OCI sets records in STAC benchmarks for financial risk analysis

February 12, 2024 | 4 minute read
Text Size 100%:

In this blog post, we discuss the performance measured at Oracle Cloud Infrastructure (OCI) with the industry standard STAC-A2 benchmark. The system was independently audited by the Strategic Technology Analysis Center (STAC). STAC and all STAC names are trademarks or registered trademarks of the Strategic Technology Analysis Center.

Background

Financial services institutions are facing the daunting task of balancing traditional market pressures, such as regulatory compliance and uncertain markets, with emerging challenges, such as new competitors and evolving customer expectations. These institutions require computational platforms that process large datasets and offer configurable solutions, freeing their top talent to focus on innovation and customer-oriented solutions. OCI is uniquely positioned to accelerate these high-performance computing (HPC) workloads.

Until recently, HPC solutions haven’t effectively scaled dynamically on demand, with limiting factors such as cost-efficiency and reliable, flexible high-performance platforms. Financial services institutions are now experiencing a breakthrough with the new generation of HPC, capable of powering demanding applications like the Monte Carlo Method (MCM) on demand at economically viable costs. A critical tool in financial analysis, MCM enables institutions to simulate millions of scenarios, stress-test portfolios, and manage a broad spectrum of financial risks more effectively.

STAC-A2 is the technology benchmark standard based on financial market risk analysis and was designed by quants and technologists from some of the world’s largest banks. STAC-A2 reports the performance, scaling, quality, and resource efficiency of any technology stack that can handle the workload. The benchmark is a Monte Carlo estimation of Heston-based Greeks for a path-dependent, multiasset option with early exercise. The workload can be a proxy extended to price discovery, market risk calculations, such as sensitivity Greeks, profit and loss, and value at risk (VaR) in market risk. and counterparty credit risk (CCR) workloads, such as credit valuation adjustment (CVA), and margins that financial institutions calculate for trading and risk management.

Benchmarks 

STAC performed STAC-A2 Benchmark tests at OCI using an eight-node cluster using an HPC Terraform template available on the Oracle Cloud Marketplace. The cluster consists of compute nodes based on the OCI HPC-optimized BM.Optimized3.36 bare metal servers, with dual 3rd generation Intel Xeon, formerly codenamed “Ice Lake”, 18-core processors, 512-GB memory, and 3.2-TB NVMe with a high-speed 100-Gbps network. These STAC-A2 tests were performed using the Intel-authored STAC-A2 Pack for Intel oneAPI (Rev N), which used the Intel oneAPI 2022.3 Base and HPC Toolkits, that includes the Intel MPI and oneTBB libraries. In all, the STAC-A2 specifications delivered over 200 test results, which are summarized in the STAC Report.

The report includes the following notable results:

  • Compared to all publicly reported, nonaccelerated cloud solutions to date, this solution set the following records:
    • The best price-performance: 19,788 options per USD over one hour and three days of continuous use (STAC-A2.β2.HPORTFOLIO.PRICE_PERF.[BURST | PERIODIC])
    • The highest throughput: 119.2 options per second (STAC-A2.β2.HPORTFOLIO.SPEED)
    • The fastest warm time (24.3 ms) in the baseline Greeks benchmark (STAC-A2.β2.GREEKS.TIME.WARM)
    • The fastest cold (2.98 s) and warm (1.96 s) times in the large Greeks benchmarks (STAC-A2.β2.GREEKS.10-100K.1260.TIME.[COLD | WARM])
    • The most correlated assets (145) and Monte Carlo paths (7,000,000) simulated in 10 minutes (STAC-A2.β2.GREEKS.[ MAX_ASSETS | MAX_PATHS ])
  • Compared to the previously most price-performant, nonaccelerated cloud solution (SUT ID INTC221006b), this solution demonstrated a 1.3-times price-performance advantage in options per USD for one hour and three days of continuous use1
  • Compared to a 10-server nonaccelerated cloud-based solution using the previous generation of CPUs, Intel Xeon Scalable Processor, formerly codenamed Cascade Lake, (SUT ID INTC210331), this eight-server solution demonstrated the following results:
    • Over 1.5times price-performance advantages in options per USD for one hour and three days of continuous use (]STAC-A2.β2.HPORTFOLIO.PRICE_PERF.[BURST | PERIODIC])
    • Over 1.4-times the speed in warm runs of the baseline Greeks benchmark (STAC-A2.β2.GREEKS.TIME.WARM)
    • Over 1.2-times the speed in warm runs of the large Greeks benchmark (STAC-A2.β2.GREEKS.10-100K.1260.TIME.[COLD | WARM])
    • Over 1.1-times the speed in cold runs of the large Greeks benchmark (STAC-A2.β2.GREEKS.10-100K.1260.TIME.[COLD | WARM])

Conclusion

The financial industry’s key concerns have been pricing and risk calculation, which rely heavily on the latest technologies for instantaneous calculations and real-time decision-making. Pricing and risk calculation, algorithmic trading model development, and backtesting need a robust scalable environment with the fastest interconnects and networking.

In a previous blog post, OCI demonstrated the best in class cloud GPU performance for STAC-A2 using the OCI Compute GPU instance, BM.GPU4.8. OCI offers financial services customers elastic and comprehensive HPC capabilities with CPUs, GPUs, ultra-high-speed networking and storage for both MCM-based analyses and emerging artificial intelligence and machine learning-based analyses and analytics. We encourage readers to evaluate their financial services workloads at Oracle Cloud Infrastructure with a 30-day free trial.

 

Martin Feyereisen


Previous Post

Automate secret generation and rotation with OCI Secret Management

FREDERICK BOSCO | 3 min read

Next Post


Announcing the availability of OCI Full Stack Disaster Recovery in OC1 regions worldwide

Greg King | 1 min read