Developer Stories

In this lively discussion, HPC Dan, a professor emeritus from the University of Utah and a notable voice in high-performance computing, shares insights from his impressive journey across academia and Microsoft. He delves into the challenges facing resource allocation in HPC amid an AI boom, including supply chain issues with GPUs. Dan emphasizes the need for collaboration among HPC, cloud services, and government for innovation and investment in the future. He also discusses the vital role of education in developing a skilled workforce and the importance of mentorship in scientific advancement.

Today we talk to Michela Taufer, Dongarra Professor of High Performance Computing at the University of Tennessee Knoxville. Michela leads the Global Computing Laboratory, and you’ve probably seen her on the stage of ISC and several high performance computing events! It was great to talk to Michela to get her perspective on leading a laboratory, democratizing access to science, and interesting work going on in our community.

As we hit the five year mark for the Developer Stories Podcast, we want to pause to celebrate a milestone that not many podcasts achieve - the 100th episode! 🎉 To celebrate, we have an immensely special guest - Andrew Jones from Microsoft, a leader in the high performance computing community that is “currently planning the future of HPC & AI at Microsoft Azure” and writes about trends in his blog and is a prominent voice in our community. If you are interested in the future of HPC, including strategy, culture, history, and the technology itself, and especially in the context of the cloud and converged computing, I know you will be interested to listen to this episode! We especially have some fun at the end as we play a game, weaving what the world might look like with our AI future… 😆️ From this passionate author to you, Happy 100 episodes, and enjoy!

Today we chat with Wileam Phan, a performance engineer that first encountered high performance computing in Indonesia during his senior year, and has followed his interest through today and working on the performance tool HPCToolkit. In this episode, we talk about different strategies for performance analysis, the Exascale computing project, culture in the United States, and a little about open source.

Jakob Luettgau is a researcher at the French Institute for Research in Computer Science and Automation at Inria, with an interest in storage, workflows, and machine learning, amongst other things. In this episode, we talk about different patterns of I/O, storage for cloud and high performance computing, and more general ideas about the convergence between the spaces. We hope that you enjoy the episode, and happy summer listeners! 🌞️

I recently spoke with Jay Lofstead about the challenges of starting a career in HPC, and this morning listened to a panel on “Emerging Talent in HPC.” I found the panel was missing a valuable perspective, and wanted to make a quick recording of that perspective. This isn’t an official Developer Stories episode, but rather some associated ramblings of a dinosaur, and ideas or thoughts that I think are useful to say.

Jay Lofstead is a Principal Member of the Technical Staff at Sandia National Laboratories, with extensive experience in storage technologies, High Performance Computing, and reproducibile practices. In this episode, we talk about the challenges of navigating the graduate school and career space, because the decisions you make can be very strategic, and they often need to be customized on the level of the individual. You also might be taken with Jay’s love of ice cream, how he chooses a flavor, and how he has created community around HPC and getting ice cream. Here’s a cool hint - “Try the gray stuff, it’s delicious!” And if you are a member of the HPC community, you are invited to come too. This episode is immensely useful for early career folks looking to make a next step, mid- to later- career folks that want to help, and anyone that likes to eat ice cream.

Today on Developer Stories I talk with Felix LeClair, an open source developer and BLAS Floating-point guru that works at Environment and Climate Change Canada. Felix is someone to watch - I have never met someone with as much passion about hardware and chip design as him. In this episode, we talk briefly about the history of HPC Social, the growing place to talk with your friends that are also HPC practitioners, and then delve into chips and precision. Is science in trouble with the move toward lower precision chips for the AI/ML use case? What is utilization? Why are we talking about feeding the beast? Felix has some interesting insights on these ideas that you might want to hear. This episode was immensely fun to record, Felix is a fanastic human, and I hope that you enjoy listening!

I’m thrilled to be interviewing Alan Sill, Managing Director, of the High Performance Computing Center at Texas Tech University, President of the Open Grid Forum, Co-Director of the National Science Foundation Cloud and Autonomic Computing Center, and my co-founder for HPC Social. Alan is someone that I could sit down with and listen to stories for hours, and that’s reflected in this episode - we talk for over an hour on his start in Physics to exploration of high performance computing and impact that “still doesn’t get him a discount at Starbucks.” This is a great episode to capture a community member with both deep impact and broad experience, and for wisdom about leadership through technical excellence and social components. I hope that you enjoy it!

Today we talk to Ramiro Berrelleza, Founder and CEO of Okteto about his story of falling in love with computers as a kid, and pursuring a non-traditional path that has led him from industry to consulting to leading companies. We have a lot of fun talking about web technologies, and our shared experiences. This episode is full of great stories that range from discovering the weather in different parts of the United States to finding belonging in a new place.