llvm-project/mlir/test
MaheshRavishankar 5ca20851e4 [mlir][Linalg] Improve the logic to perform tile and fuse with better dependence tracking.
This change does two main things
1) An operation might have multiple dependences to the same
   producer. Not tracking them correctly can result in incorrect code
   generation with fusion. To rectify this the dependence tracking
   needs to also have the operand number in the consumer.
2) Improve the logic used to find the fused loops making it easier to
   follow. The only constraint for fusion is that linalg ops (on
   buffers) have update semantics for the result. Fusion should be
   such that only one iteration of the fused loop (which is also a
   tiled loop) must touch only one (disjoint) tile of the output. This
   could be relaxed by allowing for recomputation that is the default
   when oeprands are tensors, or can be made legal with promotion of
   the fused view (in future).

Differential Revision: https://reviews.llvm.org/D90579
2020-11-12 00:25:24 -08:00
..
Analysis [mlir] Fix post-dominance between blocks of different regions. 2020-11-11 11:20:53 -08:00
Bindings/Python Disable implicit nesting on parsing textual pass pipeline 2020-11-11 19:21:51 +00:00
CAPI [mlir][CAPI] Proposal: Always building a libMLIRPublicAPI.so (re-apply). 2020-11-08 16:57:51 -08:00
Conversion [MLIR][SPIRVToLLVM] Enhanced conversion for execution mode 2020-11-10 18:33:54 +03:00
Dialect [mlir][Linalg] Improve the logic to perform tile and fuse with better dependence tracking. 2020-11-12 00:25:24 -08:00
EDSC Remove global dialect registration 2020-10-24 00:35:55 +00:00
Examples [MLIR] Change FuncOp assembly syntax to print visibility inline instead of in attrib dict. 2020-11-09 11:08:08 -08:00
IR [mlir] Use assemblyFormat in AllocLikeOp. 2020-11-11 10:27:20 +01:00
Pass Disable implicit nesting on parsing textual pass pipeline 2020-11-11 19:21:51 +00:00
SDBM Remove global dialect registration 2020-10-24 00:35:55 +00:00
Target Revert "Use uint64_t for branch weights instead of uint32_t" 2020-10-31 00:25:32 -07:00
Transforms [mlir] Rework DialectConversion inlineRegionBefore 2020-11-11 10:42:33 +01:00
Unit Add build files and update README. 2019-03-30 11:23:22 -07:00
lib [mlir][Linalg] Improve the logic to perform tile and fuse with better dependence tracking. 2020-11-12 00:25:24 -08:00
mlir-cpu-runner [mlir][std] Add ExpandOps pass. 2020-11-09 21:58:28 +01:00
mlir-cuda-runner [MLIR] Adding gpu.host_register op and lower it to a runtime call. 2020-08-10 22:46:17 +02:00
mlir-linalg-ods-gen [mlir][Linalg] Side effects interface for Linalg ops 2020-11-05 09:00:28 +00:00
mlir-opt Separate the Registration from Loading dialects in the Context 2020-08-19 01:19:03 +00:00
mlir-reduce [MLIR] Change FuncOp assembly syntax to print visibility inline instead of in attrib dict. 2020-11-09 11:08:08 -08:00
mlir-rocm-runner [AMDGPU] Fix ROCm unit test memref initialization 2020-09-17 09:48:05 -07:00
mlir-spirv-cpu-runner [mlir][gpu] NFC: Change gpu.launch_func ops to custom format. 2020-10-29 21:21:30 +01:00
mlir-tblgen [mlir] Allow omitting spaces in assemblyFormat with a `` literal. 2020-11-11 09:34:43 +01:00
mlir-vulkan-runner [mlir][gpu] NFC: switch occurrences of gpu.launch_func to custom format. 2020-10-22 17:27:19 +02:00
APITest.h Mass update the MLIR license header to mention "Part of the LLVM project" 2020-01-26 03:58:30 +00:00
CMakeLists.txt Add basic Python bindings for the PassManager and bind libTransforms 2020-11-10 19:55:21 +00:00
lit.cfg.py Add a basic C API for the MLIR PassManager as well as a basic TableGen backend for creating passes 2020-11-04 06:36:31 +00:00
lit.site.cfg.py.in [MLIR][mlir-spirv-cpu-runner] A SPIR-V cpu runner prototype 2020-10-26 09:09:29 -04:00