Above the Fold
I (Peter) am working on a project to detect the particle emission from black holes. The project centers on a complex transport calculation to determine how many particles from the black hole end up near Earth after diffusing across 45 billion kilometers of the Solar System, which is filled with disordered magnetic fields. I started with a toy model using the Mathematica package on my laptop, which morphed into a process that requires 16 jobs that take 20 CPU hours per run on my Mac. Ultimately, the project will require dozens or hundreds of runs – time to move off the laptop.
I’ve decided to use this project to make a head-to-head comparison between running on Enagaging and a cloud provider (to be named maybe later). I consider this a small/medium-sized project akin to most of the computing that happens on Engaging. In comparing between the cloud and Engaging, I am of course interested in cost, but also interoperability (how easily code and data port between my laptop, Engaging and the cloud environment); wall clock performance (the elapsed time per run); and ease of use (if this succeeds, our UROP student will have to learn to run jobs).
My first tasks are to get set up on Engaging – especially setting up checkpointing – and to get an account and access to the cloud service. More to come…
ORCD Seed Fund
The MIT Office of Research Computing and Data (ORCD) is continuing our strategic Seed Fund initiative. The Seed Fund is one part of a strategic effort supporting infrastructure for applied AI/ML research, creating a powerful AI/ML platform for all of MIT.
The deadline for this round of Seed Fund proposals is October 1, 2025 at 12PM ET. See ORCD’s website for more details and application instructions.
ORCD Student Jobs
ORCD is looking for students to support our Research Community Facing team in the Fall 2025 semester as Research Computing Help Desk Support student employees.
See the job description file for more details. Applications must be submitted through the Student Financial Services portal (Job ID: 5172).
What We’re Reading
- The Massachusetts Technology Collaborative has issued an RFP to support the design, development, and launch of the Data Commons Collaborative, a foundational initiative under the Massachusetts AI Hub. This engagement aims to centralize access to data for AI development, research, and innovation.
- What lights up when you use ChatGPT–this article got a lot of press.
- Passings: Rai Weiss and John Joannopoulos both passed away this month. Both were giants at MIT.
- Arts Corner: MIT’s Student Lending Art Program and why MIT students write poetry
- Multiple curious pre-print articles this month on AI “automating” research?
- Interesting lessons in how not to deploy AI in a class!
ORCD Fall Training Schedule
ORCD will be teaching multiple classes to members of the MIT community during the fall of 2025 at MIT.
- Introduction to Engaging:
- Tuesday, September 9, 2-4 PM
- Wednesday, September 24, 2-4 PM
- Tuesday, October 14, 2-4 PM
- Wednesday, November 12, 10 AM - 12PM
- Introduction to Parallel Programming Series:
- Parallel Programming with OpenMP: Wednesday, October 22, 10:30 AM - 12:30 PM
- Distributed Computing with MPI: Wednesday, October 22, 2 - 4 PM
- GPU Programming with CUDA: Thursday, October 23, 10:30 AM - 12:30 PM
- Distributed Deep Learning: Thursday, October 23, 2 - 4 PM
See the 2025 fall classes page on ORCD’s website for more information and signup instructions.