When hardware stops getting faster, software has to get smarter.
Aqua is Aquon's computation engine. It reads the machine it runs on and adapts to it, at runtime, with nothing configured by the caller.
One engine, many machines
The same code, a different route on every machine.
Aqua profiles the hardware when it starts, classifies what it is working with, and routes each operation to the strategy that machine can run fastest. Pick a machine and watch the route change. The caller never writes a line of configuration.
Aqua profiles the machine
Aqua detects constrained resources and selects a conservative parallel strategy.
and routes the work
On a constrained cloud, Aqua routes to conservative parallelism and cache-blocked kernels.
Nothing is configured by the caller. Your code never changes; Aqua re-routes underneath.
Three things have to be true at once.
It decides at runtime.
The machine is discovered when the program starts, not assumed when the code was written.
Aqua also carries a declarative layer: you state what to optimize and the constraints it must respect, rather than how to compute it. That is the surface our patent application covers.
A computation engine for the math at the core of modern finance.
Portfolio construction means solving large, heavily constrained optimization problems quickly and correctly. Aqua is built for exactly that. On Zeton's 10-year production backtest, over half a million solver calls, these are the results.
5–15×
faster
vs our own pre-Aqua implementation, on our own backtest, not a third-party system. Upper end depends on available cores.
1.3x to 1.9x
across general workloads
across 209 tests on five hardware configurations
1 in 1,000,000,000
correctness parity
across 36 of 36 configurations, zero regressions
Current application · convex optimization
Proven where the math is hardest.
Portfolio construction is a constrained optimization problem, solved over and over. A ten-year weekly walk-forward backtest makes more than half a million solver calls in a single run.
- Same formulation
- Same solvers and estimators
- Same risk model
The same solvers, routed to the path the machine in front of you runs fastest, and run substantially faster because of it.
Aqua does not change that math. On our own production backtest that is 5 to 15 times faster than the implementation it replaced, measured across 36 configurations, with every financial metric matching the reference to within one part in a billion and zero correctness regressions.
Measured against our own pre-Aqua implementation on our own backtest, not against a third-party system. The upper end of the range is dependent on available cores.