[mlir] Add op printing flag to skip regions #77726
Merged
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
kuhar
requested review from
antiagainst,
MaheshRavishankar,
Hardcode84,
joker-eph,
Mogball,
qedawkins and
River707
|
@llvm/pr-subscribers-mlir-core Author: Jakub Kuderski (kuhar) ChangesThe new flag, Example: // -----// IR Dump Before CSE (cse) //----- //
func.func @<!-- -->bar(%arg0: f32, %arg1: f32) -> f32 {...}
// -----// IR Dump Before Canonicalizer (canonicalize) //----- //
func.func @<!-- -->bar(%arg0: f32, %arg1: f32) -> f32 {...}The main use-case is to be triage compilation issues (crashes, slowness) on very deep pass pipelines and with very large IR files, where printing IR is prohibitively slow otherwise. Unlike the pass statistics mechanism, the new flag can be used as an 'interactive' progress indicator. Full diff: https://github.com/llvm/llvm-project/pull/77726.diff 3 Files Affected:
diff --git a/mlir/docs/PassManagement.md b/mlir/docs/PassManagement.md
index 95a38207b7f854..a303e350d28b6b 100644
--- a/mlir/docs/PassManagement.md
+++ b/mlir/docs/PassManagement.md
@@ -1302,6 +1302,27 @@ func.func @simple_constant() -> (i32, i32) {
}
```
+* `mlir-print-ir-skip-regions`
+ * Do not print op regions.
+ * This option is useful to when we are interested in the names of the
+ executed passes but do not want to inspect the IR, e.g., when the
+ module is very large and printing is prohibitively slow.
+
+```shell
+$ mlir-opt foo.mlir -mlir-disable-threading -pass-pipeline='func.func(canonicalize,cse)' -mlir-print-ir-after-all -mlir-print-ir-skip-regions
+// -----// IR Dump Before CSE (cse) //----- //
+func.func @bar(%arg0: f32, %arg1: f32) -> f32 {...}
+
+// -----// IR Dump Before Canonicalizer (canonicalize) //----- //
+func.func @bar(%arg0: f32, %arg1: f32) -> f32 {...}
+
+// -----// IR Dump Before CSE (cse) //----- //
+func.func @simple_constant() -> (i32, i32) {...}
+
+// -----// IR Dump Before Canonicalizer (canonicalize) //----- //
+func.func @simple_constant() -> (i32, i32) {...}
+```
+
## Crash and Failure Reproduction
The [pass manager](#pass-manager) in MLIR contains a builtin mechanism to
diff --git a/mlir/lib/Pass/PassManagerOptions.cpp b/mlir/lib/Pass/PassManagerOptions.cpp
index ffc53b7e3ed023..da7f27fb75a29c 100644
--- a/mlir/lib/Pass/PassManagerOptions.cpp
+++ b/mlir/lib/Pass/PassManagerOptions.cpp
@@ -58,6 +58,12 @@ struct PassManagerOptions {
llvm::cl::desc("When printing IR for print-ir-[before|after]{-all} "
"always print the top-level operation"),
llvm::cl::init(false)};
+ llvm::cl::opt<bool> printSkipRegions{
+ "mlir-print-ir-skip-regions",
+ llvm::cl::desc(
+ "When printing IR for print-ir-[beforef|after]{-all} do not output "
+ "regions. This is useful to indicate pass pipeline progress."),
+ llvm::cl::init(false)};
/// Add an IR printing instrumentation if enabled by any 'print-ir' flags.
void addPrinterInstrumentation(PassManager &pm);
@@ -119,10 +125,14 @@ void PassManagerOptions::addPrinterInstrumentation(PassManager &pm) {
if (!shouldPrintBeforePass && !shouldPrintAfterPass)
return;
+ OpPrintingFlags opPrintingFlags;
+ if (printSkipRegions)
+ opPrintingFlags.skipRegions();
+
// Otherwise, add the IR printing instrumentation.
