Kumudha Narasimhan Engineering Manager (Research) @ Codeplay Software

My area of interest is Compilers and High performance Computing. Primarily I focus on optimizing compiler and middleware stack for DNNs and AI to improve performance and portability.

Portfolio

Publications


  • A performance analysis of leading many-core technologies for Cellular Automata execution
    Alessio De Rango, Donato D'Ambrosio, Alfonso Senatore, Giuseppe Mendicino, Kumudha Narasimhan, Mehdi Goli, Rod Burns
    International Workshop on Algorithms, Models and Tools for Parallel Computing on Heterogeneous Platforms (HeteroPar) @ Euro-Par, August 2023 [pdf]
  • Accelerating Neural Networks using Open Standard Software on RISC-V
    Kumudha Narasimhan, Mehdi Goli
    International workshop on RISC-V for HPC @ ISC 2023 (RISC-V HPC), May 2023 [pdf], [slides]
  • User-Driven Online Kernel Fusion for SYCL
    Victor Perez, Lukas Sommer, Victor Lomüller, Kumudha Narasimhan, Mehdi Goli
    ACM Transactions on Architecture and Code Optimization (TACO), Nov 2022 [pdf]
  • Toward Performance Portability of AI Graphs Using SYCL
    Kumudha Narasimhan, Ouadie El Farouki, Mehdi Goli, Muhammad Tanvir, Svetlozar Georgiev, Isaac Ault
    Performance, Portability, and Productivity in HPC (P3HPC) @ Supercomputing (SC), Nov 2022 [pdf]
  • Towards performance portability of AI models using SYCL-DNN
    Muhammad Tanvir, Kumudha Narasimhan, Mehdi Goli, Ouadie El Farouki, Svetlozar Georgiev, Isaac Ault
    International Workshop on OpenCL (IWOCL), May 2022 [pdf]
  • Improving Performance of SYCL Applications on CPU Architectures using LLVM-directed Compilation Flow
    Pietro Ghiglio, Uwe Dolinsky, Mehdi Goli, Kumudha Naraimshan
    Programming Models and Applications for Multicores and Manycores (PMAM) @ PPoPP, Apr 2022 [pdf], [slides]
  • A Practical Tile Size Selection Model for Affine Loop Nests
    Kumudha Naraimshan, Aravind Acharya, Abhinav Baid, Uday Bondhugula
    ACM International Conference on Supercomputing (ICS), Jun 2021 [pdf], [slides], talk ([short], [long])
  • Towards Cross-Platform Performance Portability of DNN Models using SYCL
    Mehdi Goli, Kumudha Naraimshan, Ruyman Reyes et. al.
    Performance, Portability, and Productivity in HPC (P3HPC) @ Supercomputing (SC), Nov 2020 [pdf]
  • Optimizing Dense Matrix Computations with PolyMage
    Kumudha KN
    Masters Thesis, July 2018 [pdf], [slides]
  • Optimizing Geometric Multigrid Method Computation using a DSL Approach
    Vinay Vasista, Kumudha Naraimshan, Siddharth Bhat, Uday Bondhugula
    Supercomputing (SC), Nov 2017 [pdf]

Talks


  • A SYCL Extension for User-Driven Online Kernel Fusion
    International Workshop on OpenCL (IWOCL), Apr 2023 [link]
  • The oneAPI Software Abstraction for Accelerated Computing
    Panel at Supercomputing (SC), Nov 2022 [link]
  • The oneAPI Software Abstraction for Heterogeneous Computing
    oneAPI Developer Summit at Supercomputing (SC), Nov 2021 [link]
  • SYCL Performance and Portability
    Excalibur-SLE, Exascale HPC for System Level Engineering, Jul 2020 [link]
    SYCL Summer Sessions, Aug 2020 [link]

Professional Service



About

Past

I completed my Masters by Research degree from CSA, Indian Institute of Science working with Prof. Uday Reddy

I completed my Bachelors with Gold medal hons. in CS from M. S. Ramaiah Institute Of Technology

I worked at Goldman Sachs as an Analyst in the Technology division for three years (2012-2015)

Hobbies (click to expand)

I learn Bharatnatyam (Indian classical dance) and try to give live performances once in a while.

In my spare time I read crime fiction novels. One of my favorites is Prisoner of Birth by Jeffrey Archer
Also, like listening to classical Indian music - Yesudas is my favorite singer

I also love to try new Mehandi designs. Reduces my stress and keeps me calm



More details about my work is in my Resume

Contact Me

In case you have any questions for me.    Please fill out the below form.

I will try to get back to you within 24 hours.