Blyx v0.1.0-alpha is here! Read the release notes ->
EMPIRICAL BENCHMARKS • HARDWARE TEST SUITE

Blyx Performance Metrics

Empirical hardware benchmark measurements across matrix math, actor message throughput, compiler speed, and binary size.

Execution Time (lower is better)

1000x1000 Matrix Multiplication Speed

Blyx is 2.8x faster than Python NumPy

Blyx v0.1.0-alpha12.4 ms
C++ GCC -O311.8 ms
Rust 1.80 -O312.1 ms
Python NumPy (OpenBLAS)34.7 ms
Messages / sec (higher is better)

Actor Message Throughput

142 Million lock-free messages per second

Blyx Actor Runtime142M / sec
Erlang / OTP38M / sec
Go Channels52M / sec
Akka Scala45M / sec
Cold Build Seconds (lower is better)

Compiler Throughput (100,000 LOC)

Cold build in 4.2 seconds; incremental rebuild in 0.3s

Blyx (blyxc)4.2 s
Go (gc)3.8 s
Rust (rustc)18.4 s
C++ (clang++)14.2 s
Binary Size in KB (lower is better)

Standalone Binary Footprint

Hello World binary is only 48 KB

Blyx Binary48 KB
C Binary (strip)40 KB
Rust Binary380 KB
Go Binary2,100 KB
Benchmark Methodology & Environment
All benchmarks measured on x86_64 Linux (Intel Core i9-13900K @ 5.8 GHz, 64GB DDR5 RAM, Ubuntu 24.04 LTS).
Blyx compiler version v0.1.0-alpha built with BIR SSA passes and LLVM backend (-O3 optimization level).