We learn from our alumni in this interaction series, often technically, sometimes semi-technically.

Madhur Amilkanthwar NVIDIA, Pune

Madhur Amilkanthwar is a Senior software engineer at NVIDIA, Pune. He graduated from IIT Madras with an M.S. (By Research) degree advised by Dr. Shankar Balachadran. He joined NVIDIA after graduating from IIT Madras in the PTX compiler group where he contributed to several key features and drove a complicated set of components meant for customers. He also worked with the Nervana group at Intel as a key compiler engineer to drive core compiler development for the Machine Learning ASIC Nervana was building for training workloads. He made significant contributions to the overall compiler and also received the important Department Recognition Award. Further, he worked with AMD with the RoCM compiler group to mature and lead key compiler support for AMD's GPUs. The contributions were aimed towards facilitating AMD's compiler to be adopted suitably for performance objectives. He recently started his next stint at NVIDIA again and is with the Grace Superchip compiler team. He has spent most of his work life in and around LLVM infrastructure. Madhur has published papers in ICS, GPGPU, and PPoPP. His work on uncoalasced access pattern detection received a silver medal in the student research competition held with ICS. He has served as a program committee member on the LLVM-HPC workshop (Co-located with SC) in 2022 and 2023.

The compiler is a complicated piece of software and historically it has been compared with dragons. In this talk, I will give an overview of the compiler, the current popular compiler in the world today, and the adoption of compilers in the industry. I will share my experiences and relate it today to why compilers are still relevant today. The talk aims to give enough perspective to new graduates about compilers and convince them why compilers offer interesting core computer science problems.

Organizers