Thirteenfold Systems — C++/WebAssembly website baseline measurements
Measured: 2026-09-09T21:59:38.893Z
Selected source commit: d05161ab2c897d4adebccc4fb367441ad99c5bc7
Build receipt: 2026-09-09T21:52:34.794Z; all 83 source digests matched that commit.
The receipt itself contains no commit field; the commit association was established by hashing git show COMMIT:PATH bytes.
Toolchain: Ubuntu clang version 18.1.3 (1ubuntu1); Ubuntu LLD 18.1.3.
Measurement runtime: v24.19.0 on linux x64; AMD EPYC 9V74 80-Core Processor.

BUILD ARTIFACTS (bytes; gzip is a local level-9 estimate)
server/engine.wasm: 365677 raw; 51798 locally gzipped. Server deployment artifact, not a browser download.
server/index.js: 12605 raw; 4656 locally gzipped. Server deployment artifact, not a browser download.
client/assets/site.17cd14e29919bea7.js: 12597 raw; 5229 locally gzipped. Browser stylesheet or script asset.
client/assets/site.fd23149c4bf6f043.css: 43118 raw; 9175 locally gzipped. Browser stylesheet or script asset.
The wasm engine is deployed to the server. A visitor receives HTML, CSS and browser JavaScript; this is not a browser wasm download.

WASM BOUNDARY
Imports: 0. Initial linear memory: 2,097,152 bytes (2 MiB). Declared maximum: 4,194,304 bytes (4 MiB).
Memory after measured requests: 2097152 bytes. Bounded input capacity: 8192 bytes.
These bounds concern wasm linear memory, not total Node/Worker process memory.
Exported interface: memory (memory), __wasm_call_ctors (function), memcpy (function), memset (function), tfs_input (function), tfs_input_capacity (function), tfs_reset (function), tfs_set (function), tfs_clock (function), tfs_flags (function), tfs_run (function), _ZN6BufferILm1000000EE3addEPKc (function), _ZN6BufferILm1000000EE4jsonE4View (function), _ZN6BufferILm1000000EE4htmlE4View (function), _ZN6BufferILm1000000EE3addE4View (function), tfs_kind (function), tfs_body (function), tfs_body_size (function), tfs_headers (function), tfs_headers_size (function), __dso_handle (global), __data_end (global), __stack_low (global), __stack_high (global), __global_base (global), __heap_base (global), __heap_end (global), __memory_base (global), __table_base (global)
Operations: 1 route/render; 2 pilot-field validation; 3 management token/action shape validation; 4 health-row render; 5 asset-cache policy.
Only operation 1 was measured. No real form data, service requests or authentication were used.

LOCAL WARM REQUEST LOOP
100 warmup requests per route; 7 batches of 300 measured sequential requests per route.
/: 17085 HTML bytes; median batch mean 0.0875 ms/request; batch-mean range 0.0828–0.1161 ms/request.
/applications: 17169 HTML bytes; median batch mean 0.0809 ms/request; batch-mean range 0.0795–0.0826 ms/request.
/trace: 14713 HTML bytes; median batch mean 0.0780 ms/request; batch-mean range 0.0758–0.0940 ms/request.
/contact: 9797 HTML bytes; median batch mean 0.0633 ms/request; batch-mean range 0.0620–0.0727 ms/request.
The loop includes input encoding, memory resets, wasm routing/rendering, output decoding, Headers construction and success assertions.
It excludes networking, TLS, Cloudflare scheduling/service I/O, the full HTTP host adapter, Response construction, concurrency and browser work.
Batch means are not per-request p50/p95 latency. Shared-host timing varies; no comparison or performance superiority is established.
Module compile single observation: 3.7129 ms; instance creation: 0.4138 ms. Cache effects are uncontrolled; not cold-start results.

REPRODUCE
node measure-baseline.mjs ./baseline
Requires Node >=22 and the selected baseline output files, build receipt and provenance.json. Output includes exact hashes and raw timing batches.

LIMITATIONS
Local gzip numbers do not establish HTTP compression, actual transferred bytes, mobile loading speed, Core Web Vitals, real-user performance, security qualification or a product release.
This immutable v2.5.0 baseline will differ from later builds containing new content. Re-run the script against a selected later build to obtain updated sizes.