Skip to content

efficiency: dispatch regions in a loop in process_event_impl - #745

Merged
kris-jusiak merged 1 commit into
boost-ext:masterfrom
infn-ke:kenth/dispatch_loop
Nov 20, 2026
Merged

kris-jusiak merged 1 commit into
boost-ext:masterfrom
infn-ke:kenth/dispatch_loop

Conversation

@infn-ke

@infn-ke infn-ke commented Nov 19, 2026

Copy link
Copy Markdown
Contributor

process_event_impl unrolls one dispatch call per orthogonal region. Every region uses the same dispatch table, so a loop over current_state_ keeps one call site and drops the fold-expression branch.

Clang 19.1.7, -Oz, -mcpu=cortex-a53, one machine with three regions: the translation unit that instantiates it shrank by 992 bytes.

process_event_impl unrolls one dispatch call per orthogonal region.
Every region uses the same dispatch table, so a loop over
current_state_ keeps one call site and drops the fold-expression
branch.

Clang 19.1.7, -Oz, -mcpu=cortex-a53, one machine with three regions:
the translation unit that instantiates it shrank by 992 bytes.
@kris-jusiak
kris-jusiak merged commit d2a7ec1 into boost-ext:master Nov 20, 2026
7 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.

2 participants