| name | tile-code-planar-architectures |
| description | Strictly local tile-code architectures for quantum error correction on 2D planar lattices. Tile codes are planar qLDPC codes with weight-6 stabilizers and open boundary conditions, offering up to 4x encoding efficiency vs. surface code. Uses SWAP-based routing schemes for syndrome extraction with nearest-neighbor interactions only, matching surface code connectivity. Activation: tile code architecture, planar qLDPC codes, SWAP routing syndrome extraction, quantum error correction architecture, surface code alternative, qLDPC on 2D lattice. |
| metadata | {"arxiv_id":"2607.05897","published":"2026-07-09","tags":["quantum","error-correction","qLDPC","tile-codes","planar-architecture","syndrome-extraction","surface-code"]} |
Tile-Code Planar Quantum Error Correction Architectures
Description
Tile codes are a family of planar quantum low-density parity-check (qLDPC) codes with weight-6 stabilizers and open boundary conditions, offering encoding efficiency $kd^2/n$ up to 4x the surface code. This skill covers the design of strictly local tile-code architectures on 2D planar lattices using only nearest-neighbor interactions.
Core Architecture
Tile Code Properties
- Stabilizer weight: 6 (vs. 4 for surface code)
- Boundary conditions: Open (planar)
- Encoding efficiency: Up to 4x surface code ($kd^2/n$)
- Connectivity: 2D square lattice with nearest-neighbor only (matches surface code)
SWAP-Based Routing
- Exhaustive search algorithm finds SWAP-based routing schemes for syndrome extraction
- Routes qubits through 2D lattice to enable multi-body stabilizer measurements
- Four tile-code families analyzed with explicit routed syndrome-extraction circuits
Performance Trade-offs
- Without routing constraint: Threshold 0.23%-0.31% (SI1000 noise model)
- With routing constraint: Threshold 0.11%-0.13% (reduction factor 2-3x)
- Crossover point: At physical error rate $p^* \approx 0.08%$, tile codes become more qubit-efficient than surface code
- Below crossover: Tile code advantage grows monotonically as physical error rate decreases
Usage Patterns
Pattern 1: Architecture Selection
- Determine target physical error rate of hardware
- If $p < 0.08%$ (SI1000 model): tile codes offer better qubit efficiency
- If $p > 0.08%$: surface code may be preferable due to higher threshold
- Evaluate routing overhead: factor of 2-3x threshold reduction must be offset by encoding efficiency gain
Pattern 2: Syndrome Extraction Circuit Design
- Select tile-code family from the four analyzed families
- Use exhaustive search algorithm to find optimal SWAP-based routing
- Construct explicit routed syndrome-extraction circuits
- Decode with BP+OSD (Belief Propagation + Ordered Statistics Decoding)
- Estimate circuit-level threshold under target noise model
Pattern 3: Resource Footprint Analysis
- Calculate physical qubits per logical qubit for tile code at target code distance
- Compare against surface code baseline
- Account for SWAP gate overhead in syndrome extraction
- Include routing-induced error rate increase in total error budget
Pitfalls
- Threshold penalty underestimation: Routing reduces threshold by 2-3x. Design margins must account for this, not just the nominal threshold.
- Crossover rate dependency: The 0.08% crossover point is specific to the SI1000 noise model. Different noise models will shift this point. Always recalculate for your hardware's noise characteristics.
- BP+OSD decoder complexity: BP+OSD is more complex than minimum-weight perfect matching used for surface codes. Factor in decoder latency and computational resources.
- Four-family limitation: Only four tile-code families were analyzed. Other families may offer different trade-offs — don't assume these four are exhaustive.
- 2D lattice constraint: This architecture is specifically for 2D square lattices. 3D or non-square lattices may enable different routing schemes with lower overhead.
Related Skills
quantum-error-correction-methods — umbrella skill for QEC methodologies
surface-code-lattice-surgery — surface code with lattice surgery operations
vine-codes-qldpc — another qLDPC code family for comparison