Dell Technologies HPC Community Event
About the Event
High performance computing systems have many elements, all of which need to perform well and in balance, but actual solution of mathematical computations is achieved in silicon: in the processors and accelerators at heart of HPC systems. We now measure performance in teraflops, petaflops, and even exaflops, and we now optimize the performance for calculations ranging from 64-bit floating point operations in simulations down to 8-bit integer operations in high throughput inferencing. What does it take to achieve this computational performance and how do we build and connect processors and accelerators to power the expanding scale and diversity of HPC workloads? How do these computation elements connect to each other, memory, storage and networks? Garima Kochhar, HPC expert and distinguished engineer in the CTIO office at Dell Technologies, will give an accelerated ;-) introduction to processors and accelerators, including the emerging technologies for connecting them, the trends in performance and power, and more.
About the Speaker
Garima Kochhar, Distinguished Engineer, Dell Technologies
Garima Kochhar is a Distinguished Engineer working on High-Performance Computing systems at Dell Technologies. She has worked on several areas of HPC including system architectures, compute performance tuning, high-speed interconnects, parallel file systems and storage, and cluster deployment and management techniques. Her current focus areas include processor architecture and system-level performance optimization. Her responsibilities include technology analysis, solution design, and performance analysis. She is part of Dell’s Integrated Solutions and Ecosystems CTIO team.