code wiki / _hdl_build / nx_simd_sse_shuf_gate.nx
nx_simd_sse_shuf_gate.nx
buildroot/runtime/_hdl_build/nx_simd_sse_shuf_gate.nx
about
nx_simd_sse_shuf_gate.nx -- SOVEREIGN 128-bit pshufd run-KAT (rung R2+R3:
3-OPERAND assembler + packed dword shuffle), on REAL x86 silicon, NO gcc/
qemu/binutils. AUTHOR=ORGAN, no-false-green.
Proves (a) the assembler now parses a THIRD operand (the prerequisite for
pshufd AND all VEX/EVEX 256/512-bit ops), and (b) pshufd routes lanes by its
imm control byte. src=[11,22,33,44] (4x u32):
POS: pshufd $2 -> dst lane0 = src[2] = 33 -> exit(33).
NEG: pshufd $0 -> dst lane0 = src[0] = 11 -> exit(11).
GREEN iff pos==33 AND neg==11 AND pos!=neg (distinct => the imm genuinely
selects the lane, and op0=imm/op1=src/op2=dst threaded correctly). Completes
the sovereign 128-bit integer SIMD surface. 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
| 15 | const ASM_TOOL: *u8 = "_offc/nxasm_x86_main.elf" |
| 16 | const SV_LOG: *u8 = "knowledge/status/simd_vec_kat.log" |
functions
| 18 | func gw(fd: i64, s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } sys_write(fd, s, n); return 0 } |
| 19 | func gwn(fd: i64, v: i64) -> i64 |
| 30 | func g_run(path: *u8, argv: *i64) -> i64 |
| 50 | func asm_and_run(spath: *u8, elfpath: *u8) -> i64 |
| 64 | func write_shuf_s(path: *u8, sel: i64) -> i64 |
| 85 | func g_emit(fd: i64, pos: i64, neg: i64, ok: i64) -> i64 |
| 93 | func main() -> i64 |