MMA routes host-GPU transfers over multiple available paths to deliver 4.62x higher peak bandwidth and lower latencies in LLM serving without hardware or driver changes.
Boosting large-scale parallel training efficiency with c4: A communication-driven approach
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.DC 2representative citing papers
Reversa turns legacy systems into confidence-marked, code-traceable operational specifications via a multi-agent pipeline, shown only in an incomplete COBOL-to-Go ATM case study.
citing papers explorer
-
MultiPath Memory Access: Breaking Host-GPU Bandwidth Bottlenecks in LLM Services
MMA routes host-GPU transfers over multiple available paths to deliver 4.62x higher peak bandwidth and lower latencies in LLM serving without hardware or driver changes.
-
EPIC: Abstraction and Polymorphism of In-Network Collectives on Ethernet
Reversa turns legacy systems into confidence-marked, code-traceable operational specifications via a multi-agent pipeline, shown only in an incomplete COBOL-to-Go ATM case study.