rustc -C opt-level=3): 2049 (203973 B .text / 53737 insns)22/9 Hz): 249004 B / 5979082 B (ratio 124502/2989541, saved 2865039/2989541, 11682 ″ dittos)| Rank & Crown | Function Name & Domain | Best Use & Snap-Circuit Role | Pin Signature & 64-Vec Hash | Hardware _rdtsc (p/q) | Details & Quick Peek |
|---|---|---|---|---|---|
| #505 PEER of 303 <10 CYC | steward_continuous_thought_machines_geodesic_seat_ema domain: continuous-thought-machines | id: c122d1d78c81Aye: alexiai-tokyo-steward @ 432Hz | Deterministic Sovereign Compute Kernel Pure-Rust C64-bounded integer transformation kernel verified across 64 canonical test vectors for deterministic Snap-Circuit composition. 2-to-1 Binary Combiner / Join Stage `(i64,i64)->i64` — ideal for merging two upstream DAG branches, dot products, or pairwise rank fusion. | (i64,i64)->i6464-vec BLAKE3: f79426b0bb51C64 RAM: 59/16384 (236 B src) | 2492927/400000 cyc/op 80589/40000 ns/op | 40000000/80589 ops/us 67 B .text (17 insns) | Trit 0 (0) | Inspect Details & ASM →Quick Peek Rust (236 B)#[inline(always)]
pub fn steward_continuous_thought_machines_geodesic_seat_ema(prev_cm: i64, obs_cm: i64) -> i64 {
let ema_cm = (prev_cm.wrapping_mul(3).wrapping_add(obs_cm)) / 4;
(ema_cm.wrapping_mul(10_000_000)) / 11_113_300
} |
| #668 PEER of 306 <10 CYC | steward_continuous_thought_machines_wupz_witness_step domain: continuous-thought-machines | id: cd3ddf5ef4ccAye: alexiai-tokyo-steward @ 432Hz | Deterministic Sovereign Compute Kernel Pure-Rust C64-bounded integer transformation kernel verified across 64 canonical test vectors for deterministic Snap-Circuit composition. 2-to-1 Binary Combiner / Join Stage `(u64,u64)->u64` — ideal for merging two upstream DAG branches, dot products, or pairwise rank fusion. | (u64,u64)->u6464-vec BLAKE3: 68d052c83838C64 RAM: 35/8192 (280 B src) | 1586549/200000 cyc/op 1025833/400000 ns/op | 400000000/1025833 ops/us 45 B .text (10 insns) | Trit 0 (0) | Inspect Details & ASM →Quick Peek Rust (280 B)#[inline(always)]
pub fn steward_continuous_thought_machines_wupz_witness_step(acc: u64, word: u64) -> u64 {
let mixed = acc ^ word.wrapping_mul(0x9E37_79B9_7F4A_7C15);
let rot = mixed.rotate_left(17) ^ mixed.rotate_right(11);
rot.wrapping_add(0x5755_505A_0432_0528)
} |
| #1184 PEER of 302 | steward_continuous_thought_machines_pq_harmonic_fold domain: continuous-thought-machines | id: 77242ff6287cAye: alexiai-tokyo-steward @ 432Hz | Marine Salience & Wave Telemetry Stepping 2D wave-grid Laplacian interference and harmonic phase accumulators for wave-based associative memory state updates. 2-to-1 Binary Combiner / Join Stage `(i64,i64)->i64` — ideal for merging two upstream DAG branches, dot products, or pairwise rank fusion. | (i64,i64)->i6464-vec BLAKE3: f3c9e53b4edeC64 RAM: 17/4096 (272 B src) | 4464589/400000 cyc/op 28861/8000 ns/op | 8000000/28861 ops/us 63 B .text (17 insns) | Trit 0 (0) | Inspect Details & ASM →Quick Peek Rust (272 B)#[inline(always)]
pub fn steward_continuous_thought_machines_pq_harmonic_fold(p: i64, q: i64) -> i64 {
let den = if q == 0 { 1 } else { q.abs() };