pm.enableIRPrinting(shouldPrintBeforePass, shouldPrintAfterPass,
printModuleScope, printAfterChange, printAfterFailure,
- llvm::errs());
+ llvm::errs(), opPrintingFlags);
}
void mlir::registerPassManagerCLOptions() {
diff --git a/mlir/test/Pass/ir-printing.mlir b/mlir/test/Pass/ir-printing.mlir
index 048b721ba6d530..cc2f277d76ee52 100644
--- a/mlir/test/Pass/ir-printing.mlir
+++ b/mlir/test/Pass/ir-printing.mlir
@@ -3,6 +3,7 @@
// RUN: mlir-opt %s -mlir-disable-threading=true -pass-pipeline='builtin.module(func.func(cse,canonicalize))' -mlir-print-ir-after=cse -o /dev/null 2>&1 | FileCheck -check-prefix=AFTER %s
// RUN: mlir-opt %s -mlir-disable-threading=true -pass-pipeline='builtin.module(func.func(cse,canonicalize))' -mlir-print-ir-after-all -o /dev/null 2>&1 | FileCheck -check-prefix=AFTER_ALL %s
// RUN: mlir-opt %s -mlir-disable-threading=true -pass-pipeline='builtin.module(func.func(cse,canonicalize))' -mlir-print-ir-before=cse -mlir-print-ir-module-scope -o /dev/null 2>&1 | FileCheck -check-prefix=BEFORE_MODULE %s
+// RUN: mlir-opt %s -mlir-disable-threading=true -pass-pipeline='builtin.module(func.func(cse,canonicalize))' -mlir-print-ir-before-all -mlir-print-ir-skip-regions -o /dev/null 2>&1 | FileCheck -check-prefix=BEFORE_ALL_SKIP_REGIONS %s
// RUN: mlir-opt %s -mlir-disable-threading=true -pass-pipeline='builtin.module(func.func(cse,cse))' -mlir-print-ir-after-all -mlir-print-ir-after-change -o /dev/null 2>&1 | FileCheck -check-prefix=AFTER_ALL_CHANGE %s
// RUN: not mlir-opt %s -mlir-disable-threading=true -pass-pipeline='builtin.module(func.func(cse,test-pass-failure))' -mlir-print-ir-after-failure -o /dev/null 2>&1 | FileCheck -check-prefix=AFTER_FAILURE %s
@@ -56,6 +57,16 @@ func.func @bar() {
// BEFORE_MODULE: func @foo()
// BEFORE_MODULE: func @bar()
+// BEFORE_ALL_SKIP_REGIONS: // -----// IR Dump Before{{.*}}CSE (cse) //----- //
+// BEFORE_ALL_SKIP_REGIONS-NEXT: func @foo() {...}
+// BEFORE_ALL_SKIP_REGIONS: // -----// IR Dump Before{{.*}}Canonicalizer (canonicalize) //----- //
+// BEFORE_ALL_SKIP_REGIONS-NEXT: func @foo() {...}
+
+// BEFORE_ALL_SKIP_REGIONS: // -----// IR Dump Before{{.*}}CSE (cse) //----- //
+// BEFORE_ALL_SKIP_REGIONS-NEXT: func @bar() {...}
+// BEFORE_ALL_SKIP_REGIONS: // -----// IR Dump Before{{.*}}Canonicalizer (canonicalize) //----- //
+// BEFORE_ALL_SKIP_REGIONS-NEXT: func @bar() {...}
+
// AFTER_ALL_CHANGE: // -----// IR Dump After{{.*}}CSE (cse) //----- //
// AFTER_ALL_CHANGE-NEXT: func @foo()
// AFTER_ALL_CHANGE-NOT: // -----// IR Dump After{{.*}}CSE (cse) //----- //
|
|
@llvm/pr-subscribers-mlir Author: Jakub Kuderski (kuhar) ChangesThe new flag, Example: // -----// IR Dump Before CSE (cse) //----- //
func.func @<!-- -->bar(%arg0: f32, %arg1: f32) -> f32 {...}
// -----// IR Dump Before Canonicalizer (canonicalize) //----- //
func.func @<!-- -->bar(%arg0: f32, %arg1: f32) -> f32 {...}The main use-case is to be triage compilation issues (crashes, slowness) on very deep pass pipelines and with very large IR files, where printing IR is prohibitively slow otherwise. Unlike the pass statistics mechanism, the new flag can be used as an 'interactive' progress indicator. Full diff: https://github.com/llvm/llvm-project/pull/77726.diff 3 Files Affected:
diff --git a/mlir/docs/PassManagement.md b/mlir/docs/PassManagement.md
index 95a38207b7f854..a303e350d28b6b 100644
--- a/mlir/docs/PassManagement.md
+++ b/mlir/docs/PassManagement.md
@@ -1302,6 +1302,27 @@ func.func @simple_constant() -> (i32, i32) {
}
```
+* `mlir-print-ir-skip-regions`
+ * Do not print op regions.
