The benchmark software most commonly used to test engineering workstations has an update that improves its ability to test computers running more than 64 logical processors. This will be of value to those testing the high-end performance required for such applications as VR/AR development, artificial intelligence and machine learning programming, and digital twin modeling in very large projects.
The SPECworkstation benchmark is a product of the Standard Performance Evaluation Corporation (SPEC), a non-profit corporation that creates, maintains, and promotes a set of standardized benchmarking tools for testing computer systems.
SPECworkstation 3.0.4 has new multi-threading code allowing it to test computational workloads that scale beyond 64 logical processors. It supports the Microsoft feature called Processor Groups. Microsoft defines processor groups as “a static set of up to 64 logical processors that is treated as a single scheduling entity.”
SPECwpc chair Alex Shows says the update allows SPECworkstation to record more accurate scores on workstations using Processor Groups by accounting for “improved partitioning and affinitization. Processor affinity enables the binding and unbinding of a process or a thread to a CPU or a range of CPUs, so that the process or thread will execute only on the designated CPU or CPUs.”
More information on the update is available in an article published by SPEC: https://www.linkedin.com/pulse/new-specworkstation-version-304-enables-scaling-beyond-bob-cramblitt/?trackingId=kK28iTySRGeslS86rnU4ng%3D%3D