Nift Labs
Nift Labs / benchmark collection
Eight experiments, one evidence policy: correct output, controlled runs, raw samples and conclusions with boundaries.
01 / Published reports
Startup-sensitive snippets, sustained computation, and file/data workloads. Compare all participating languages with checked outputs and retained samples.
Read the benchmark →From 100 to 10,000 pages: full builds, Nift incremental and targeted builds, elapsed time, and memory—with output equivalence and state boundaries attached.
Read the benchmark →Startup and RC costs, native scripting, process scaling and mixed 100k-file tasks. Bash, Zsh, Fish, Nushell and Nift, with checked outputs and explicit native/Unix-tool boundaries.
Read the benchmark →GNU Make, Ninja and Nift execute the same native graph. No-op and incremental scaling, generated inputs, compiler-heavy builds and tiny actions—with exact rebuilt sets and output hashes.
Read the benchmark →Nift, Turbo and Nx execute controlled project workflows and tiny tasks. Exact execution sets, file invalidation, concurrency, targeting, failure recovery and local-state boundaries.
Read the benchmark →Five engines assemble identical documents. Fresh invocation, preparation, reuse, loops, partials and memory, with exact byte checks and a candid practitioner assessment.
Read the benchmark →Five runtimes solve equivalent concurrency problems: scheduling, waits, CPU scaling, bounded pools, queues, shared state, subprocesses and memory. Exact correctness and explicit model boundaries.
Read the benchmark →Five runtimes traverse, inspect, fingerprint and transform identical project trees. Exact outputs, controlled mutations, warm-state boundaries and separate syscall/allocation diagnostics.
Read the benchmark →These are bounded implementation experiments on dated shared Linode nodes, with controlled CPU affinity and per-series worker limits. Fresh application state is not a machine-cold boot. Correctness is checked for every sample. Toolchains, commands, machine inventory and raw evidence accompany each report.
Minimal generated-page throughput and microbenchmarks do not replace real-site migration experiments. Read the workloads and limitations before quoting a result.
All eight benchmark reports use Nift v4.12.0 final candidate, frozen source c14d06835bca515a5281e41da0e40219c2d60ef0. Seven suites retain their independently audited final recertification datasets. The website-generator report adds a fresh same-size four-vCPU cohort at 100, 1,000 and 10,000 pages: Nift uses all workers and Hugo uses its default parallelism. Worker and state boundaries accompany every report; historical evidence remains archived separately.
Eight-suite final recertification evidence · New all-core website measurements.
Task runners: source-built Nift 4.12.0 final candidate, Turbo 2.11.7 and Nx 23.3.0 on one fresh node. Controlled project tasks, exact execution checks and explicit local-state/cache boundaries. Task workflow, native capability gaps and frozen evidence.
Five pinned engines assemble byte-identical documents on one fresh node: Nift 4.12.0, Jinja2 3.1.6, Nunjucks 3.2.4, Handlebars 4.7.10 and LiquidJS 10.30.0. Fresh invocation, preparation, compiled reuse and source reuse have separate measurement boundaries. Template composition, correctness gates and frozen evidence.