Post by Asymmetry Computing

Every index fund has to re-solve a massive optimization problem every time its benchmark changes. The tool everyone reaches for — Gurobi — is accurate, but it's built on methods that can't trade accuracy for speed. So at scale, solve time becomes a real cost line. We built PRISM to fix that. Same optimum as Gurobi (weight-RMSE ≤ 0.00076). A median 25× faster on real 2025 S&P 500 reconstitutions, up to 865× at scale. Every solution feasibility-certified — 15/15 stacks verified, 5/5 infeasible cases caught. We're not hiding the edges either: two small-N tiers where we're slower, hard-cardinality problems where exact methods still win. Benchmarks are reproducible and the losses are disclosed. That's on purpose — it's the only way quant teams will actually trust a new solver. If you run rebalancing, transition management, or portfolio optimization at scale, we'd like to run this on your own data. asymmetrycomputing.com #QuantFinance #PortfolioOptimization #GPUComputing #IndexFunds #FinTech #ETF #DirectIndexing