Increase chunk sizes to improve performance and parallelism #4080

Open
opened 2023-12-08 15:54:13 +00:00 by itamarst · 0 comments
itamarst commented 2023-12-08 15:54:13 +00:00
Owner
  1. Processing data in larger chunk sizes can reduce overhead of all the little Python function calls for each chunk.
  2. Larger chunk sizes for functions that release GIL add additional opportunities for utilizing parallelism by shifting to a thread pool. Larger chunks of work are more efficient for this scheme.
1. Processing data in larger chunk sizes can reduce overhead of all the little Python function calls for each chunk. 2. Larger chunk sizes for functions that release GIL add additional opportunities for utilizing parallelism by shifting to a thread pool. Larger chunks of work are more efficient for this scheme.
tahoe-lafs added the
unknown
normal
enhancement
n/a
labels 2023-12-08 15:54:13 +00:00
tahoe-lafs added this to the Performance and Benchmarking milestone 2023-12-08 15:54:13 +00:00
Sign in to join this conversation.
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference: tahoe-lafs/trac-2024-07-25#4080
No description provided.