First look at 3rd Generation Intel Xeon Scalable processor performance with Abaqus

TotalCAE has been running some benchmarks with the latest 3rd generation Intel Xeon Scalable 8380 processors (Ice Lake) to compare performance to the 2nd generation Intel Xeon Scalable 6226R processors (Cascade Lake Refresh).

All tests used 32 CPU cores (16 cores per socket) along with Abaqus 2021 and its new Hybrid Message Parallel Feature

Abaqus has a standard list of classical benchmarks such as s4b, and e1 that are a bit small for today’s powerful hardware, but we used these as they are publicly available and well known.

s4b – 32 CPU – Abaqus 2021

The s4b benchmark is a nonlinear static analysis that simulates bolting a cylinder head onto an engine block. s4b is a 5 million Degree of Freedom (DOF) model. This was run on 32 cores, 16 cores per socket on both systems.

Intel 2nd Generation Intel Xeon Scalable processors: 309 Seconds

Intel 3rd Generation Intel Xeon Scalable processors: 203 Seconds

This is a 34% improvement for the new Intel 3rd Generation Intel Xeon Scalable processors.

e1 -32 CPU- Abaqus 2021

This benchmark consists of a passenger car impacting a rigid wall at 25 mph. This was run on 32 cores, 16 cores per socket on both systems.

Intel 2nd Generation Intel Xeon Scalable processors: 480 Seconds

Intel 3rd Generation Intel Xeon Scalable processors: 342 Seconds

This is a 28% improvement for the new Intel 3rd Generation Intel Xeon Scalable processors.

Real World Model – 32 CPU – Abaqus 2021

This test is real-world model and not a synthetic benchmark. The data is below

Intel 2nd Generation Intel Xeon Scalable processors: 27747 Seconds

Intel 3rd Generation Intel Xeon Scalable processors: 20759 Seconds

Nearly a 25% increase with the 3rd generation Intel Xeon Scalable processor based system, a nice performance boost.

TotalCAE looks forward to deploying the new Intel 3rd Generation Intel Xeon Scalable processors to clients through our TotalCAE Managed HPC Cluster and Cloud offerings.