nx_parallelism_census.nx
buildroot/runtime/nx_parallelism_census.nx
about
nx_parallelism_census.nx -- the PARALLELIZATION WORKLIST: walk the tree and flag COMPUTE-HEAVY organs (deep
nested loops = the hot-path signature: matmul / convolution / codec macroblock loops / hashing rounds) that
would benefit from multi-core once NishiLang threading is unblocked. Every organ is single-threaded today
(thread-parallelism is toolchain-blocked, 2026-07-06), so this names WHERE the wins are when threads land.
Structural signal = max concurrent `while`-nesting depth (comment/string-aware). Depth>=3 = a candidate.
HONEST: a candidate-generator, not proof of compute-boundedness (a deep parse/IO loop can also nest);
confirm with nx_parallelism_probe on the actual run. Sovereign (getdents64 walk, no shell). expect_exit: 0
dependencies 3 imports · 0 importers
imports: nx_syscalls.nxnx_itoa_lib.nxnx_dir.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
| 12 | const PC_CAP: i64 = 16384 |
| 13 | const PC_BUF: i64 = 1048576 |
functions
| 15 | func pc_puts(s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } sys_write(1, s, n); return 0 } |
| 20 | func pc_putn(v: i64) -> i64 { nxi_out(v); return 0 } |
| 21 | func pc_alnum(c: u8) -> i64 called by 1: pc_max_loop_depth |
| 28 | func pc_match(buf: *u8, i: i64, total: i64, lit: *u8) -> i64 called by 1: pc_max_loop_depth |
| 33 | func pc_read(path: *u8, buf: *u8) -> i64 |
| 41 | func pc_max_loop_depth(buf: *u8, total: i64) -> i64 |
| 72 | func main() -> i64 |