Landing: 8bce5e413a15

← All landings

Project / Subsystem

gcc / gcc/alpha

Date

2026-09-21

Author

Matt Turner

Commit

8bce5e413a15abb756a1f6d0bc2f6741b6463b3e

Source

github

Perf win

Breaking

View on GitHub →

All attributes

project: gcc
subsystem: gcc/alpha
patch_id: —
commit_hash: 8bce5e413a15abb756a1f6d0bc2f6741b6463b3e
source_type: github
headline: alpha: Implement FP classification built-in optimizations.
tldr: Fixes FP classification on alpha to avoid crashes with subnormal numbers.
author: Matt Turner
outcome: committed
performance_win: false
breaking_change: false
series_id: gcc:alpha: implement the fp classification optabs
series_parts: []
tags: • alpha
• optimization
• bugfix
• floating-point
discussion_id_link: —
bugzilla_pr: —
url: https://github.com/gcc-mirror/gcc/commit/8bce5e413a15abb756a1f6d0bc2f6741b6463b3e
date: 2026-09-21T00:00:00.000Z

GCC expands FP classification built-ins like isnan and isfinite into direct FP comparisons on alpha. These comparisons can trigger exceptions for subnormal operands, leading to program termination. This change replaces these direct comparisons with checks on the floating-point encoding in an integer register, preventing crashes with subnormal values. It also prevents unnecessary rewriting of these built-ins by the compiler.