H200 in Practice: Kimi K3 Throughput Tuning and FP8 KVCache
hyhieu226 · x · 2026-08-01
A developer shared practical insights on tuning the Kimi K3 model for throughput-oriented workloads on H200 GPUs.
- Architecture Choice: TP+EP (Tensor Parallelism + Expert Parallelism) outperforms DP+EP (Data Parallelism + Expert Parallelism) for K3. Since only about 1/4 of K3's layers use gated MLA, DP attention provides less benefit from avoiding KV-cache duplication.
- Memory Optimization: Discussed the application of FP8 KVCache and highlighted the "magic number" trick for balancing precision and memory footprint.
Related event: Kimi K3 Throughput Optimization on H200 Clusters(3 posts)→
More from Infra
- Reflection's 501B MoE Beam hits 463 tok/s on 8x GB200 with NVFP4 at day 0 — zhyncs42 · 2026-10-16
- Nemotron 3.5 Bolt runs 16 concurrent streams on a single DGX Spark — NVIDIAAI · 2026-10-16
- Why your computer can't store 0.2: an IEEE 754 primer — Franc0Fernand0 · 2026-10-16
- Sabaki: MIT-licensed typed decision router runs on CPU in milliseconds — Fox3000foxy · 2026-10-16
- BitRun v0.2 runs 8B ternary LLMs on CPU in 2GB RAM with live wattage telemetry — The_VibePreneur · 2026-10-16
- MiniCPM-V 4.5 MoE reportedly hits 13+ tokens/s on iPhone with just 3GB memory — heyshrutimishra · 2026-10-16