Talks
I enjoy preparing and giving talks about my work. So, here’s a page specifically for the talks I gave with short description and available materials for them, if any. A more concise version of this list also exist in my CV.
Please contact me if you want to set up a talk in your institution or company about my work on Chapel and Arkouda.
LUMI User Coffee Breaks | October 15, 2025
In this one, I wanted to focus on some of the accomplishments of Chapel users to motivate the language, rather than its syntax, though I covered some of the key features as well. The key message was that Chapel makes the path from a laptop to a supercomputer much easier
Scalable Parallel Programming with Chapel: From Multicore CPUs to GPU-Powered Supercomputers (a.k.a What Chapel Users Get Done and How)
Slides & VideoIn this one, I wanted to focus on some of the accomplishments of Chapel users to motivate the language, rather than its syntax, though I covered some of the key features as well. The key message was that Chapel makes the path from a laptop to a supercomputer much easier
Los Alamos National Laboratory - Advances in Applied Computer Science Invited Speaker Series | June 18, 2025
How Chapel can be used to program most common parallel architectures, using the same set of first-class language featuers. Includes code examples, benchmark results and user stories
Chapel's Batteries-Included Approach for Portable Parallel Programming
SlidesHow Chapel can be used to program most common parallel architectures, using the same set of first-class language featuers. Includes code examples, benchmark results and user stories
NASA Ames Research Center | February 20, 2025
This is a joint talk with our colleagues Eric Laurendeau and Karim Zayni from Polytechnique Montreal. We covered what Chapel is and how it helps CHAMPS team to build their CFD solver
High-Performance, Productive Programming using Chapel with Examples from the CFD Solver CHAMPS
Slides & VideoThis is a joint talk with our colleagues Eric Laurendeau and Karim Zayni from Polytechnique Montreal. We covered what Chapel is and how it helps CHAMPS team to build their CFD solver
HPE Inner Sourcing Summit II | October 4, 2024
In this talk that was internal to HPE, we discussed ways in which fellow HPE employees can contribute to the Chapel ecosystem. As Chapel is an open-source language, the content of the talk did not include much proprietary discussion. The slides that are available above were only lightly edited before making them public.
The Chapel Parallel Programming Language and its Ecosystem
SlidesIn this talk that was internal to HPE, we discussed ways in which fellow HPE employees can contribute to the Chapel ecosystem. As Chapel is an open-source language, the content of the talk did not include much proprietary discussion. The slides that are available above were only lightly edited before making them public.
HPE Developer Meetup | July 31, 2024
A joint effort with my colleague Jade Abraham. I talked about Chapel's vendor-neutral GPU programming support, and Jade gave a live demo
Vendor-Neutral GPU Programming in Chapel
Slides & VideoA joint effort with my colleague Jade Abraham. I talked about Chapel's vendor-neutral GPU programming support, and Jade gave a live demo
AMD Developer Sync | May 31, 2024
I was invited to give a talk at this internal AMD event. One of the more in-depth technical discussions on Chapel's GPU support
Making Parallel Programming and GPUs More Accessible with Chapel
SlidesI was invited to give a talk at this internal AMD event. One of the more in-depth technical discussions on Chapel's GPU support
HPE Tech Talks | January 9, 2024
This is an internal HPE talk. A slightly earlier version of the talk I gave at AMD couple of months later
High-level, Vendor-Neutral GPU Programming Using Chapel
SlidesThis is an internal HPE talk. A slightly earlier version of the talk I gave at AMD couple of months later
SIAM PP22 Minisymp. on Code Gen. and Transformation in HPC on Heterogeneous Platforms | February 26, 2022
Quite possibly, the earliest talk I gave on Chapel's GPU support. I am not sure whether at that point we had a full-blown runtime implementation. Luckily, this is a code generation venue, and the talk is more about how the Chapel compiler automatically generates GPU kernels from regular data-parallel loops in Chapel.
Generating GPU Kernels from Chapel's Features for Parallelism and Locality
SlidesQuite possibly, the earliest talk I gave on Chapel's GPU support. I am not sure whether at that point we had a full-blown runtime implementation. Luckily, this is a code generation venue, and the talk is more about how the Chapel compiler automatically generates GPU kernels from regular data-parallel loops in Chapel.