Posts by Lirong Zhang

DP Attention and TBO for DeepSeek-V4 on MI355X

Running DeepSeek-V4 efficiently requires solving two intertwined problems: how to parallelize MoE communication across GPUs, and how to hide that communication behind useful compute. The dominant approach is Expert Parallel with all2all backends like DeepEP. This solves both problems, but it also requires specialized kernels, topology assumptions, and careful expert placement.

Read more ...


ATOMesh: Unlocking AMD Hardware for Scalable LLM Serving

Large language model serving is moving from single-engine optimization to full-stack distributed inference. Production deployments must handle high concurrency, long-context prefill, latency-sensitive decode, KV cache store pressure, and multi-node GPU utilization at the same time. On AMD Instinct GPUs, the key opportunity is to connect ROCm-native kernels, communication libraries, inference engines, and distributed orchestration into one scalable serving stack.

Read more ...