+ * This option is useful to when we are interested in the names of the
+ executed passes but do not want to inspect the IR, e.g., when the
+ module is very large and printing is prohibitively slow.
+
+```shell
+$ mlir-opt foo.mlir -mlir-disable-threading -pass-pipeline='func.func(canonicalize,cse)' -mlir-print-ir-after-all -mlir-print-ir-skip-regions
+// -----// IR Dump Before CSE (cse) //----- //
+func.func @bar(%arg0: f32, %arg1: f32) -> f32 {...}
+
+// -----// IR Dump Before Canonicalizer (canonicalize) //----- //
+func.func @bar(%arg0: f32, %arg1: f32) -> f32 {...}
+
+// -----// IR Dump Before CSE (cse) //----- //
+func.func @simple_constant() -> (i32, i32) {...}
+
+// -----// IR Dump Before Canonicalizer (canonicalize) //----- //
+func.func @simple_constant() -> (i32, i32) {...}
+```
+
## Crash and Failure Reproduction
The [pass manager](#pass-manager) in MLIR contains a builtin mechanism to
diff --git a/mlir/lib/Pass/PassManagerOptions.cpp b/mlir/lib/Pass/PassManagerOptions.cpp
index ffc53b7e3ed023..da7f27fb75a29c 100644
--- a/mlir/lib/Pass/PassManagerOptions.cpp
+++ b/mlir/lib/Pass/PassManagerOptions.cpp
@@ -58,6 +58,12 @@ struct PassManagerOptions {
llvm::cl::desc("When printing IR for print-ir-[before|after]{-all} "
"always print the top-level operation"),
llvm::cl::init(false)};
+ llvm::cl::opt<bool> printSkipRegions{
+ "mlir-print-ir-skip-regions",
+ llvm::cl::desc(
+ "When printing IR for print-ir-[beforef|after]{-all} do not output "
+ "regions. This is useful to indicate pass pipeline progress."),
+ llvm::cl::init(false)};
/// Add an IR printing instrumentation if enabled by any 'print-ir' flags.
void addPrinterInstrumentation(PassManager &pm);
@@ -119,10 +125,14 @@ void PassManagerOptions::addPrinterInstrumentation(PassManager &pm) {
if (!shouldPrintBeforePass && !shouldPrintAfterPass)
return;
+ OpPrintingFlags opPrintingFlags;
+ if (printSkipRegions)
+ opPrintingFlags.skipRegions();
+
// Otherwise, add the IR printing instrumentation.
