Skip to content

PoC of lowering compilation time using Python threading - #3

Open
p-wysocki wants to merge 11 commits into
alexsu52:developfrom
p-wysocki:debug_async_compilation
Open

PoC of lowering compilation time using Python threading#3
p-wysocki wants to merge 11 commits into
alexsu52:developfrom
p-wysocki:debug_async_compilation

Conversation

@p-wysocki

Copy link
Copy Markdown

Changes

Related tickets

119274

current_group.operations, current_group.quantizers, quantized_model, quantized_model_graph
)

modified_models.append(modified_model)

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

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

A general comment is that the proposed solution is not memory optimal. What I mean is that this solution requires storing as many copies of the model in memory as there are groups in groups_to_rank. For some models, the number of groups for ranking is in the hundreds. Probably, algorithm can be crashed by memory for huge models.

@alexsu52
alexsu52 self-requested a review September 21, 2023 05:59
alexsu52 pushed a commit that referenced this pull request Apr 8, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants