July 2026
Intermediate to advanced
536 pages
10h 39m
English
Quantitative trading research is computationally intensive. Factor research, portfolio optimization, and risk modeling require processing millions of rows of market data, fitting models across large cross-sections of assets, and solving complex optimization problems. These workflows are bottlenecked by CPU-bound operations that dominate most quant research codebases.
Graphics processing units (GPUs) solve this by executing thousands of parallel operations simultaneously. Unlike CPUs, which have 8-20 cores optimized for sequential tasks, modern GPUs have thousands of cores designed for data-parallel workloads. This architecture makes GPUs ideal for the vectorized DataFrame operations, matrix ...
Read now
Unlock full access