vLLM now matches DeepSeek's in-house DeepInfer engine on DeepSeek-V4.1-Flash decode
lmoroney · x · 2026-10-25
The vLLM team reports vLLM now matches DeepInfer, DeepSeek's in-house serving engine, on DeepSeek-V4.1-Flash decode across batch sizes 8–256. They got there by pulling in DeepSeek's open-source kernels plus heavy prefill/decode tuning, running DEP32 where DeepInfer used EP64DP64. Fun detail: DeepInfer was originally built on vLLM—open source came full circle. Advice: rerun your own benchmark with real prompt lengths on latest vLLM before maintaining a second serving stack.
More from Infra
- Sail equivalence proof and riscv-coq cross-checks uncover real RISC-V spec bugs — xennygrimmato_ · 2026-10-27
- xAI's Colossus 1 hits ~14.6 exaflops FP64, ~5x Top500 leader — cis_female · 2026-10-27
- Lambda runs 19 B200 nodes in a 16-node power budget, boosting throughput 24% — TheZachMueller · 2026-10-27
- Podcast: Micron's HBM roadmap, Meta's AI infra, and NVIDIA's supply chain lock-in — BenBajarin · 2026-10-27
- SGLang MegaKernel + PD Pushes the TCO-Latency Frontier on GB300, SemiAnalysis Benchmarks Show — AccBalanced · 2026-10-27
- Scaling self-hosted Convex: compiling TypeScript Convex functions to WASM — leventov · 2026-10-27