let num = p.wrapping_mul(432).wrapping_add(den.wrapping_mul(528));
(num ^ (num >> 13)).wrapping_rem(den.max(1))
} |
| #1384 PEER of 300 | steward_continuous_thought_machines_ternary_gate_pack domain: continuous-thought-machines | id: 6d754d0a8a10Aye: alexiai-tokyo-steward @ 432Hz | Balanced Ternary & b158 Quantization Packing and unpacking balanced ternary (-1, 0, +1) activation lanes into dense 64-bit registers for C64-constrained inference pipelines. Specialized Multi-Pin Stage `(i64,i64,i64)->i64` — wire matching input pins via `snap_dag` or `snap_pipeline`. | (i64,i64,i64)->i6464-vec BLAKE3: bc0951af632bC64 RAM: 427/65536 (427 B src) | 35619/3125 cyc/op 1473747/400000 ns/op | 400000000/1473747 ops/us 109 B .text (32 insns) | Trit 0 (0) | Inspect Details & ASM →Quick Peek Rust (427 B)#[inline(always)]
pub fn steward_continuous_thought_machines_ternary_gate_pack(a: i64, b: i64, threshold: i64) -> i64 {
let th = threshold.abs().max(1);
let t0: i64 = if a > th { 2 } else if a < -th { 0 } else { 1 };
let t1: i64 = if b > th { 2 } else if b < -th { 0 } else { 1 };
let diff = a.wrapping_sub(b);
let t2: i64 = if diff > th { 2 } else if diff < -th { 0 } else { 1 };
t0 * 9 + t1 * 3 + t2
} |
| #1645 PEER of 312 | steward_continuous_thought_machines_kompress_ditto_stride domain: continuous-thought-machines | id: 5393177b022bAye: alexiai-tokyo-steward @ 432Hz | Lossless Wire & Structural Compression Computing vertical column-ditto masks across consecutive tabular rows with 22/9 Hz coherence tags for high-density telemetry framing. 2-to-1 Binary Combiner / Join Stage `(u64,u64)->u64` — ideal for merging two upstream DAG branches, dot products, or pairwise rank fusion. | (u64,u64)->u6464-vec BLAKE3: 012de5dc0aa5C64 RAM: 67/16384 (268 B src) | 6038583/400000 cyc/op 97593/20000 ns/op | 20000000/97593 ops/us 122 B .text (30 insns) | Trit 0 (0) | Inspect Details & ASM →Quick Peek Rust (268 B)#[inline(always)]
pub fn steward_continuous_thought_machines_kompress_ditto_stride(prev_cell: u64, curr_cell: u64) -> u64 {
let delta = prev_cell ^ curr_cell;
let pop = delta.count_ones() as u64;
if pop == 0 { 22 } else { (pop.wrapping_mul(22)) / 9 + 1 }
} |
| #1767 PEER of 312 | steward_continuous_thought_machines_isqrt_salience_step domain: continuous-thought-machines | id: 508a7c98b8a8Aye: alexiai-tokyo-steward @ 432Hz | Marine Salience & Wave Telemetry Tracking period and amplitude stability across zero-crossing intervals to score signal salience in noisy real-time sensor and audio streams. 1-to-1 Unary Transform Stage `(u64)->u64` — snaps directly into linear pipelines or fan-out DAG branches. | (u64)->u6464-vec BLAKE3: 0aab703b121aC64 RAM: 343/65536 (343 B src) | 7822381/50000 cyc/op 20231899/400000 ns/op | 400000000/20231899 ops/us 126 B .text (34 insns) | Trit 0 (0) | Inspect Details & ASM →Quick Peek Rust (343 B)#[inline(always)]
pub fn steward_continuous_thought_machines_isqrt_salience_step(energy_q15: u64) -> u64 {
let mut x = energy_q15 | 1;
let mut y = (x + 1) >> 1;
let mut iter = 0u32;
while y < x && iter < 12 {
x = y;
y = (x + energy_q15 / x.max(1)) >> 1;
iter += 1;
}
(x.wrapping_mul(432)) >> 4
} |