Skip to content

Conversation

@kfirtoledo
Copy link
Collaborator

This PR adds the main storage offload CUDA module for the filesystem offloading connector.
Included in this PR:

  • Adds storage_offload.cu
  • Implements the high-level CUDA logic that orchestrates block movement between GPU and storage
  • Includes the initialization and cleanup routines for resources used by the module
  • Adds setup.py configuration required to build and load the CUDA extension
  • Adds debug_utils.hpp for debug logging and timing measurement

@kfirtoledo kfirtoledo requested review from orozery and removed request for elevran and vMaroon November 23, 2025 11:55
@kfirtoledo kfirtoledo force-pushed the connector_storage_cuda branch 2 times, most recently from b9db7c8 to d4b8e4c Compare November 28, 2025 08:22
Signed-off-by: Kfir Toledo <kfir.toledo@ibm.com>
@kfirtoledo kfirtoledo force-pushed the connector_storage_cuda branch from d4b8e4c to 8a75202 Compare December 4, 2025 12:09
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