Public results of STAC-A2 on Intel Haswell-based systems

September 8, 2014

Intel has commissioned audited STAC-A2 tests of systems using the Intel Xeon Processor E5-2600 v3 series, code named "Haswell EP". The reports are linked below. Qualified STAC subscribers may also access the source code and binaries used in these tests, as well as micro-detailed configuration information for the systems tested.

1. A white box with 2 x Intel Xeon E5-2699 v3 @ 2.30 GHz (Haswell EP) CPUs and 1 x Intel Xeon Phi 7120A Co-Processor (Knights Corner) card. This was the fastest of any system published to date, in warm runs of the end-to-end Greeks benchmark (STAC-A2.ß2.GREEKS.TIME.WARM). In the same benchmark, this system was 22% faster than a system with 2 CPUs plus 2 GPUs (SUT ID NVDA131118). It also had 46% higher asset capacity (STAC-A2.ß2.GREEKS.MAX_ASSETS) and 53% higher paths capacity (STAC-A2.ß2.GREEKS.MAX_PATHS) than the GPU-based system.

2. A white box with 2 x Intel Xeon E5-2699 v3 @ 2.30 GHz (Haswell EP) CPUs (and no Phi co-processor). This system was 30% faster in STAC-A2.ß2.GREEKS.TIME.WARM than an equivalent system with 2 Ivy Bridge processors running the same implementation source code. And this 2-CPU system was only 12% slower than the 2 CPU/2 GPU system mentioned above, while demonstrating a 46% higher asset capacity.

STAC-A2 is the user-developed benchmark standard based on financial market risk analysis. Developed 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 is able to handle the workload (Monte Carlo estimation of Heston-based Greeks for a path-dependent, multi-asset option with early exercise).

These tests follow on the heels of tests we performed on Intel Ivy Bridge servers in the spring of 2014.