PLSECompilers @ Georgia Tech

Programming Languages/Compilers/Software Engineering/Formal Verification @gatech_scs

[PLSECompilers @ Georgia Tech @PLCompilersGT]

Presenting the lighting talk for our paper at #OOPSLA2023 (@splashcon ) - "Beacons: An End-to-End Compiler Framework for Predicting and Utilizing Dynamic Loop Characteristics"

https://t.co/be0ZdIWaHa

Feel free to drop by at the poster session, and the talk tomorrow!

[PLSECompilers @ Georgia Tech @PLCompilersGT]

Congrats to all the authors:

Girish, @SharjeelK4754, @AbirBodhisatwa1, Chris, Chao, Ada, Santosh
&
Fangke, Jisheng, @vsarkar_gt

Feel free to drop by if you are attending the conference at Cascais!

[PLSECompilers @ Georgia Tech @PLCompilersGT]

We are pleased to announce the acceptance of two papers in #OOPSLA23 (@splashcon), Portugal:

Beacons: An End-to-End Compiler Framework for Predicting & Utilizing Loop Characteristics

Concrete Type Inference for Code Optimization using Machine Learning using SMT solving

[PLSECompilers @ Georgia Tech @PLCompilersGT]

The Com-CAS paper showed how to perform smart cache apportioning accounting for applications' dynamic phase behaviors, obtaining speedup of 20% over SOTA.

[PLSECompilers @ Georgia Tech @PLCompilersGT]

We're delighted to announce the acceptance of two papers in PACT'2022, Chicago:

ReACT: Redundancy-Aware Code Generation for Tensor Expressions by Tong Zhou, @vsarkar_gt

Com-CAS: Effective Cache Apportioning Under Compiler Guidance by @AbirBodhisatwa1 @SharjeelK4754 #pact2022

[PLSECompilers @ Georgia Tech @PLCompilersGT]

The ReACT paper showed how to eliminate a few common types of redundant computations when generating code for sparse tensor operations, which resulted in 1.05x - 100x speedups over SOTA. This was a collaboration with folks @PNNLab