Technical Requirements | Options Analysis Suite

Last reviewed: May 17, 2026 by Options Analysis Suite Research.

Technical Requirements

Browser Requirements

• Chrome: Version 113+ (Recommended for WebGPU)
• Edge: Version 113+ (WebGPU support)
• Firefox: Version 88+ (WebGPU support in development)
• Safari: Version 18+ (WebGPU in Technology Preview)

System Requirements

• Minimum RAM: 4GB
• Recommended RAM: 8GB or more
• Processor: Dual-core 2.0GHz or faster
• Display: 1920x1080 or higher recommended

GPU Acceleration

WebGPU acceleration is production-ready for Monte Carlo simulations and model calibration, delivering significant performance improvements for path-dependent option pricing, complex Greeks calculations, and parameter fitting.

• Supported Browsers: Chrome 113+ and Edge 113+ with WebGPU enabled
• Monte Carlo: GPU-accelerated path generation for pricing and Greeks (10,000+ paths simultaneously)
• Model Calibration: GPU-accelerated objective function evaluation for Heston, Jump Diffusion (Merton/Kou/Bates), Variance Gamma, and Local Volatility (Dupire eSSVI)
• Hardware: Dedicated GPU recommended (NVIDIA or AMD with 2GB+ VRAM)
• Fallback Chain: Automatic GPU → Web Worker → CPU fallback when WebGPU is unavailable
• WebGPU PDE Solver: Adjoint method computes all Greeks in a single forward+backward pass

WebGPU Model Calibration

GPU-accelerated calibration uses WGSL compute shaders to evaluate objective functions in parallel, delivering 10-50x speedups over CPU for large option chains.

• Heston: Parallel characteristic function integration for stochastic volatility calibration
• Jump Diffusion: GPU batch pricing for Merton (log-normal jumps), Kou (double-exponential), and Bates (Heston + jumps)
• Variance Gamma: FFT-based characteristic function evaluation with GPU parallelization
• Local Volatility (Dupire eSSVI): Full GPU pipeline with Huber loss, per-expiry normalization, and arbitrage-free surface generation
• CPU Parity: GPU results validated against CPU reference within tolerance gates (objective < 0.1% difference)
• QC Metrics: Butterfly/calendar arbitrage detection, fit quality metrics (RMSE, max error), and parameter diagnostics

Network Requirements

• Bandwidth: 10 Mbps minimum for real-time data
• Latency: Under 100ms to data providers recommended
• WebSocket support: Required for live updates

Why WebGPU Matters Here

WebGPU is the browser-native GPU compute API, and it's what lets a web app run the kind of parallel workloads that used to require a desktop install: Monte Carlo path generation across 10,000+ paths, characteristic-function calibration over wide parameter spaces, and the adjoint method for computing all Greeks in a single PDE pass. On supported browsers with a dedicated GPU, calibration that takes 30 seconds on CPU runs in 1–3 seconds on GPU. On integrated GPUs the speedup is smaller but still meaningful (5–10×). The fallback chain (GPU → Web Worker → CPU) means the platform still works on any reasonably current browser; WebGPU just makes the heavy workflows materially faster.

Mobile and Tablet Compatibility

The platform runs on modern mobile browsers (iOS Safari 18+, Chrome on Android 113+), but the design target is desktop or large tablet. The full Greeks-by-strike heatmaps, the 3D volatility surface viewer, and the multi-pane Strategy Builder are all built for a wide viewport; on phone screens, layouts collapse to single-column with a scrollable chart area, which works for spot-checks and quick reads but not for sustained workflow. If you're doing real research, plan around a 1440×900 or larger display.

What Happens on Older Hardware

Below the recommended specs, the platform still works but with three predictable degradations. First, WebGPU paths will fall back to Web Worker or CPU: Monte Carlo pricing and calibration will be slower, often by 5–20×. Second, very large 3D surfaces (200+ strikes × 8+ expirations) may stutter on integrated graphics. Third, browser-side memory pressure can build up across long sessions; a periodic refresh keeps the cache sane. None of these are fatal; the analytical results are identical, just slower or less smooth. But if you're choosing hardware, the recommended specs are calibrated to what makes the workflows feel responsive.

This page is part of the Options Analysis Suite documentation hub. Browse the glossary for term definitions.