dotnet/roslyn

Compiler should optimize common boolean return patterns

Open

#34,726 opened on Apr 3, 2019

View on GitHub
 (9 comments) (0 reactions) (0 assignees)C# (4,257 forks)batch import
Area-CompilersCode Gen Qualityhelp wanted

Repository metrics

Stars
 (20,414 stars)
PR merge metrics
 (Avg merge 6d 17h) (256 merged PRs in 30d)

Description

Version Used: Rosyln master branch from 2nd April 2019

Steps to Reproduce: [Sharplab Demo]

  1. Have this code: public bool Proper(object o) { return o is object; }
  2. Also, have this unoptimized version public bool ShouldBeOptimized(object o) { if (o is object) { return true; } return false; }

Expected Behavior: Identical IL output, as they are semantically identical. This wouldn't be an issue but it does affect JIT output seemingly (JIT64 desktop, tested on Sharplab above) - it switches a setcc ; movzx pattern to a full branch, which is not ideal

Actual Behavior: IL generation has an additional branch and changes the generated JIT

It is a very minor issue but was recommended to create an issue for it

Contributor guide

Research direction
Identify the IL generation method for boolean returns in the Roslyn compiler; look for patterns where an if else block returning true/false can be optimized to a single comparison and conditional set instruction (e.g., setcc, movzx). Trace how the compiler emits IL for the given C# patterns and modify the optimizer to recognize and transform the unoptimized pattern.
Tech stack
csharp
Domain
compilers
Issue type
Performance
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
2
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-3 hours
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
C#
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
50