llvm-project/mlir/test/Conversion/GPUCommon
Christian Sigg 5535696c38 [mlir] Add gpu.allocate, gpu.deallocate ops with LLVM lowering to runtime function calls.
The ops are very similar to the std variants, but support async GPU execution.

gpu.alloc does not currently support an alignment attribute, and the new ops do not have
canonicalizers/folders like their std siblings do.

Reviewed By: herhut

Differential Revision: https://reviews.llvm.org/D91698
2020-11-27 09:40:59 +01:00
..
lower-alloc-to-gpu-runtime-calls.mlir [mlir] Add gpu.allocate, gpu.deallocate ops with LLVM lowering to runtime function calls. 2020-11-27 09:40:59 +01:00
lower-launch-func-to-gpu-runtime-calls.mlir [mlir][gpu] Fix leaked stream and module when lowering gpu.launch_func to runtime calls. 2020-10-29 08:40:51 +01:00
lower-wait-to-gpu-runtime-calls.mlir [mlir][gpu] Add lowering to LLVM for `gpu.wait` and `gpu.wait async`. 2020-10-21 18:20:42 +02:00
memory-attrbution.mlir [mlir] switch the modeling of LLVM types to use the new mechanism 2020-08-04 14:29:25 +02:00