Solver

SolvLRDE

Replace time-stepping for linear systems. Cost independent of horizon and stiffness.

Every time-stepper marches through hundreds of small steps. SolvLRDE evaluates the answer at any time T directly via a Talbot-contour integral in the Laplace domain — no marching, no horizon dependence, no stiffness penalty. One LU factorization reused across every strike, scenario, and parameter. Seven vertical SKUs from derivatives pricing to power-grid EMTP.

Contour integral bypassing the time-stepping march, collapsing directly to the final answer.

Imagine yourself in these moments. Same product, different industries.

Hedge-fund pricing

The vol surface that takes three minutes.

Your vol surface has 6,000 cells — 200 strikes times 30 maturities. Every time the market moves, the surface has to be re-priced. Today the inner loop is a time-stepper: Crank-Nicolson, BDF, ADI. It marches through hundreds of small steps. The 6,000-cell surface takes three minutes. The trader waits.

Imagine a pricing engine that computes the answer at T directly — one contour integral, one LU factorization reused across every strike in the same vol bucket. The 6,000-cell surface takes two seconds. The trader doesn’t wait.

Power grid

The transient that crashed the model.

A grid operator simulating electromagnetic transients on a 10,000-bus network. Each EMTP time step is a linear solve. The simulation marches through thousands of steps at microsecond resolution. A transformer switching event changes the stiffness by four orders of magnitude. The time-stepper chokes.

SolvLRDE evaluates the settled state directly. No marching. Cost is independent of stiffness. The same solver handles gentle dynamics and the worst switching transient.

Defense radar V&V

The 10,000-cell campaign.

A phased-array antenna simulation team runs V&V across scan angle × frequency × element calibration. Thousands of cells. Each is a time-stepper marching from t=0 to the settled state. The campaign takes 45 minutes of single-core scipy. SolvLRDE does it in under 90 seconds with 4–6 orders of magnitude better precision.