Yang Chen 4d0ae7c9da [inductor] support _scaled_dot_product_flash_attention fallback (#110085) ... Summary: This PR supports _scaled_dot_product_flash_attention fallback kernel. Note that in the abi_compatible mode, we retrieve outputs by passing output argument pointers rather than relying on std::get. It also fixes an issue related to dynamic shapes, where we wrongfully query undefined dynamic symbols. Test Plan: ci Reviewed By: frank-wei Differential Revision: D49620191 Pull Request resolved: https://github.com/pytorch/pytorch/pull/110085 Approved by: https://github.com/desertfire		2023-09-27 00:09:56 +00:00
..
aoti_runtime	[aotinductor] Rename aot_runtime to aoti_runtime (#110007)	2023-09-26 00:46:54 +00:00
cuda	[inductor] enable mypy checking in torch/_inductor/codegen/cpp.py (#109729)	2023-09-25 22:53:05 +00:00
__init__.py
common.py	[inductor] enable mypy checking in torch/_inductor/codegen/cpp.py (#109729)	2023-09-25 22:53:05 +00:00
cpp_prefix.h	inductor: support masked load for cpu path (#107670)	2023-08-25 21:11:09 +00:00
cpp.py	[inductor] enable mypy checking in torch/_inductor/codegen/cpp.py (#109729)	2023-09-25 22:53:05 +00:00
triton_foreach.py	[Inductor] Generalize inductor triton backend device agnostic (#109486)	2023-09-24 07:49:20 +00:00
triton_utils.py	Basic fp8 support in Inductor (#109168)	2023-09-23 04:41:41 +00:00
triton.py	[inductor] Fix triton compiler error in multilayer any (#109325)	2023-09-26 12:29:29 +00:00
wrapper.py	[inductor] support _scaled_dot_product_flash_attention fallback (#110085)	2023-09-27 00:09:56 +00:00