LRU cache

Hand-rolled LRU (doubly-linked list + map) servicing 100k mixed get/put operations.

Runtime Compile time

Compile time · median (cold)

median of 10 runs

Native Gocompiled

180 ms109×

Piko interpbytecode VM

1.66 msbaseline

CPython 3.13bytecode VM

449 µs0.27×

PyPy 7.3tracing JIT

368 µs0.22×

Ttengobytecode VM

313 µs0.19×

Sscriggobytecode VM

unsupported in scriggo

n/a

Mmvmbytecode VM

461 µs0.28×

YyaegiAST walker

567 µs0.34×

Full statistics

Runner	N	Compile	Runtime	P95	Stddev	RSS	vs piko	Status
Native Gocompiled	10	180 ms	16.2 ms	16.7 ms	527 µs	69 MiB	109×	OK
Piko interpbytecode VM	10	1.66 ms	155 ms	155 ms	351 µs	183 MiB	1.00×	OK
CPython 3.13bytecode VM	10	449 µs	98.7 ms	101 ms	1.92 ms	n/a	0.27×	OK
PyPy 7.3tracing JIT	10	368 µs	15.0 ms	16.3 ms	516 µs	n/a	0.22×	OK
tengobytecode VM	10	313 µs	174 ms	177 ms	5.21 ms	872 MiB	0.19×	OK
scriggobytecode VM	0	n/a	n/a	n/a	n/a	n/a	n/a	unsupported
mvmbytecode VM	10	461 µs	323 ms	361 ms	20.8 ms	63 MiB	0.28×	OK
yaegiAST walker	10	567 µs	966 ms	996 ms	23.9 ms	66 MiB	0.34×	OK

Workload & symmetry rules

Workload

A cache of size 4096 sees a 100,000-op trace of 70% reads / 30% writes. Implement using a hash map plus a doubly-linked list for recency. Print the final hit/miss totals.

Symmetry rules

Hand-rolled linked list; no OrderedDict, no functools.lru_cache.
Generic map / dict for indexing.

Source code

piko / Go
piko_source.go
native Go
native_main.go
CPython / PyPy
cpython.py
tengo
script.tengo