[LLVMGPU][DT] Add MaterializeDeviceEncodingPass to LLVMGPU passes behind flag #19849
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Max191
requested review from
MaheshRavishankar,
qedawkins,
kuhar and
Groverkss
as code owners
kuhar
reviewed
Jan 29, 2025
…ind flag Signed-off-by: Max Dawkins <[email protected]>
Max191
force-pushed
the
llvmgpu-materialize-device-encodings
branch
from
86792d5 to
965eeb6
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR adds the
MaterializeDeviceEncodingPassto LLVMGPU/Passes.cpp behind a new flagiree-llvmgpu-experimental-data-tiling. The flag's default value is false, because the codegen for data tiling ops on GPU is not yet working and performant for all cases. Some of the work is in flight, but it will likely take some time before data tiling codegen is ready to be flipped on by default. For now, the flag allows developers to enable the late materialization codegen path on LLVMGPU.To effectively use the late materialization path, some additional non-default flags need to be set:
--iree-opt-data-tiling=false(to turn off the early materialization data tiling path)--iree-dispatch-creation-experimental-data-tiling=true(to turn on the late materialization data tiling path)--iree-dispatch-creation-pad-factor=128(the current default is based on CPU materialization)