3 papers
cs.DC2026
Piper: A Programmable Distributed Training System
Megan Frisella, Shubham Tiwari, Andy Ruan +5
Large-scale model training increasingly relies on composing multiple parallelism strategies, such as data, pipeline, and expert parallelism, together with memory-saving optimizatio…
cs.ET2025
Jenga: Responsive Tiered Memory Management without Thrashing
Rohan Kadekodi, Haoran Peng, Gilbert Bernstein +2
A heterogeneous memory has a single address space with fast access to some addresses (a fast tier of DRAM) and slow access to other addresses (a capacity tier of CXL-attached memor…
cs.PL2025
Exo 2: Growing a Scheduling Language
Yuka Ikarashi, Kevin Qian, Samir Droubi +3
User-schedulable languages (USLs) help programmers productively optimize programs by providing safe means of transforming them. Current USLs are designed to give programmers exactl…