Skip to main content

Isayah Reed Triumphs While Connecting the World

reed

Isayah Reed is a senior technical program manager for HPC/AI benchmarking at Microsoft. Prior to joining Microsoft, Reed spent seven years at Intel as a performance analyst working on HPC hardware and interconnects, where he eventually filed a patent for an automated load-balancing methodology across distributed computation and communication patterns.

Early in his career, Reed spent a lot of time working on activities tangentially related to inclusion. He volunteered in his community doing various things, such as coding sessions at a Girls Club in rural Massachusetts, visiting local universities, participating in career panels, and, in his words, “other generic let’s-fill-the-STEM-pipeline endeavors.” Reed “eventually understood that these things [were] not compatible with [his] personal strengths,” and that “the median software engineer can execute any of the aforementioned volunteer activities just as well as himself.”  

Instead, Reed opted toward finding an outlet aligned with his personal qualities. “There were times when I noticed that I had the skills, experience, and perspective to undividedly contribute,” he said. This revelation provided the motivating factor for his time away from the corporate world doing various HPC projects.

In 2021, Reed took a career break for a few months to move abroad and dedicate more time to things he is passionate about outside of work, including fitness, learning, and traveling to experience new cultures. During that time, he lived in East Africa doing volunteer work in HPC capacity-building. Reed built low-cost, low-powered HPC mini-clusters and created basic HPC learning content for local researchers, which he continues to do today. This experience also expanded Reed’s interest in the prevailing problem of accessibility and inclusivity within HPC.

“While a large entry barrier for today’s aspiring computer scientist and computational scientist is access to specialized equipment, this was not a problem during my undergraduate years studying computer science,” Reed explained. “When I first arrived at the University of Illinois, very few students had coding experience; almost no one owned a laptop; and specialized equipment, such as FPGAs, GPUs, and HPC clusters, were not commonly used by most engineers.”

According to Reed, this allowed everyone to be on approximately equal footing with respect to learning. 

“These days, it is very common for computer science students to begin their university studies after years of coding (by the way, I hate coding),” he said. “But, this is not necessarily true in HPC, where the resource gap is large and possibly growing. 

“I was fortunate to have both a university and research advisor that provided me with several resources to explore what was still a niche specialty at the time. This has motivated me to work on improving access—to equipment, people, resources, etc.—and removing as many barriers as possible for inclusion,” Reed continued. “For example, at the moment, I run a small but powerful HPC cluster from home and allow free compute time for researchers and students in rural areas around the world to use it as a sandbox for learning and debugging. The purpose of my volunteer work is to improve inclusivity for rural areas without access to HPC equipment, as I strongly believe there should be as few barriers as possible for joining, participating in, and contributing to HPC, as well as the wider fields of science and engineering.”

Lowering the Entry Barrier to HPC

For his I Am HPC profile, Isayah Reed further shares how he became part of this community and how he envisions its future.

Isayah Reed

Senior Technical Program Manager, HPC/AI Benchmarking, Microsoft

Q: What single event most made you realize you wanted a career in HPC/computing?

Reed: During my college years, I was looking for a research advisor in computer architecture and was offered a project that looked into heterogeneous computing with GPUs. This was well before GPUs were a “thing.” Few people knew how to use them. I was fascinated by the problem statements and different trade-offs that come from offloading certain tasks to different computing units. My advisor, Maria Garzaran, recommended some HPC courses; I fell in love with the field; and I never looked back.

Q: What do you consider your biggest contribution to the HPC/computing community?

Reed: Outside of my day job, I try to focus my efforts on expanding accessibility and lowering the entry barrier to HPC. For years I’ve run a mini HPC cluster at home, to which I allow free access for several students, researchers, and enthusiasts in less developed countries. I try to help them learn and practice HPC or perform HPC research. I also donate my older equipment to various research teams and universities in these countries. Things like mid-tier GPUs and interconnect switches are typically inaccessible to people outside of western countries due to challenges such as price, maintenance, or availability. I strongly believe anyone should be able to enter HPC without needing access to an expensive government-funded or university supercomputer or spending a ton of money for time on a cloud provider.

Q: In the past 35 years, what is the most significant overlooked breakthrough that has impacted the field in your eyes?

Reed: Introduction of the dual core processor in the early 2000s. It completely changed the approach to software design and started the ascendance of parallel programming as a specialty. Even today, many applications have unnecessarily lower performance due to not making efficient use of available cores.

Q: What would you like to see change about, within, or among the HPC/computing community?

Reed: Tearing down walls between the different HPC sub-groups. We are missing opportunities to solve certain problems by failing to collaborate across different fields of study. I’m seeing, for example, that the AI people all know each other, while the fluid dynamics people also know each other. But, I don’t see much cross-collaboration between the two groups. There are some HPC computing issues that other sub-disciplines unnecessarily struggle with—issues that I, myself, could easily resolve in a few seconds. Yet, they rarely reach out for collaboration.

Who Is Your HPC Hero?

Send your suggestions to the SC23 Communications Team. Share their name and a brief description of why they’re your HPC hero.

Stay Up to Date

Sign up to receive the SC newsletter in your inbox.

Information provided is treated in accordance with ACM & IEEE privacy policies.

Back To Top Button