Browse documentation
DocsBenchmarks

Benchmarks

Current parsing and streaming comparisons, presented with enough context to interpret the numbers.

01

TMX materialized catalog

FrameworkLanguageMedianThroughputPeak cold RSS
LokitRust / Python0.240s208,143 units/s80.4 MiB
Okapi FrameworkJava0.263s189,437 units/s269.2 MiB
translate-toolkitPython0.652s76,586 units/s135.7 MiB

Workload: 50,000 units from a 10.75 MB TMX corpus. These values are retained from the current site while the benchmark suite is rerun.

02

XML streaming comparison

WorkloadLokit medianOkapi materialize medianRatio
TMX synchronous stream0.196s0.263s1.34x
XLIFF synchronous stream0.226s0.309s1.37x

Streaming and materialization have different output contracts. Keep that distinction visible when interpreting cross-framework ratios.