Landing: 5cc9ef64063c
Project / Subsystem
gcc / builtins
Date
2026-09-21
Author
Matt Turner
Commit
5cc9ef64063c45cac1d67fcdaba9258fbaf21a8c
Source
github
Perf win
Yes
Breaking
No
All attributes
- project
- gcc
- subsystem
- builtins
- patch_id
- —
- commit_hash
- 5cc9ef64063c45cac1d67fcdaba9258fbaf21a8c
- source_type
- github
- headline
- builtins: build fpclassify on the classification optabs [PR66462]
- tldr
- Built-in fpclassify now uses specialized classification optabs for better accuracy and to avoid spurious exceptions.
- author
- Matt Turner
- outcome
- committed
- performance_win
- true
- breaking_change
- false
- series_id
- gcc:builtins: build fpclassify on the classification optabs [pr66462]
- series_parts
- []
- tags
-
- • builtins
- • floating-point
- • optimization
- discussion_id_link
- —
- bugzilla_pr
- —
- date
- 2026-09-21T00:00:00.000Z
The __builtin_fpclassify function now leverages target-specific classification optabs, such as isnan, isinf, and isnormal, when available. This change replaces direct floating-point comparisons, which could lead to incorrect results or trap on signaling NaNs and subnormal numbers on certain architectures like AArch64 and LoongArch. The new implementation uses bit manipulation for improved accuracy.