Skip to content
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

Nkpatel/conv op non tile multiple shard widht #15742

Merged
merged 14 commits into from
Dec 5, 2024

Conversation

nkpatel-tt
Copy link
Contributor

@nkpatel-tt nkpatel-tt commented Dec 5, 2024

Ticket

Same PR as
#11435

Problem description

With tile multiple shard width, all cores might not be utilised which can degrade performance.

What's changed

Add non-tile multiple shard width support.

Checklist

  • Post commit CI passes Link passing with clang-tidy
  • Blackhole Post commit (if applicable)
  • Model regression CI testing passes (if applicable)
  • Device performance regression CI testing passes (if applicable)
  • New/Existing tests provide coverage for changes

Signed-off-by: Nilaykumar Patel <[email protected]>

tensor util changes

Signed-off-by: Nilaykumar Patel <[email protected]>

conv2d changes

Signed-off-by: Nilaykumar K Patel <[email protected]>

Normal test cases working with updated conv block config.

Signed-off-by: Nilaykumar K Patel <[email protected]>

kernel changes

Signed-off-by: Nilaykumar K Patel <[email protected]>

Resolve hang

Signed-off-by: Nilaykumar K Patel <[email protected]>

Clean up debug statements.

Signed-off-by: Nilaykumar K Patel <[email protected]>

Clean up debug statements and functions.

Signed-off-by: Nilaykumar K Patel <[email protected]>

Solve WS test hang.

Signed-off-by: Nilaykumar K Patel <[email protected]>

Update input offset calculations based on alignment.

Signed-off-by: Nilaykumar K Patel <[email protected]>

Align input matrix M for 1x1 conv according to new changes.

Signed-off-by: Nilaykumar K Patel <[email protected]>

Fix after rebase

Signed-off-by: Nilaykumar K Patel <[email protected]>

Resolve Comments.

Signed-off-by: Nilaykumar K Patel <[email protected]>

Remove Debug statements and commented code.

Signed-off-by: Nilaykumar K Patel <[email protected]>

Changes with rebase.

Signed-off-by: Nilaykumar K Patel <[email protected]>

Undo some changes after rebase-clenup

Signed-off-by: Nilaykumar K Patel <[email protected]>

Make pipeline work. Clean up needed.

Signed-off-by: Nilaykumar K Patel <[email protected]>

Update input offset calculations based on alignment. Resolves maxpool and model pipeline failures.

Signed-off-by: Nilaykumar K Patel <[email protected]>

Enable variable to support non-tile multiple width.

Signed-off-by: Nilaykumar K Patel <[email protected]>

Divide tiles among cores instead of total height

Signed-off-by: Nilaykumar K Patel <[email protected]>

Modify Condition

Signed-off-by: Nilaykumar K Patel <[email protected]>

Change after latest rebase.

Signed-off-by: Nilaykumar K Patel <[email protected]>

Modify test cases to accomodate small number of cores.

Signed-off-by: Nilaykumar Patel <[email protected]>

Resolve Yolo failure.

Signed-off-by: Nilaykumar Patel <[email protected]>

Add support for mulit-device tensor for weight and bias tensors.

Signed-off-by: Nilaykumar K Patel <[email protected]>

Add comment for prepare weight and bias matrix and modify condition.

Signed-off-by: Nilaykumar Patel <[email protected]>

Remove debug statements.

Signed-off-by: Nilaykumar Patel <[email protected]>

Address review comments

Signed-off-by: Nilaykumar Patel <[email protected]>
Signed-off-by: Nilaykumar Patel <[email protected]>
Signed-off-by: Nilaykumar Patel <[email protected]>
Signed-off-by: Nilaykumar Patel <[email protected]>
Signed-off-by: Nilaykumar Patel <[email protected]>
Signed-off-by: Nilaykumar Patel <[email protected]>
Merge branch 'main' into nkpatel/conv_op_non_tile_multiple_shard_widht
Copy link
Contributor

@shwetankTT shwetankTT left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@blozano-tt
Copy link
Contributor

I’m merging this, because it was already approved, and I see the clang-tidy fix.

Thanks

@blozano-tt blozano-tt merged commit d6c5a99 into main Dec 5, 2024
120 checks passed
@blozano-tt blozano-tt deleted the nkpatel/conv_op_non_tile_multiple_shard_widht branch December 5, 2024 14:51
@nkpatel-tt
Copy link
Contributor Author

I’m merging this, because it was already approved, and I see the clang-tidy fix.

Thanks

Thanks bryan 👍

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants