nx_nofloat_qwen_fastgen_i32_gate.nx
buildroot/runtime/nx_nofloat_qwen_fastgen_i32_gate.nx
about
nx_nofloat_qwen_fastgen_i32_gate.nx -- Stage 1 of the decode-speed arc: the dequant-once cache narrowed
i64 -> i32. Q16 weights are real*65536 with |real|<32768, so they fit i32 EXACTLY -> half the weight
bandwidth (2.73->1.43 GB/token) with ZERO precision loss. Teeth: (a) narrow overflow count == 0 (proves
every value fit i32 => lossless by construction), (b) tokens still [12095,13,1084,374] bit-exact vs the
i64 fastgen path, (c) faster than 404 ms/token. No hw writes (Rule 26). expect_exit: 0 license_tier: ORIGINAL
dependencies 10 imports · 0 importers
imports: nx_syscalls.nxnx_tier.nxnx_le.nxnx_tensor.nxnx_gguf.nxnx_gguf_load.nxnx_gguf_meta.nxnx_nofloat_llm.nxnx_nofloat_tok.nxnx_gate_verdict.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
| 17 | func fi_puts(s: *u8) -> i64 { var n: i64=0; while s[n]!=(0 as u8){n=n+1} sys_write(1,s,n); return 0 } |
| 18 | func fi_num(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 as u8;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 } |
| 19 | func fi_slen(s: *u8) -> i64 { var n: i64=0; while s[n]!=(0 as u8){n=n+1} return n } called by 1: main |
| 21 | func main() -> i64 |