Benchmarks

Milo compiles via LLVM with -O2 — the same backend as C and Rust. On most workloads it lands within noise of C.

Results

MiloCGo

matmul 512×512

12ms

12.8ms

13.2ms

binarytrees depth 18

3ms

3.9ms

10.5ms

quicksort 2M f64

34.7ms

35.7ms

34.7ms

startup empty main

1.2ms

1.2ms

1.5ms

stringops 100k concat

3.2ms

3.1ms

6.5ms

fib(42)

20.8ms

18.4ms

21.6ms

sieve to 10M

2.5ms

2.1ms

3.4ms

maplookup 100k

4.4ms

3.3ms

5ms

grep -c 5MB

5.5ms

2.1ms

4ms

json parse+walk 1MB

7.1ms

1.6ms*

9.7ms

Apple M-series, macOS. *C uses yyjson; Go and Milo use their stdlibs.

Notes

• Values < 1.0x mean Milo is faster than C (usually within measurement noise)
• 4 benchmarks match or beat C; 3 more within 20%
• Slower entries (grep, maplookup) have known hot spots — not fundamental limits
• Go's GC overhead shows clearly in allocation-heavy benchmarks (binarytrees 3.5x slower)
• JSON: flat pool parser with source-offset strings — ~6 allocations for any document size. Beats Go; remaining gap vs C is stdlib-vs-yyjson

Reproduce

./benchmarks/run.sh