Add a more complete example to mlir-reduce docs #116085
Open
+48
−9
Conversation
This file contains 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
|
@llvm/pr-subscribers-mlir Author: Jeremy Kun (j2kun) ChangesThis PR was started because I noticed bugs in the docs:
Then I wanted to test my fixes to be sure I got it right, and I realized that the doc has no properly runnable example. I tried adapting some tests but the use of the @jpienaar do you see anything obvious I'm doing wrong in my example that would cause this stack trace? Full diff: https://github.com/llvm/llvm-project/pull/116085.diff 1 Files Affected:
diff --git a/mlir/docs/Tools/mlir-reduce.md b/mlir/docs/Tools/mlir-reduce.md
index b60cdd799ba0c6..e01b57ef0bab45 100644
--- a/mlir/docs/Tools/mlir-reduce.md
+++ b/mlir/docs/Tools/mlir-reduce.md
@@ -26,27 +26,66 @@ to the tree traversal strategy. The different strategies may lead to different
results and different time complexity. You can run as
`-reduction-tree='traversal-mode=0'` to select the mode for example.
+### Example MLIR input
+
+```mlir
+// query-test.mlir
+func.func @func1() {
+ // A func can be pruned if it's not relevant to the error.
+ return
+}
+
+func.func @func2() -> i32 {
+ %0 = arith.constant 1 : i32
+ %1 = arith.constant 2 : i32
+ %2 = arith.constant 2.2 : f32
+ %3 = arith.constant 5.3 : f32
+ %4 = arith.addi %0, %1 : i32
+ %5 = arith.addf %2, %3 : f32
+ %6 = arith.muli %4, %4 : i32
+ %7 = arith.subi %6, %4 : i32
+ %8 = arith.fptosi %5 : f32 to i32
+ %9 = arith.addi %7, %8 : i32
+ return %9 : i32
+}
+```
+
### Write the script for testing interestingness
-As mentioned, you need to provide a command to `mlir-reduce` which identifies
-cases you're interested in. For each intermediate output generated during
-reduction, `mlir-reduce` will run the command over the it, the script should
-returns 1 for interesting case, 0 otherwise. The sample script,
+You need to provide a command to `mlir-reduce` which identifies cases you're
+interested in. For each intermediate output generated during reduction,
+`mlir-reduce` will run the command over the it, the script should returns 1 for
+interesting case, 0 otherwise. For the IR above, a sample script might simply
+look for the presence of the `arith.fptosi` operation. A more realistic script
+would check for the presence of a particular kind of error message in stderr.
```shell
-mlir-opt -convert-vector-to-spirv $1 | grep "failed to materialize"
-if [[ $? -eq 1 ]]; then
+# query-test.sh
+# `2>&1` redirects stderr (where errors and diagnostics are printed) to stdout
+# so it can be piped to grep.
+mlir-opt $1 2>&1 | grep "arith.fptosi"
+
+# grep's exit code is 0 if the queried string is found
+if [[ $? -eq 0 ]]; then
exit 1
else
exit 0
fi
```
-The sample usage will be like, note that the `test` argument is part of the mode
-argument.
+### Running the example
+
+The sample usage will be as follows, noting that the `test` argument is part of
+the mode argument.
```shell
-mlir-reduce $INPUT -reduction-tree='traversal-mode=0 test=$TEST_SCRIPT'
+mlir-reduce query-test.mlir -reduction-tree='traversal-mode=0 test=query-test.sh'
+```
+
+The output:
+
+```
+TODO: the usage above produces a stack trace
```
## Available reduction strategies
|
j2kun
force-pushed
the
mlir-reduce-docs
branch
from
4c5b388 to
b970411
Compare
jpienaar
reviewed
Mar 7, 2025
|
@jpienaar I'm happy to merge this as-is, if you think it would be an improvement over what's there (even with the TODO?), but I still have questions about how to properly test it. Do you have an example that I can reproduce that only uses upstream dialects? |
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 PR was started because I noticed bugs in the docs:
Then I wanted to test my fixes to be sure I got it right, and I realized that the doc has no properly runnable example. I tried adapting some tests but the use of the
testdialect gave me some issues, so I felt it would be more straightforward to add a test that only uses builtin dialects. Because I could not find an upstream error to demonstrate, I instead just made the test check for the presence of a particular op, and the example I came up with hits a stack trace in the reduce tool:@jpienaar do you see anything obvious I'm doing wrong in my example that would cause this stack trace?