Implementing a new lowering pass

What is a lowering pass?

In MLIR, a lowering pass rewrites operations from one dialect into operations of a lower-level dialect — a step on the journey from abstract IR to executable code. snn-mlir ships one reference lowering, SNNToLinalg, which converts each snn op into standard linalg/arith operations that any MLIR-based CPU backend can consume.

This is the extension point for hardware developers. To target your own accelerator, you write a pass that lowers the snn ops to your representation — whether that's another MLIR dialect, intrinsics, or calls into a hardware runtime. The same network.mlir can then be compiled for a custom FPGA target or an ASIC implementation without touching the frontend or the dialect definition. lib/Conversion/SNNToLinalg/SNNToLinalg.cpp is the reference implementation to copy from.

1. Create the pass files

include/SNN/Conversion/SNNToMyBackend.h
lib/Conversion/SNNToMyBackend/SNNToMyBackend.cpp
lib/Conversion/SNNToMyBackend/CMakeLists.txt

2. Declare your pass in the header

#include "mlir/Pass/Pass.h"
#include <memory>

namespace snn {
  std::unique_ptr<mlir::Pass> createConvertSNNToMyBackendPass();
  void registerConvertSNNToMyBackendPass();
} // namespace snn

3. Implement a rewrite pattern per op

#include "SNN/SNNOps.h"

struct LowerLinear : public OpRewritePattern<snn::LinearOp> {
  using OpRewritePattern::OpRewritePattern;

  LogicalResult matchAndRewrite(snn::LinearOp op,
                                PatternRewriter &rewriter) const override {
    // Replace op with your backend calls
    rewriter.eraseOp(op);
    return success();
  }
};

4. Wire up the pass

struct ConvertSNNToMyBackendPass
    : public PassWrapper<ConvertSNNToMyBackendPass, OperationPass<ModuleOp>> {

  StringRef getArgument() const override { return "convert-snn-to-mybackend"; }

  void runOnOperation() override {
    RewritePatternSet patterns(&getContext());
    patterns.add<LowerLinear, LowerRescale, LowerCubaLIF>(&getContext());

    ConversionTarget target(getContext());
    target.addIllegalDialect<snn::SNNDialect>();
    target.addLegalDialect</* your dialects */>();

    if (failed(applyPartialConversion(getOperation(), target, std::move(patterns))))
      signalPassFailure();
  }
};

5. Register in CMake

Use add_mlir_conversion_library() — see lib/Conversion/SNNToLinalg/CMakeLists.txt as a template.