Skip to content

[REFACTOR] Add sequential kernel data stacking with GPU memory optimization strategies#57

Merged
Leguark merged 1 commit intomainfrom
memory
Apr 1, 2026
Merged

[REFACTOR] Add sequential kernel data stacking with GPU memory optimization strategies#57
Leguark merged 1 commit intomainfrom
memory

Conversation

@Leguark
Copy link
Copy Markdown
Member

@Leguark Leguark commented Mar 26, 2026

Changes

  • Introduced _build_stacked_kernel_data_parallel and refactored _build_stacked_kernel_data to enhance parallelization
  • Added multiple selective stacking strategies (_stack_sub_struct_split, _stack_sub_struct_pinned, _stack_sub_struct_gpu_first) for improved flexibility
  • Enhanced GPU memory handling with explicit profiling zones for optimized tensor operations
  • Improved fail-safe concatenation logic to handle various backend configurations
  • Replaced parallel kernel data building with sequential approach using _stack_sub_struct_split strategy
  • Added pinned memory optimization for high-speed CPU-to-GPU transfers in _stack_sub_struct_pinned

Copy link
Copy Markdown
Member Author

Leguark commented Apr 1, 2026

Merge activity

  • Apr 1, 5:23 PM UTC: A user started a stack merge that includes this pull request via Graphite.
  • Apr 1, 5:24 PM UTC: Graphite rebased this pull request as part of a merge.
  • Apr 1, 5:29 PM UTC: @Leguark merged this pull request with Graphite.

@Leguark Leguark changed the base branch from wqef_improvements to graphite-base/57 April 1, 2026 17:23
@Leguark Leguark changed the base branch from graphite-base/57 to main April 1, 2026 17:24
…tegies

- Introduced `_build_stacked_kernel_data_parallel` and refactored `_build_stacked_kernel_data` to enhance parallelization.
- Added multiple selective stacking strategies (`_stack_sub_struct_split`, `_stack_sub_struct_pinned`, `_stack_sub_struct_gpu_first`) for improved flexibility.
- Enhanced GPU memory handling with explicit profiling zones for optimized tensor operations.
- Improved fail-safe concatenation logic to handle various backend configurations.
@Leguark Leguark merged commit ea51595 into main Apr 1, 2026
2 checks passed
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.

1 participant