Expand description
High-performance photonic network-on-chip primitives.
Accelerates the hot paths in the Python photonic_noc bridge:
- Waveguide routing — O(N²) Manhattan distance on mesh
- Crosstalk analysis — O(N²) pairwise channel coupling
- MZI transfer matrix — 2×2 unitary matrix cascade
- Power budget — path loss accumulation
Structs§
- Crosstalk
Bank Result - Crosstalk
Pair Result - Crosstalk
Result - Crosstalk result per channel.
- Power
Budget Result - Path power budget result.
- Waveguide
Result - A routed waveguide segment.
Constants§
Functions§
- analyze_
crosstalk - Analyze inter-channel crosstalk for WDM channels.
- analyze_
crosstalk_ bank - Analyse crosstalk in a uniform parallel-waveguide bank. Adjacent pairs
(gap = g) are the dominant term; next-nearest (gap = 2g) are included
as the largest secondary term — Marcatili 1969 predicts that all other
pairs are at least
exp(-2·g/L_decay)smaller still. - analyze_
crosstalk_ pairs - Per-pair crosstalk for arbitrary waveguide geometry.
pairscarries(idx_a, idx_b, gap_nm, coupling_length_um)per pair. Evaluated in parallel via Rayon — this is the O(N²) path the commercial layout tools call after full pair enumeration. - analyze_
power_ budget - Analyze power budget for all waveguide paths.
- cascade_
mzi - Cascade N MZI stages by multiplying transfer matrices.
- complex_
mat_ 🔒mul - Multiply two 2×2 complex matrices (each stored as 8 f64s).
- mzi_
transfer_ matrix - MZI 2×2 transfer matrix elements (complex).
- pair_
coupling 🔒 - route_
waveguides - Route waveguides on a mesh topology from an adjacency matrix.
- thermal_
power_ for_ phase - Compute electrical power (mW) needed for a phase shift.