Die Yield & Cost-per-Good-Die Simulator

Set a die size, wafer, and defect environment, then run it — the simulation executes on the ChipFoundryServices distributed compute pool. It steps the die grid to count how many whole dies pack onto the wafer inside the edge-exclusion ring (gross-die-per-wafer), applies four classic defect-limited yield models (Poisson, Murphy, negative-binomial clustering, and Seeds) to find the surviving fraction, and reports good-die-per-wafer, packing efficiency and the cost-per-good-die — the number that actually sets the price of a GPU or accelerator die. Big reticle-limit AI dies live or die on defect density. Reduced-order educational model. See also the transistor I-V, thermal, interconnect RC, CMP planarization and lithography simulators and the compute-pool status.

AI GPU (reticle) Mobile SoC Mature 200 mm
Wafer map — whole dies fit inside the edge-exclusion ring (green); edge-clipped dies are lost (red)
Yield vs die size — four defect models; marker is your die at the current D0
Developer API — same simulation over HTTP (load-balanced across the pool): 
curl -X POST https://www.chipfoundryservices.com/edge/yield \
  -H "Content-Type: application/json" \
  -d '{"die_width_mm":26,"die_height_mm":33,"wafer_diameter_mm":300,"defect_density_d0":0.12,
       "clustering_alpha":3.0,"yield_model":3,"edge_exclusion_mm":3,"scribe_lane_mm":0.10,"wafer_cost_usd":16000}'
Returns JSON with outputs (die_area_mm2, die_area_cm2, gross_die_per_wafer, defects_per_die, yield_percent, good_die_per_wafer, packing_efficiency_percent, cost_per_good_die_usd, cost_per_gross_die_usd, yield_poisson_percent, yield_murphy_percent, yield_negbinom_percent, yield_seeds_percent, model_name, verdict), the full profile (80-point yield_curve, wafer packing geometry), the serving node, and compute_ms.