code wiki / _hdl_build / nx_gpu_matcher.nx

nx_gpu_matcher.nx

buildroot/runtime/_hdl_build/nx_gpu_matcher.nx

3516 B46 linesdepth 2pulls 2 transitivereach 0 importersview sourcekind tooltopic gpu
docsdependenciesstructsconstsfunctions

about

nx_gpu_matcher.nx -- sovereign GPU-provider MATCHER for our exact sovereign-development spec (verified 2026-07-14). The axis that actually filters providers is NOT the 5090 or CUDA (install those yourself as root) -- it is raw /dev/nvidia* + nvidia-smi -lgc, which need VM-passthrough or physical METAL (you OWN the driver), NOT a container (root-in-container shares the host driver -> clock/device denied). "root alone" is the trap. Data below verified against live provider pages + NVIDIA docs. Belongs on the compare-plane as /compare/gpu-sovereign.

dependencies 1 imports · 0 importers

nx_syscalls.nx nx_gpu_matcher.nx

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

main puts sys_write row puts ↻ putn sys_mmap sys_write ↻ putn ↻

structs

none

consts

7const NP: i64 = 7
9const REQ_ALL: i64 = 255

functions

11func puts(s: *u8) -> i64 { var n: i64=0; while s[n]!=(0 as u8){n=n+1} sys_write(1,s,n); return 0 }
called by 2: rowmain calls 1: sys_write
12func putn(v: i64) -> i64 { let b: *u8=sys_mmap(28); var m: i64=v; if m<0{m=0-m;sys_write(1,"-" as *u8,1)}; let t: *u8=sys_mmap(28); var k: i64=0; if m==0{t[0]=48;k=1}; while m>0{t[k]=(48+(m%10)) as u8;m=m/10;k=k+1}; var i: i64=0; while i<k{b[i]=t[k-1-i];i=i+1}; sys_write(1,b,k); return 0 }
called by 2: rowmain calls 2: sys_mmapsys_write
14func row(nm: *u8, met: i64, cents: i64, note: *u8) -> i64
called by 1: main calls 2: putsputn
31func main() -> i64
calls 3: putsrowputn