Skip to content

[MLIR][Affine] Switch default option for affine data copy generate pass #129445

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Mar 3, 2025
Merged

[MLIR][Affine] Switch default option for affine data copy generate pass #129445

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion mlir/include/mlir/Dialect/Affine/Passes.td
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,7 +27,7 @@ def AffineDataCopyGeneration : Pass<"affine-data-copy-generate", "func::FuncOp">
/*default=*/"1",
"Fast memory space identifier for copy generation (default: 1)">,
Option<"generateDma", "generate-dma", "bool",
/*default=*/"true", "Generate DMA instead of point-wise copy">,
/*default=*/"false", "Generate DMA instead of point-wise copy">,
Option<"minDmaTransferSize", "min-dma-transfer", "int",
/*default=*/"1024",
"Minimum DMA transfer size supported by the target in bytes">,
Expand Down
2 changes: 1 addition & 1 deletion mlir/lib/Dialect/Affine/Transforms/AffineDataCopyGeneration.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -86,7 +86,7 @@ struct AffineDataCopyGeneration

/// Generates copies for memref's living in 'slowMemorySpace' into newly created
/// buffers in 'fastMemorySpace', and replaces memory operations to the former
/// by the latter. Only load op's handled for now.
/// by the latter.
std::unique_ptr<OperationPass<func::FuncOp>>
mlir::affine::createAffineDataCopyGenerationPass(
unsigned slowMemorySpace, unsigned fastMemorySpace, unsigned tagMemorySpace,
Expand Down
4 changes: 2 additions & 2 deletions mlir/test/Dialect/Affine/affine-data-copy.mlir
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
// RUN: mlir-opt %s -split-input-file -affine-data-copy-generate="generate-dma=false fast-mem-space=0 skip-non-unit-stride-loops" | FileCheck %s
// RUN: mlir-opt %s -split-input-file -affine-data-copy-generate="fast-mem-space=0 skip-non-unit-stride-loops" | FileCheck %s
// Small buffer size to trigger fine copies.
// RUN: mlir-opt %s -split-input-file -affine-data-copy-generate="generate-dma=false fast-mem-space=0 fast-mem-capacity=1" | FileCheck --check-prefix=CHECK-SMALL %s
// RUN: mlir-opt %s -split-input-file -affine-data-copy-generate="fast-mem-space=0 fast-mem-capacity=1" | FileCheck --check-prefix=CHECK-SMALL %s

// Test affine data copy with a memref filter. We use a test pass that invokes
// affine data copy utility on the input loop nest.
Expand Down
4 changes: 2 additions & 2 deletions mlir/test/Dialect/Affine/dma-generate.mlir
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,5 @@
// RUN: mlir-opt -allow-unregistered-dialect %s -split-input-file -affine-data-copy-generate="generate-dma fast-mem-space=2 skip-non-unit-stride-loops" -verify-diagnostics | FileCheck %s
// RUN: mlir-opt -allow-unregistered-dialect %s -split-input-file -affine-data-copy-generate="generate-dma fast-mem-capacity=16 fast-mem-space=2" | FileCheck %s --check-prefix FAST-MEM-16KB
// RUN: mlir-opt -allow-unregistered-dialect %s -split-input-file -affine-data-copy-generate="generate-dma=1 fast-mem-space=2 skip-non-unit-stride-loops" -verify-diagnostics | FileCheck %s
// RUN: mlir-opt -allow-unregistered-dialect %s -split-input-file -affine-data-copy-generate="generate-dma=1 fast-mem-capacity=16 fast-mem-space=2" | FileCheck %s --check-prefix FAST-MEM-16KB

// We run most test cases with -copy-skip-non-unit-stride-loops to allow testing
// DMA generation at inner levels easily - since the DMA generation would
Expand Down