This Rock Stars of HPC series is about the men and women who are changing the way the HPC community develops, deploys, and operates the supercomputers and social and economic impact of their discoveries.
Karan Batta, Senior Program Manager for Azure HPC & Big Compute at Microsoft
From software developer at a small start-up in New Zealand, to Senior Program Manager at one of the largest multinational technology companies in the US, Karan Batta has led a career touched by HPC – even if he didn’t always realize it at the time. As the driving force behind the GPU Infrastructure vision, roadmap and deployment in Microsoft Azure, Karan Batta is one of the Rock Stars of HPC.
insideHPC: Have you always had a passion for HPC?
Karan Batta: Actually I didn’t know what HPC was until I joined GreenButton, a start-up that specialized in moving high performance and process-heavy applications to the cloud. I have, however, always been amazed by the years of research and work some people devote themselves to, and the level of sophistication required for it to be meaningful. For me, HPC meant working on the most cutting-edge technologies, whether that was hardware, software or platform. It also equated to my passion for helping people smarter than I am solve harder problems, from trying to cure cancer to teaching a car to drive itself. My real passion is helping others make use of my technology.
insideHPC: And what’s been the highlight for you in your career?
Batta: Being there when GreenButton was acquired by Microsoft was a true highlight for me. This was back in 2014, when the Azure Cloud was still in its infancy. Even then Microsoft could see that a shift towards the cloud was coming and so began gathering all the right pieces to build a robust platform. Knowing that the company saw the value in GreenButton was tremendous, and we came to see that we’d been working within HPC without realizing it was HPC. We worked on large-scale computing, such as DNA sequencing to rendering animated films.
Karan Batta presenting at SC16
insideHPC: Has your perception changed now you’ve got to know HPC?
Batta: I don’t think it has changed much as my initial sense was always that there’s a lot of complexity that goes into HPC. I do think people take HPC to be an interesting term that encompasses anything that requires large-scale compute resources or huge data sets. I’ve always believes that people do HPC without realizing it as it’s moved far beyond the image of managing clusters or grids. Perceptions are changing and HPC is becoming a more generalized concept of a need for launching large-scale compute using on-demand resources.
insideHPC: And what impact is that having on the industry?
Batta: Even going back to the early days of GreenButton, we’d talk about the fact we wanted to have the little guy be able to compete with the big guys – that’s the impact we wanted to see, and that’s what’s starting to happen. A team of 10 people working in a garage to make an animated movie can start to compete with the Disney and Pixars of the world because they can have a supercomputer with the click of a button. HPC is going to be more accessible and widely available, and the impact of that will be an increase in competition out in the world. Honestly, no one goes out and buys a data center when they can have HPC resources in their back pocket. That’s a huge deal and I can see it happening right now.
insideHPC: You’ll be sharing your views at the upcoming GPU Technology Conference 2017. Other than your talk on HPC and Deep Learning on Azure, what are you most looking forward to?
Batta: Last year, GTC focused on advances like VR and autonomous cars, and I feel like this year will be a leap frog jump in that direction. I don’t think we should just expect to see hardware innovations; I think there will be wide acceptance of AI software from quite diverse industries. Nvidia will surely have some interesting hardware announcements up its sleeves, and I’m excited to see how far things are going on the automotive side too.
—
Karan Batta is Senior Program Manager for Azure HPC & Big Compute at Microsoft. In addition to working on high scale services & GPU Infrastructure in Azure, he also leads accelerated workloads with a focus in large-scale rendering and visual effects with customers like Jellyfish Pictures, Autodesk and others.
At GTC 2017, Karan will present two talks, first entitled “Accelerated Compute Workloads on Azure” and second titled “High-End Design and Visualizations on Azure”. He will discuss how to scale traditional HPC-based applications or workloads in Azure using NVIDIA Tesla-based GPUs and Azure’s low-latency networking. He will also present case studies of deep learning and AI workloads using GPUs in Azure.
See Karan Batta speak at the GPU Technology Conference in Santa Clara, CA on May 8-11 – Learn More.
