SANTA CLARA, CALIFORNIA
The advent of graphics processing units (GPU) is helping the financial industry keep pace with regulations. Following the financial crisis that hit bottom in 2008, the government unloaded a raft of regulations on financial institutions that overloaded central processing units (CPU). Whereas banks used to run a single valuation adjustment daily, they were now required to regularly update multiple, large-scale, complex simulations on billions of quickly-changing data points. GPUs were first used for sharper, faster 3D game renderings. They are capable of parallel-processing iterative functions sometimes a hundred times faster than CPUs. Some banks now employ thousands of GPUs because quants want to develop algorithms with competitive edges instead of fighting computer issues. GPUs’ capacity for deep learning helps algorithmic traders identify new strategies, asset managers optimize portfolios, and investment bankers develop more accurate Monte Carlo simulations.