mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-05-14 10:09:41 +00:00
3772cf9dd4
* [CK_BUILDER] Integrate GPU reference as ConvAlgorithm
Add GPU reference as a ConvAlgorithm specialization, enabling:
- Unified Builder API for reference and optimized kernels
- Future ckProfiler integration for validation
- First step toward numerical validation in Builder tests
Changes:
- Add ConvAlgorithmSpecialization::REFERENCE enum
- Add ConvAlgorithm_Reference struct
- Add IsReferenceAlgorithm concept
- Create 3 reference factories (Forward, BwdData, BwdWeight)
- Wire into conv_dispatcher
- Add proof-of-concept test (passing)
Test result: Can instantiate reference through Builder API
* Add GPU reference execution tests
- Reference kernel executes through Builder (459ms)
- Both reference and optimized can instantiate
- Tests passing
Next: Implement utilities for comparison
* Optimized Builder kernel execution works
- MakeArgument pattern implemented
- Builder-generated kernel executes successfully
- Tests passing (451ms execution)
Next: Add comparison
* VALIDATION COMPLETE: Builder == Reference
Builder-generated kernel output matches GPU reference!
Test: Validate_Optimized_vs_Reference_Forward_2D_FP16
Result: PASS ✓
This proves CK Builder generates correct code!
* Update to new Builder API
All tests passing
* Rename test file for clarity
test_builder_kernel_execution -> test_builder_kernel_validation
* Add all 3 directions support
- Forward, Backward Data, Backward Weight
- All reference factories working
- Dispatcher wired for all directions
- 9 tests passing
Tests:
- test_reference_execution: 3 tests (all directions)
- test_optimized_execution: 3 tests (all directions)
- test_builder_kernel_validation: 3 tests (fwd validated, bwd placeholders)
* Add backward direction support
- Backward data and weight dispatcher wiring
- Fix factories for new API
- All 3 directions tested
- 9 tests passing
* Refactor: Change IsReferenceAlgorithm from concept to consteval function
Address review feedback: Use consteval function in dispatcher instead of
concept, matching the pattern for other algorithms (Tile, XDL, WMMA, DL).
- Remove IsReferenceAlgorithm concept from conv_algorithm_concepts.hpp
- Add IsReferenceAlgorithm() consteval function to conv_dispatcher.hpp
- Update dispatcher to use function call: IsReferenceAlgorithm<T>()
- Remove redundant algorithm checks from reference factory requires clauses
All tests passing (9/9).
* Move Tile algorithm check outside direction block to support all directions
* Implement MakeInvokerPointer interface and add random input validation
- Implement full Argument/Invoker structs for old CK interface (not just nullptr)
- Refactor with reference_common.hpp to reduce code duplication
- Add random input validation tests: Builder vs direct GPU reference (all directions)
- Fix layout: GNHWC -> NHWGC to match reference kernel expectations
- All 12 tests pass with IDENTICAL results on random input
* Move ConvAlgorithm_Reference to test/impl/conv_algorithm_types.hpp
Keep types.hpp for data types only (enums), move algorithm descriptors
to conv_algorithm_types.hpp as suggested by review.
* Add static_assert to ensure reference factories only accept PassThrough operations
Reference implementation doesn't support fused elementwise operations.
Add compile-time validation to fail early with clear error message if
non-PassThrough operations are specified on input, weight, or output.
* Add InstanceTraits support for reference kernels
- Store SIGNATURE/ALGORITHM/VERSION in Instance for reflection
- Create shared ReferenceCommonTraits base for common properties
- Add 3 direction-specific InstanceTraits specializations in one file
- Include data type and layouts in instance_string output
* Remove optimized kernel validation tests from reference-only branch
* Use existing layout helper and organize reference tests
Use LayoutToCK from conv_tensor_layout.hpp and move reference InstanceTraits
test to validation folder.
* Merge develop branch
Fix DataType switch for new mixed precision types.
* Fix comment spacing for CI
* Convert IsReferenceAlgorithm from function to concept
* Add reference tests to CI smoke tests
* Consolidate 3 reference factories into single unified factory
---------
Co-authored-by: Ville Pietilä <188998872+vpietila-amd@users.noreply.github.com>
[ROCm/composable_kernel commit: a0acc83a72]
Builder
This directory contains the experimental builder feature for composable_kernel.
- Status: In development (October - December 2025)
Overview
The builder provides a high-level, semantically-clear interface for constructing composable kernel operations, with an initial focus on convolution kernels for MIOpen. It leverages modern C++20 features (such as POD structs as non-type template parameters, concepts, and designated initializers) to simplify kernel instantiation and improve developer experience.
This project is a prototype for a more general builder pattern for all of composable_kernel (CK) and CKTile, but is currently limited to formalizing the interface between MIOpen and CK.
Design descriptions
Directory Structure
include/ck_tile/builder/Core builder headers and public API.include/ck_tile/builder/reflectReflection mechanism.include/ck_tile/builder/factoryCompile-time dispatch from builder descriptors to our exisitng specialized convolution kernel implementations.test/Unit tests and example usage of the builder pattern.CMakeLists.txtCMake configuration for building the experimental builder and its tests.
CMake Configuration
To enable the experimental builder, configure your build with:
cmake \
-D CMAKE_PREFIX_PATH=/opt/rocm \
-D CMAKE_CXX_COMPILER=/opt/rocm/bin/hipcc \
-D CMAKE_BUILD_TYPE=Release \
-D GPU_TARGETS="gfx942" \
-D CK_EXPERIMENTAL_BUILDER=ON \
-D CMAKE_CXX_STANDARD=20 \
-G Ninja \
..
Building and Testing
The builder test suite is organized into two main categories:
Smoke Tests (Fast Unit Tests)
Quick unit tests that verify the builder's internal logic without compiling GPU kernels. These complete in under 1 second total and are suitable for frequent execution during development.
ninja smoke-builder
Regression Tests (Integration Tests)
Integration tests that compile actual GPU kernels to verify that the builder generates valid, compilable code. These are more expensive than smoke tests (can take minutes to compile) but cover more fuctionality. )
ninja regression-builder
Running All Tests
To build and run the complete test suite:
ninja check-builder
Building Individual Tests
To build and run a specific test:
ninja test_ckb_conv_builder && bin/test_ckb_conv_builder
Test Organization
- Smoke tests: Fast feedback during active development
- Regression tests: Thorough validation before submitting changes
- Factory tests: Expensive tests that build all MIOpen kernels (included in regression tests)
When adding new tests, please follow the convention where the CMake build target starts with a prefix test_ckb. This allows filtering of CK Builder tests from the full CK repository test suite.