Name: Ashish Panwar
Type: User
Company: Indian Institute of Science
Bio: I am interested in systems research oriented towards maximizing hardware resource utilization -- particularly around operating systems and memory management.
Location: Bangalore, India
Blog: https://apanwariisc.github.io/
Ashish Panwar's Projects
Resource-adaptive cluster scheduler for deep learning training.
Personal Webpage of Anirban Laha
AutoML library for deep learning
Collective Knowledge repository to support artifact evaluation and reproducibility initiatives:
Samples for CUDA Developers which demonstrates features in CUDA Toolkit
Personal Testing
A set of examples around pytorch in Vision, Text, Reinforcement Learning, etc.
Tracing page fault statistics (e.g., avg. page fault time) using ftrace in Linux
This repository contains data indexes from NIST's Genome in a Bottle project.
The implementation of HawkEye, our research system: "HawkEye: Efficient Fine-grained OS Support for Huge Pages" from ASPLOS 2019.
This is the implementation of our research system Illuminator that was published in ASPLOS 2018 with the title "Making Huge Pages Actually Useful".
This is the canonical git mirror of the LLVM subversion repository. The repository does not accept github pull requests at this moment. Please submit your patches at http://reviews.llvm.org.
Experimental ground for optimizing memory of pytorch models
SDNet
Reference models and tools for Cloud TPUs.
This is a user-space tool that profiles all running applications under a specified user, and periodically outputs their MMU Overhead i.e., the fraction of CPU Cycles spent in servicing TLB misses. This tool was used to assist our research system HawkEye (published in ASPLOS'19) to achieve fair huge page allocation across multiple applications.