This commit is contained in:
drazi
2025-06-07 18:00:39 +08:00
committed by GitHub
parent 5a287538c2
commit cc23f6d1e9

View File

@@ -18,7 +18,7 @@
- [Ampere GEMM](https://github.com/NVIDIA/cutlass/tree/main/examples/python/CuTeDSL/ampere/tensorop_gemm.py)
- [FlashAttention-2 implementation targeting Ampere and Ada class GPUs (SM80, SM86, SM89)](https://github.com/NVIDIA/cutlass/tree/main/examples/python/CuTeDSL/ampere/flash_attention_v2.py)
- [SmemAllocator to facilitate shared memory allocation and management](https://github.com/NVIDIA/cutlass/tree/main/examples/python/CuTeDSL/ampere/smem_allocator.py)
- [C-structure based customized interface between JIT function and user codes](https://github.com/NVIDIA/cutlass/tree/main/examples/python/CuTeDSL/cute/jit_argument.py)
- [C-structure based customized interface between JIT function and user codes](https://github.com/NVIDIA/cutlass/tree/main/examples/python/CuTeDSL/cute/ffi/jit_argument.py)
* [Educational notebooks for getting started with CuTe DSL](https://github.com/NVIDIA/cutlass/tree/main/examples/python/CuTeDSL/notebooks)
* API updates
- Fixed API mismatch in class ``cute.runtime.Pointer``: change ``element_type`` to ``dtype`` to match ``typing.Pointer``