code wiki / _hdl_build / nx_js_sota_scorecard.nx
nx_js_sota_scorecard.nx
buildroot/runtime/_hdl_build/nx_js_sota_scorecard.nx
about
nx_js_sota_scorecard.nx -- HONEST placement of the sovereign Nishi JS engine vs the 2026 SOTA engines
(V8, JavaScriptCore/WebKit, SpiderMonkey) and the honest interpreter-class peer (QuickJS). This is an
ARCHITECTURAL scorecard, NOT a benchmarked head-to-head: we CANNOT run JetStream2/Speedometer3 (they need
a full browser + complete language + WASM + workers), and every speedup we've measured is vs OUR OWN
tree-walker, NOT vs V8 -- so this card states tiers/features honestly and names what we have not measured.
Competitor levels are GROUNDED in the sovereign research corpus knowledge/fetched/jsvm_*.raw (v8.dev
Sparkplug/Maglev/TurboFan/Ignition, JavaScriptCore FTL/B3, SpiderMonkey Warp/Ion, JetStream). Our levels
are GROUNDED in our own measured gates (parity 110/110, conf 74/74, jit 49/49, shape 8/8, poly 3/3,
realbench 5/5, consumer-swap 17/17). Level scale: 0 NONE, 1 PARTIAL, 2 HAVE, 3 MATURE/HARDENED.
A LIAR-KILL asserts the card must NOT rate us at-or-above SOTA overall (we are behind) -- if it ever does,
the numbers were gamed and the gate goes RED. expect_exit: 0 license_tier: ORIGINAL
dependencies 1 imports · 0 importers
imports: nx_syscalls.nx
imported by: nobody (leaf or entry point)
call flow from main pre-order; caps 40 nodes / depth 6 declared; ↻ = already shown
structs
| none |
consts
| none |
functions
| 14 | func sc_w(s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } sys_write(1, s, n); return 0 } |
| 15 | func sc_n(v: i64) -> i64 |
| 29 | func lvl(v: i64) -> i64 |
| 37 | func row(axis: *u8, v8: i64, jsc: i64, sm: i64, qjs: i64, nx: i64, sov: i64, tal: *i64) -> i64 |
| 68 | func main(argc: i64, argv: *i64) -> i64 |