Static copy
Copied from docs/en/evidence/cross-language.md in https://github.com/poppy-team/aipo-lang (MIT). Pinned to revision ce47c4e070404fc41722ccd8e2b7b808b99062a8, blob 31575dfc7a9d96ebbe30c38d78284ae3c602d835. The source repository remains canonical; this copy is refreshed through a sync pull request, not live.
Cross-Language Benchmarks
Aipo provides a standardized benchmarking harness (aipo-bench --compare) evaluating wall-clock execution speed and memory footprint (Peak RSS) against established runtime ecosystems.
Benchmark Scope
Our goal is measuring real-world workload execution costs without obscuring architectural distinctions between interpreters, JIT compilers, and ahead-of-time compiled native code:
- Aipo VM In-Process: Compilation occurs before the clock starts; measures VM instantiation, call frame allocation, and bytecode instruction dispatch.
- Aipo CLI / VM (
aipo run): Cold process startup (source reading, lexing, parsing, semantic analysis, bytecode emission, and execution). - Aipo → JavaScript (Node.js): Emitted JavaScript executed on the V8 engine via
aipo-js. - Reference Runtimes: Lua 5.4, LuaJIT, Wren 0.4.0, Luau 0.739, CPython 3.12, PyPy 8.0.0, Ruby, and native Rust (algorithmic baseline).
Six Canonical Workloads
All benchmarks compute a standardized checksum:<value> to verify that optimizing compilers did not elide legitimate computation:
| Workload | Measured Domain | Primary Evaluation Goal |
|---|---|---|
arithmetic | Tight loop with function calls and intensive math | Basic dispatch overhead and call frame mechanics |
collections | Dynamic list construction and traversal | Heap allocation pressure and container indexing |
fields | Struct with 6 fields, mutation, and method calls | Property lookups and monomorphic field caching |
strings | Incremental ASCII string concatenations | Text buffer growth and reallocation characteristics |
recursion | Deep recursive calculation of Fibonacci(24) | Call stack depth and pure invocation overhead |
startup | Minimal process initialization and exit | Cold startup time and standard library bootstrap |
Reproducing the Results
To reproduce these measurements locally:
# Compile binaries in optimized release mode
cargo build --release -p aipo-cli -p aipo-bench
# Execute the complete comparison suite
target/release/aipo-bench --compare --compare-json target/cross-language.json
# Quick developer smoke run
target/release/aipo-bench --compare --compare-quick