The 2022 SPECtacular Award Winners

For several years, Simon Eismann has made significant contributions to the SPEC Research Group’s (RG) Cloud Working Group and to the RG DevOps Performance Working Group. He led the study of performance of microservices and stability for serverless applications. He also led the effort to create a SPEC RG Technical Report reviewing serverless use cases, which led to the publication of related peer-reviewed papers. Simon has been an active participant in the organization of events that benefit the SPEC research community, served as an evaluator of artifacts, and contributed the TeaStore microservice reference application to the SPEC Research Tool Repository.

Vojětch Horký has done excellent work in his more than six years as the SPEC Research Group’s (RG) Release Manager. He manages not only tool submissions and the SPEC RG Tool Repository, but also the entire SPEC RG website. In the past, Vojtěch led the migration of the SPEC RG website from TYPO3 to its new platform, contributed to the revision of SPEC RG membership information and FAQ, migrated and consolidated all ICPE web sites to the SPEC server, and represented RG on the Editorial Committee.

David Schmidt led the Virtualization Committee in the creation of the SPECvirt Datacenter 2021 benchmark. His leadership as Chair kept the committee’s work focused to deliver the vision for this new benchmark. David’s low-key yet energetic approach created a positive benchmark development environment and led the team on a very successful path to releasing the benchmark. In addition to chairing, David virtually wrote the Run and Reporting Rules and provided almost-daily test feedback.

International Conference on Performance Engineering (ICPE) is the flagship conference for SPEC and a very important conference for the performance evaluation and engineering world at large. The CORE ranking list is used by researchers and administrators in many areas of the world to select top conferences. Thus, having ICPE listed in this rank is important for the continued success of the conference, and for the publication of work done in performance evaluation around the world.

Hiroshi Ono contributed significant effort to the necessary benchmark candidate performance evaluation and scaling experiments for the SPEChpc 2021 Benchmark Suites. When further investigation was needed, he deepened his analysis and produced and interpreted hardware performance counter measurements to improve characterization of the benchmark candidates. The result of Hiroshi’s investigation were very important for the selection of the benchmarks candidates. His analysis not only guided how much time the committee decided to allocate to certain candidates, but also identified unknown problems to be solved before launch.

One of the important benchmark candidates for the SPEChpc 2021 Benchmark Suites, PiconGPU, is a challenging and sophisticated C++ application that stresses compilers and a software tool’s ability to support it. Just getting a successful compile with LLVM, a broadly used compiler, took a heroic effort. Jeffrey Kelling, however, persisted and was then able to provide valuable input to many SPEC members struggling to compile the code on their own compiler.

Junjie Li performed extensive experimental evaluation of the SPEChpc 2021 Benchmark Suites candidates in large-scale and diverse computational facilities. The data collected and the performance data he provided was instrumental in the continued development of the benchmark candidates, the assessment of the candidates by SPEC members, and the selection of the benchmark candidates. Junjie provided valuable access and computing time at two different datacenters. These computing resources were very valuable to the HPG committee as they increased the diversity of the platforms available for the evaluation of the benchmarks.

Mat Colgrove is the project lead for the SPEChpc 2021 Benchmark Suites. In this position, he dedicated extensive time and effort to benchmark development. He also provided leadership and hands-on work to create the many kits that were necessary during the benchmark development process. Consistent and patient, Mat showed excellent leadership for the three years this benchmark was in development, enabling and encouraging the team of volunteers and leading by example.

Chris Floyd has distinguished himself as the principal benchmark developer for the SPECvirt Datacenter 2021 benchmark. He was responsible for overseeing the adaptation of the CloudPerf test framework to the SPECvirt Datacenter 2021 harness. He was also key in designing the benchmark’s synthetic departmental workloads and porting HammerDB. Chris developed many of the necessary control and monitoring scripts needed for the benchmark. Using his deep understanding of the low-level details of the benchmark’s architecture, he authored the benchmark’s design guide and tested all the kit builds of the benchmark on VMware.

Lisa Roderick significantly contributed to the successful launch of the SPECvirt Datacenter 2021 benchmark as leads for both documentation and the benchmark’s VMware SDK. She championed the migration of the documentation to use the reStructuredText format, thus ensuring formatting consistency across all of the documentation. Lisa authored the SPECvirt Datacenter 2021 benchmark user guide, FAQ and technical support documents. As a development lead, she adapted the SDK API control scripts from VMware to conform to the requirements of the SPECvirt Datacenter 2021 harness and tested all of the benchmark kits on VMware.