Lequn Chen || abcdabcd987

Journey to 2-second Inter-node RL Weight Transfer

Harnessing 3200Gbps Network (15): Lazy Posting

Harnessing 3200Gbps Network (14): Batch Posting

Harnessing 3200Gbps Network (13): State Sharding

Harnessing 3200Gbps Network (12): CPU Core Pinning

Harnessing 3200Gbps Network (11): Multi-threading

Harnessing 3200Gbps Network (10): Pre-benchmark Warmup

Harnessing 3200Gbps Network (9): Using 32 Network Cards

Harnessing 3200Gbps Network (8): Bus Topology

Harnessing 3200Gbps Network (7): Queuing and Benchmark

Harnessing 3200Gbps Network (6): GPUDirect RDMA WRITE

Harnessing 3200Gbps Network (5): Bidirectional SEND and RECV

Harnessing 3200Gbps Network (4): Unidirectional SEND and RECV

Harnessing 3200Gbps Network (3): libfabric

Harnessing 3200Gbps Network (2): High-Performance Network System Design Philosophy

Harnessing 3200 Gbps Network (1): RDMA and EFA

Harnessing 3200 Gbps Network: A Journey with RDMA, EFA, and libfabric

Potentials of Multitenancy Fine-Tuned LLM Serving

Dissecting Batching Effects in GPT Inference

How we discovered why C++ exceptions disappear in stack trace

Coq Tricks for Beginners with Too Many Examples