[SERVER-83304] Investigate MSG_ZEROCOPY for large messages Created: 15/Nov/23  Updated: 05/Feb/24

Status: In Progress
Project: Core Server
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Improvement Priority: Major - P3
Reporter: Erin McNulty Assignee: Ryan Berryhill
Resolution: Unresolved Votes: 0
Labels: perf-8.0, perf-tiger, perf-tiger-handoff
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Related
related to SERVER-85752 Create microbenchmark for ASIO transp... Open
Assigned Teams:
Service Arch
Sprint: Service Arch 2023-12-25, Service Arch 2024-01-08, Service Arch 2024-01-22, Service Arch 2024-02-05, Service Arch 2024-02-19
Participants:

 Description   

In SERVER-81784, Mathias suggested investigating MSG_ZEROCOPY along with MSG_WAITALL:

Could also investigate using MSG_ZEROCOPY with large messages to keep the Message object alive until the kernel knows it won't need it while allowing the kernel to avoid doing its own copies. They claim it is only beneficial for messages >10KB.

Splitting this into its own ticket because I think it warrants its own investigation (after looking at it for a bit, I think we would have to determine a threshold for message size at which it would be a useful flag to set)


Generated at Thu Feb 08 06:51:48 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.