forked from flexflow/flexflow-train
-
Notifications
You must be signed in to change notification settings - Fork 0
Repo refactor kernel help test #4
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Open
lambda7xx
wants to merge
338
commits into
repo-refactor-lambda-kernel-test-issue948
Choose a base branch
from
repo-refactor-kernel-help-test
base: repo-refactor-lambda-kernel-test-issue948
Could not load branches
Branch not found: {{ refName }}
Loading
Could not load tags
Nothing to show
Loading
Are you sure you want to change the base?
Some commits from the old base branch may be removed from the timeline,
and old review comments may become outdated.
Open
Repo refactor kernel help test #4
lambda7xx
wants to merge
338
commits into
repo-refactor-lambda-kernel-test-issue948
from
repo-refactor-kernel-help-test
Conversation
This file contains hidden or 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
…lambda-utils-testing
* add unit tests for machine mapping and dp algorithm * add unit test for unity algorithm * Merge remote-tracking branch 'lockshaw/repo-refactor' into repo-refactor * fix compile errors from filter and support_interator_tag * minor fixes for compiler * Merge remote-tracking branch 'lockshaw/repo-refactor' into repo-refactor * Merge branch 'repo-refactor' into repo-refactor * Merge branch 'repo-refactor' of github.com:wmdi/FlexFlow into repo-refactor * clean up generator codes and minor fix * format * format * serial parallel composition * remove commited out codes * view MultiDiGraph as labelled * make machine mapping immutable * minor fix & format * move general codes into proper places * format * Merge remote-tracking branch 'upstream/repo-refactor' into repo-refactor * minor fix & format * minor fix Co-Authored-By: Colin Unger <[email protected]>
…transpose-OP Transpose OP
…topK-OP topK OP
* start to do the reshape operator * add init_task,forward_task, forward_task for reshape * reshape version0.1 done * update the reshape and leave the init_task and measure_operator_task * refine the reshape pr * remove the replicate * fix error in reshape * fix the type error * add init_kernel for reshape * modify the reshpae by fixing the typo error * refine the reshape * reshape * format the code * fix the error * format the code * format the code --------- Co-authored-by: Colin Unger <[email protected]> Co-authored-by: Reyna Abhyankar <[email protected]>
* add reverse op version0.1 * leave num_out_blks and others to update * first version of reverse * remove the init * fix the error --------- Co-authored-by: Colin Unger <[email protected]> Co-authored-by: Reyna Abhyankar <[email protected]>
…Repartition-OP Repartition OP
* implement the init_kernel * fix the forward_kernel and backward_kernel definition * implement the init_task_impl * implement the SimTaskBinding init_binding for measure_operator_cost * implement the register_task for init * implement the forward and backward * implement the register_task for forward and backward * implement the measure_cost_operator_cost * remove the old code * pool2d version 0.1 * fix the typo * refine the pool2d --------- Co-authored-by: Reyna Abhyankar <[email protected]>
* add init_kernel for reduction * implement the init * add forward API * add backward API * implement measure_operator_cost and Reduce version 0.1 * format the code * update the redudce * update the redudce * use exceptions in reduces.cc * fix the typo * fix the reduce.cc * format the code * add reduction size --------- Co-authored-by: Colin Unger <[email protected]> Co-authored-by: Reyna Abhyankar <[email protected]>
* start to do the Reduction * Reduction version0.1 * update the reduction * use the exceptions.h in reduction.cc * fix the input bind * fix the reduction * fix the reduction * change attrs.reduction_degree --------- Co-authored-by: Colin Unger <[email protected]> Co-authored-by: Reyna Abhyankar <[email protected]>
Co-authored-by: Zhihao Jia <[email protected]>
* batch matmul initial commit * fix FF_VISITABLE_STRUCT_NO_EQ * Merge branch 'repo-refactor' into batch_matmul * finish draft 1 batch_matmul * Merge branch 'batch_matmul' of github.com:KateUnger/FlexFlow into batch_matmul * add output and weights * format * Merge branch 'repo-refactor' of https://github.com/flexflow/FlexFlow into batch_matmul * fix DeviceSpecific * change * change * change * change * change * change * change * change * change * fix asserts * format * Merge branch 'repo-refactor' of https://github.com/flexflow/FlexFlow into batch_matmul * batch_matmul * format * Merge branch 'repo-refactor' of https://github.com/flexflow/FlexFlow into batch_matmul * add cuda * fix attention kuda * format * draft * format * reyna fixes * format * format repo-refactor * Merge branch 'form' of github.com:KateUnger/FlexFlow into batch_matmul * Merge branch 'repo-refactor' of https://github.com/flexflow/FlexFlow into batch_matmul * delete init and split register_task * finish bmm and att. * Merge branch 'repo-refactor' into batch_matmul * Merge branch 'repo-refactor' into batch_matmul * Align hip kernel * Merge branch 'repo-refactor' into batch_matmul * Merge branch 'repo-refactor' into batch_matmul * Fix attention kernels * Replace with unique ptr * Format * Merge branch 'repo-refactor' into batch_matmul * Merge branch 'repo-refactor' into batch_matmul Co-Authored-By: Colin Unger <[email protected]> Co-Authored-By: Reyna Abhyankar <[email protected]>
* Fix hip kernels for merged ops * Format * Fix combine, concat
Substitutions build
* implement the init_kernel in linear_kernels.cu * remove unuseful code * implement the forward * implement the backward_task_impl * implement the register_task * measure operator cost * leave the init_kernel float * one_ptr and remove the old code * fix the error * Update lib/kernels/src/cuda/linear_kernels.cu * fix the linear --------- Co-authored-by: Reyna Abhyankar <[email protected]>
* init draft and leave some comment * implement the forward and backward * layer norm version 0.1 * layer norm draft * fix the layer norm and leave beta, gamma * Finish layer norm * refine the layernorm * Apply suggestions from code review * fix the typo and format --------- Co-authored-by: Reyna Abhyankar <[email protected]>
* Element Unary * Fix * Fix attrs * Format * Remove comment * Merge branch 'repo-refactor' into element-unary-op * Fix endif * Move arguments and remove ElementScalarUnary * Remove ElementScalarUnary * Merge branch 'repo-refactor' into element-unary-op * Scalar Unary with inheritance * Format * Merge branch 'repo-refactor' into element-unary-op * Fix * add optional * Add op type * Add element scalar unary * Fix format and build * Update lib/op-attrs/include/op-attrs/ops/element_unary.h * Update lib/op-attrs/include/op-attrs/ops/element_unary.h * Update lib/op-attrs/include/op-attrs/ops/element_unary.h
…exflow#1321) * Remove unnecessary dependencies and allow using external installs * Cleanup before PR and format * Add FindNCCL from torch, remove unnecessary cmake files * Cleanup nccl cmake * Fix remaining optional bugs in kernels * Try nix-based CI * Fix nix-develop version * Fix flake devshell name * Move ci build code to script * Satisfy shellcheck * Try removing ccache * Attempt to skip ccache install * Move to using in-flake caches * Format
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.
Description of changes:
Related Issues:
Linked Issues:
Issues closed by this PR:
Before merging: