Landing: 806feb8e017b
Project / Subsystem
gcc / gcc/forwprop
Date
2026-07-08
Author
Konstantinos Eleftheriou
Commit
806feb8e017b243de37e4502a6fc01fb76dd1529
Source
github
Perf win
Yes
Breaking
No
All attributes
- project
- gcc
- subsystem
- gcc/forwprop
- patch_id
- —
- commit_hash
- 806feb8e017b243de37e4502a6fc01fb76dd1529
- source_type
- github
- headline
- forwprop: Match and fold the long-multiply carry form
- tldr
- Recognizes and folds a specific pattern of long-multiply summation to improve performance.
- author
- Konstantinos Eleftheriou
- outcome
- committed
- performance_win
- true
- breaking_change
- false
- series_id
- —
- series_parts
- []
- tags
-
- • optimization
- • performance
- • aarch64
- discussion_id_link
- —
- bugzilla_pr
- —
- date
- 2026-07-08T00:00:00.000Z
The forwprop pass now recognizes a schoolbook expansion of a 2N-bit unsigned multiply. It folds this pattern into a widening multiply plus a right shift for the high part and a plain MULT_EXPR for the low part. This optimization significantly reduces the instruction count for operations like those in SPEC2026’s 750.sealcrypto_r, transforming 16 instructions down to a few, improving performance.