REVIEW 6 cited by
An Extensible Software Transport Layer for GPU Networking
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
An Extensible Software Transport Layer for GPU Networking
read the original abstract
Fast-evolving machine learning (ML) workloads have increasing requirements for networking. However, host network transport on RDMA NICs is hard to evolve, causing problems for ML workloads. For example, single-path RDMA traffic is prone to flow collisions that severely degrade collective communication performance. We present UCCL, an extensible software transport layer to evolve GPU networking. UCCL decouples the data path and control path of existing RDMA NICs and efficiently runs the control-path transport on host CPUs. This software extensibility brings in transport innovations that cannot be achieved in hardware for ML workloads, e.g., a multipath transport to resolve flow collisions. ML collectives atop UCCL achieve up to 4.5x higher performance compared to existing RDMA NICs.
Forward citations
Cited by 6 Pith papers
-
OrderMoE: An expert similarity driven distributed edge MoE inference
OrderMoE reduces cross-server traffic in edge MoE inference by substituting a similar local expert for the router's remote target expert, with only small quality loss.
-
GF-DiT: Scheduling Parallelism for Diffusion Transformer Serving
GF-DiT introduces elastic GPU parallelism scheduling for DiT serving via asynchronous trajectory tasks and group-free collectives, reporting up to 6.01x throughput gains over static configurations.
-
GF-DiT: Scheduling Parallelism for Diffusion Transformer Serving
GF-DiT dynamically adapts parallelism during DiT serving via trajectory tasks and group-free collectives, reporting up to 6x throughput and 95% latency reduction versus static configurations.
-
UCCL-Zip: Lossless Compression Supercharged GPU Communication
UCCL-Zip adds lossless compression to GPU communication to reduce LLM bottlenecks while preserving exact numerical correctness.
-
TENT: A Declarative Slice Spraying Engine for Performant and Resilient Data Movement in Disaggregated LLM Serving
A telemetry-driven slice-spraying transfer engine with late-binding path selection improves LLM serving throughput and self-heals from link failures in tens of milliseconds.
-
PICO: Performance Insights for Collective Operations
PICO is a benchmarking framework for collective operations that decouples portable setup from platform execution, supplies reference MPI implementations, and shows default choices can be up to 5x slower with up to 44%...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.