Posts by Jun Kang Chow
DeepSeek-V4 Flash RL on AMD Instinct MI308X GPUs with an AITER-Backed RL Stack
- 28 October 2026
We demonstrate end-to-end DeepSeek-V4 Flash reinforcement learning with an AITER-backed AMD RL stack on 24 AMD Instinct MI308X GPUs. Our test covers the complete workflow on three MI308X OAM nodes—not only inference or an isolated training step, but long-context FP8 rollout, BF16 MoE actor training, online RDMA weight updates, evaluation, and checkpoint recovery. Training and inference run as disaggregated services: two MI308X nodes host a BF16 actor that preserves the numerical stability required by gradient computation and optimizer updates, while a dedicated third MI308X node hosts an FP8 rollout engine whose smaller per-parameter footprint leaves more GPU memory available for KV cache and longer generated sequences. Ray places and coordinates both GPU groups. This experiment demonstrates one tested three-node configuration for running both rollout inference and RL training on MI308X GPUs.
Accelerating Multimodal Inference in vLLM: The One-Line Optimization for Large Multimodal Models
- 02 January 2026
Deploying multimodal models like Qwen3-VL or InternVL at scale reveals a hidden bottleneck. While Tensor Parallelism (TP) is essential for massive language decoders, it is often overkill for vision encoders. These encoders are typically small, often just 1-5% of total model size, so there is limited compute benefit from sharding them. However, they still incur expensive all-reduce communication costs after every single layer.
The vLLM MoE Playbook: A Practical Guide to TP, DP, PP and Expert Parallelism
- 24 November 2025
Deploying large Mixture-of-Experts (MoE) models like DeepSeek-R1 efficiently isn’t just about having enough GPUs—it’s about choosing the right parallelism strategy. The wrong choice can lead to duplicated KV caches consuming 8× your memory, or communication overhead that cuts throughput in half. The right choice unlocks significantly better performance for your specific workload.