Using an NVIDIA DGX-2 system running accelerated Python libraries, which uses NVIDIA CUDA-X AI software along with NVIDIA RAPIDS and Numba machine learning software, NVIDIA just broke the previous benchmarks of a key algorithm used by hedge funds to backtest trading strategies.
Backtesting is a key step in a trading algorithm’s journey from development to deployment. Once an algorithm is developed and tested, usually on a very small subset of data, it’s ready for more robust backtesting and tuning on vast volumes of historical data.
The NVIDIA AI system used in the test delivered more than 6,000x acceleration, a speedup that has massive implications across the financial services industry. Hedge funds can now backtest 20 million trading simulations in an hour. That’s 6,000x faster than the previously set benchmark of 3,200 an hour.
The results, made public today, have been validated by the Securities Technology Analysis Center (STAC), whose membership includes more than 390 of the world’s leading banks, hedge funds, and financial services technology companies.
“Hedge funds — there are more than 10,000 of them — will be able to design more sophisticated models, stress test them harder, and still backtest them in just hours instead of days,” said John Ashley, NVIDIA’s Director of Global Financial Services Strategy. “Quants, data scientists and traders will be able to build smarter algorithms, get them into production more quickly and save millions on hardware.”
Highlights from the STAC-A3 SWEEP benchmarks included:
- Performed 6,250 times as many simulations as the previous record
- In less than 6 minutes, ran 10,000 simulations on a basket of 48 instruments
- Time to run 10,000 simulations on a basket of 48 instruments was effectively the same as the time to run 1,000
“The ability to run many simulations on a given set of historical data is often important to trading and investment firms,” said Michel Debiche, a former Wall Street quantitative analyst who is now STAC’s director of analytics research. “Exploring more combinations of parameters in an algorithm can lead to more optimized models and thus more profitable strategies.”
Feature image credit: Lorenzo Cafaro