STAC Report: STAC-A2 (derivatives risk) with and without AMD’s Encrypted State

Hardware encryption had 0-1.3% impact

4 August 2021

STAC recently performed STAC-A2 Benchmarks on two AMD-based solutions in an A/B comparison in which the only difference between solutions was whether AMD’s Secure Encrypted Virtualization – Encrypted State (SEV-ES) was enabled.

The SUTs consisted of the STAC-A2 Pack for C (Naive Implementation) Rev B running in a virtualized environment on a Dell PowerEdge R6525 server that had 2 x 8-core AMD EPYC 72F3 CPUs and 2 TiB of physical DDR4 memory. The bare-metal hypervisor VMware ESXi 7.0 Update 2 supported a single VM that had access to all of the machine's cores and 1.5 TiB of its memory, running an AMD-modified version of the SUSE Linux Enterprise Server 15 SP2 operating system. The servers had patches applied to mitigate Spectre & Meltdown security vulnerabilities.

The STAC Report on the SUT with SEV-ES enabled contains the comparisons in its Summary. The STAC Report on the SUT with SEV-ES disabled is available at the same web page.

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

AMD wished to highlight several results from this comparison:

• No change in the maximum paths or maximum assets handled
  (STAC-A2.β2.GREEKS.{MAX_PATHS/MAX_ASSETS})
• A 0.0% increase (rounded to the nearest tenth of a percent) in elapsed time for warm runs of the large Greeks benchmark
  (STAC-A2.β2.GREEKS.10-100k-1260.TIME.WARM)
• Less than 1.2% increase in elapsed time for warm and cold runs of the baseline Greeks benchmark
  (STAC-A2.β2.GREEKS.TIME.{WARM/COLD})
• Less than 1.3% reduction in throughput, energy efficiency, and space efficiency
  (STAC-A2.β2.HPORTFOLIO.{SPEED/ENERGY_EFF/SPACE_EFF})
• No change in quality benchmark results

These tests used a STAC-authored “naive” STAC Pack, which contains no hardware-specific optimizations or proprietary libraries. See the report at the link above for details. Premium subscribers have access to the source code as well as the micro-detailed configuration information for the solutions. To learn about subscription options, please contact us.