO-SAME: Front-end
// CLANG-PGO-LABEL: Functions shown:
// CLANG-PGO-SAME: 3
-// CLANG-PGO-LABEL: Maximum function count:
-// CLANG-PGO-SAME: 20
// CLANG-PGO-LABEL: Maximum internal block count:
// CLANG-PGO-SAME: 20
diff --git a/offload/test/offloading/gpupgo/pgo2.c b/offload/test/offloading/gpupgo/pgo2.c
index b75b0beaffdec..820a597bd5108 100644
--- a/offload/test/offloading/gpupgo/pgo2.c
+++ b/offload/test/offloading/gpupgo/pgo2.c
@@ -1,4 +1,5 @@
-// RUN: %libomptarget-compile-generic -fprofile-generate
+// RUN: %libomptarget-compile-generic -fprofile-generate \
+// RUN: -fprofile-update=atomic
// RUN: env LLVM_PROFILE_FILE=%basename_t.llvm.profraw \
// RUN: %libomptarget-run-generic 2>&1
// RUN: llvm-profdata show --all-functions --counts \
@@ -8,7 +9,8 @@
// RUN: %target_triple.%basename_t.llvm.profraw \
// RUN: | %fcheck-generic --check-prefix="LLVM-DEVICE"
-// RUN: %libomptarget-compile-generic -fprofile-instr-generate
+// RUN: %libomptarget-compile-generic -fprofile-instr-generate \
+// RUN: -fprofile-update=atomic
// RUN: env LLVM_PROFILE_FILE=%basename_t.clang.profraw \
// RUN: %libomptarget-run-generic 2>&1
// RUN: llvm-profdata show --all-functions --counts \
@@ -18,7 +20,8 @@
// RUN: %target_triple.%basename_t.clang.profraw | \
// RUN: %fcheck-generic --check-prefix="CLANG-DEV"
-// RUN: %libomptarget-compile-generic -Xarch_host -fprofile-generate
+// RUN: %libomptarget-compile-generic -Xarch_host -fprofile-generate \
+// RUN: -fprofile-update=atomic
// RUN: env LLVM_PROFILE_FILE=%basename_t.nogpu.profraw \
// RUN: %libomptarget-run-generic 2>&1
// RUN: llvm-profdata show --all-functions --counts \
@@ -27,7 +30,7 @@
// RUN: not test -e %target_triple.%basename_t.nogpu.profraw
// RUN: %libomptarget-compile-generic -Xarch_host -fprofile-generate \
-// RUN: -Xarch_device -fprofile-instr-generate
+// RUN: -Xarch_device -fprofile-instr-generate -fprofile-update=atomic
// RUN: env LLVM_PROFILE_FILE=%basename_t.hidf.profraw \
// RUN: %libomptarget-run-generic 2>&1
// RUN: llvm-profdata show --all-functions --counts \
@@ -38,7 +41,7 @@
// RUN: | %fcheck-generic --check-prefix="CLANG-DEV"
// RUN: %libomptarget-compile-generic -Xarch_device -fprofile-generate \
-// RUN: -Xarch_host -fprofile-instr-generate
+// RUN: -Xarch_host -fprofile-instr-generate -fprofile-update=atomic
// RUN: env LLVM_PROFILE_FILE=%basename_t.hfdi.profraw \
// RUN: %libomptarget-run-generic 2>&1
// RUN: llvm-profdata show --all-functions --counts \
@@ -59,8 +62,10 @@ int main() {
int device_var = 1;
#pragma omp target
- for (int i = 0; i < 10; i++) {
- device_var *= i;
+ {
+ for (int i = 0; i < 10; i++) {
+ device_var *= i;
+ }
}
}
@@ -78,7 +83,7 @@ int main() {
// LLVM-DEVICE-LABEL: __omp_offloading_{{[_0-9a-zA-Z]*}}_main_{{[_0-9a-zA-Z]*}}:
// LLVM-DEVICE: Hash: {{0[xX][0-9a-fA-F]+}}
// LLVM-DEVICE: Counters: 3
-// LLVM-DEVICE: Block counts: [10, 2, 1]
+// LLVM-DEVICE: Block counts: [10, {{.*}}, 1]
// LLVM-DEVICE: Instrumentation level: IR
// CLANG-HOST-LABEL: main:
@@ -97,6 +102,5 @@ int main() {
// CLANG-DEV-LABEL: __omp_offloading_{{[_0-9a-zA-Z]*}}_main_{{[_0-9a-zA-Z]*}}:
// CLANG-DEV: Hash: {{0[xX][0-9a-fA-F]+}}
// CLANG-DEV: Counters: 2
-// CLANG-DEV: Function count: 0
-// CLANG-DEV: Block counts: [11]
+// CLANG-DEV: Block counts: [10]
// CLANG-DEV: Instrumentation level: Front-end
|
jhuber6
approved these changes
Mar 20, 2025
There was a problem hiding this comment.
Normal pragma omp target is a little weird since it executes with multiple threads but masks all of them off except one. It's somewhat possible that this would interact with something? Also, AFAIK, all aligned accesses on AMDGPU should be 'stomic' at least in terms out the output. If this lowers to fine-grained memory maybe not.
|
I talked to @EthanLuisMcDonough but I want to repeat it here: We need to modify the exisiting (single-threaded!) tests to only look at the counters for user code. Those need to be stable w/o atomic accesses in a single-threaded environment. |
jdoerfert
approved these changes
May 7, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This pull request aims to improve the reliability of pgo1.c and pgo2.c