Performance Tuning

Tips for getting the most simulation throughput from CIRCT — backend selection, compilation strategies, and design patterns that simulate efficiently.

Choosing a Backend

Backend	Speed	Use Case
Ruby interpreter	Slowest	Debugging, small designs
Verilator	Fast	Large designs, standard flow
Arcilator	Fastest	Maximum throughput, LLVM-optimized
Rust native	Fast	Standalone binaries
WASM	Near-native	Browser deployment

Compilation Optimization

Enable optimization passes during compilation:

rhdl compile lib/design.rb --optimize

Key optimizations:

• Common subexpression elimination — share duplicate logic
• Dead code elimination — remove unused signals
• Constant folding — evaluate compile-time constants
• Register merging — combine equivalent registers

Design Patterns for Speed

• Minimize memory depth — smaller memories simulate faster
• Avoid unnecessary hierarchy — flat designs have less overhead
• Use enable signals — skip computation when modules are idle
• Partition clock domains — simulate independent domains separately

Profiling

Identify simulation bottlenecks:

rhdl simulate lib/design.rb --profile

This reports time spent in each module, helping you focus optimization effort where it matters.