pm.enableIRPrinting(shouldPrintBeforePass, shouldPrintAfterPass,
printModuleScope, printAfterChange, printAfterFailure,
- llvm::errs());
+ llvm::errs(), opPrintingFlags);
}
void mlir::registerPassManagerCLOptions() {
diff --git a/mlir/test/Pass/ir-printing.mlir b/mlir/test/Pass/ir-printing.mlir
index 048b721ba6d530..cc2f277d76ee52 100644
--- a/mlir/test/Pass/ir-printing.mlir
+++ b/mlir/test/Pass/ir-printing.mlir
@@ -3,6 +3,7 @@
// RUN: mlir-opt %s -mlir-disable-threading=true -pass-pipeline='builtin.module(func.func(cse,canonicalize))' -mlir-print-ir-after=cse -o /dev/null 2>&1 | FileCheck -check-prefix=AFTER %s
// RUN: mlir-opt %s -mlir-disable-threading=true -pass-pipeline='builtin.module(func.func(cse,canonicalize))' -mlir-print-ir-after-all -o /dev/null 2>&1 | FileCheck -check-prefix=AFTER_ALL %s
// RUN: mlir-opt %s -mlir-disable-threading=true -pass-pipeline='builtin.module(func.func(cse,canonicalize))' -mlir-print-ir-before=cse -mlir-print-ir-module-scope -o /dev/null 2>&1 | FileCheck -check-prefix=BEFORE_MODULE %s
+// RUN: mlir-opt %s -mlir-disable-threading=true -pass-pipeline='builtin.module(func.func(cse,canonicalize))' -mlir-print-ir-before-all -mlir-print-ir-skip-regions -o /dev/null 2>&1 | FileCheck -check-prefix=BEFORE_ALL_SKIP_REGIONS %s
// RUN: mlir-opt %s -mlir-disable-threading=true -pass-pipeline='builtin.module(func.func(cse,cse))' -mlir-print-ir-after-all -mlir-print-ir-after-change -o /dev/null 2>&1 | FileCheck -check-prefix=AFTER_ALL_CHANGE %s
// RUN: not mlir-opt %s -mlir-disable-threading=true -pass-pipeline='builtin.module(func.func(cse,test-pass-failure))' -mlir-print-ir-after-failure -o /dev/null 2>&1 | FileCheck -check-prefix=AFTER_FAILURE %s
@@ -56,6 +57,16 @@ func.func @bar() {
// BEFORE_MODULE: func @foo()
// BEFORE_MODULE: func @bar()
+// BEFORE_ALL_SKIP_REGIONS: // -----// IR Dump Before{{.*}}CSE (cse) //----- //
+// BEFORE_ALL_SKIP_REGIONS-NEXT: func @foo() {...}
+// BEFORE_ALL_SKIP_REGIONS: // -----// IR Dump Before{{.*}}Canonicalizer (canonicalize) //----- //
+// BEFORE_ALL_SKIP_REGIONS-NEXT: func @foo() {...}
+
+// BEFORE_ALL_SKIP_REGIONS: // -----// IR Dump Before{{.*}}CSE (cse) //----- //
+// BEFORE_ALL_SKIP_REGIONS-NEXT: func @bar() {...}
+// BEFORE_ALL_SKIP_REGIONS: // -----// IR Dump Before{{.*}}Canonicalizer (canonicalize) //----- //
+// BEFORE_ALL_SKIP_REGIONS-NEXT: func @bar() {...}
+
// AFTER_ALL_CHANGE: // -----// IR Dump After{{.*}}CSE (cse) //----- //
// AFTER_ALL_CHANGE-NEXT: func @foo()
// AFTER_ALL_CHANGE-NOT: // -----// IR Dump After{{.*}}CSE (cse) //----- //
|
Groverkss
reviewed
Jan 11, 2024
Groverkss
approved these changes
Jan 11, 2024
joker-eph
requested changes
Jan 11, 2024
kuhar
force-pushed
the
pm-skip-regions
branch
from
6a038ce to
9560cd2
Compare
kuhar
changed the title
joker-eph
approved these changes
Jan 11, 2024
justinfargnoli
pushed a commit
to justinfargnoli/llvm-project
that referenced
this pull request
Jan 28, 2024
The new flag, `--mlir-print-skip-regions`, sets the op printing option that disables region printing. This results in the usual `--mlir-print-ir-*` debug options printing only the names of the executed passes and the signatures of the ops. Example: ```mlir // -----// IR Dump Before CSE (cse) //----- // func.func @bar(%arg0: f32, %arg1: f32) -> f32 {...} // -----// IR Dump Before Canonicalizer (canonicalize) //----- // func.func @bar(%arg0: f32, %arg1: f32) -> f32 {...} ``` The main use-case is to be triage compilation issues (crashes, slowness) on very deep pass pipelines and with very large IR files, where printing IR is prohibitively slow otherwise.
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.
The new flag,
--mlir-print-skip-regions, sets the op printing option that disables region printing. This results in the usual--mlir-print-ir-*debug options printing only the names of the executed passes and the signatures of the ops.Example:
The main use-case is to be triage compilation issues (crashes, slowness) on very deep pass pipelines and with very large IR files, where printing IR is prohibitively slow otherwise.