Philip Taffet

About Philip

Hi, I'm a PhD student in the computer science department at Rice University. I was formerly an undergraduate student (B.S. Mathematics, B.S. Computer Science, May 2017) and a Master's student (M.S. Computer Science, May 2018) at Rice.

I love hard problems, running, thinking, programming, and reading. I'm passionate about technology and engineering.

Last summer, I interned in the Linux and R&D teams at Jump Trading exploring InfiniBand network congestion. The two summers before, I worked with Ian Karlin at Lawrence Livermore National Lab studying the impact of network congestion on HPC applications. Prior to that, I worked on HPC performance analysis with Chevron's Emerging Technology group (2016), Azure Service Fabric with the Hyper-scale Compute team of Microsoft Azure (2015), the Oxbow project with the Future Technologies Group at Oak Ridge National Lab (2014), and HPCToolkit at Rice University (2013, 2012).

Research and Publications

I currently focus on techniques for improving communication performance for HPC applications. My research advisor is John Mellor-Crummey.

SC'19: PT, JMC. Understanding congestion in high performance interconnection networks using sampling. (link)
HOTI'19: PT, JMC. Lightweight, packet-centric monitoring of network traffic and congestion implemented in P4. (link)
PMBS'19: PT, S. Rao, E. Leon, I. Karlin. Testing the Limits of Tapered Fat Tree Networks. (link)
Invited Poster at Salishan 19: PT, JMC. Understanding congestion in high performance interconnection networks using sampling. (link)
Master's Thesis: PT. Understanding congestion in high performance interconnection networks using sampling. (link)
SC'18 poster: PT, S. Rao, I. Karlin. Exploring application performance on fat-tree networks in the presence of congestion. (link)
SC'17 student poster (best poster finalist): PT, I. Karlin, E. Leon, JMC. Understanding the impact of fat-tree network locality on application performance. (link)

Other Miscellaneous Publications

ICPPW'16: L. Adhianto, PT. Addressing challenges in visualizing huge call-path traces. (link)
ACSSC'17: Y. Arefeen, PT, et al. Real-time, data-driven system to learn parameters for multisite pacemaker beat detection. (link)
Co-HPC'14: S. Sreepathi, M.L. Grodowitz, R. Lim, PT, et al. Application characterization using Oxbow toolkit and PADS infrastructure. (link)

Resume Full CV BibTeX

Selected Personal Projects

When I found that Mali does not follow the normal rules restricting Unicode characters in domain names, I knew I had to build this. It's a firework simulator game I built for HackRice 2016. Uses SignalR and websockets to allow for a real-time multiplayer firework experience.
C Preprocessor Debugger
After struggling with macro expansion and the C preprocessor, I decided to build a tool to help. This Clang plugin shows you how the preprocessor recursively expands your macros, one step at a time. I also built a browser-based version but few browsers can load a 150 MB JavaScript file.

Tasks and Calendars
A personal Universal Windows Platform app that helps me manage my integrated to-do list and calendars, all stored on SharePoint 2013. I gained experience with the SharePoint CSOM and with writing Universal Windows Apps.
Password Manager
A password manager that does not require a master password but instead uses certificates and an additional second factor when connecting from a remote network. I learned more about certificates, managing a PKI, and writing client-server applications.
Additional information and other projects available on Github or by request.

Contact