#
aiperf
Here are 4 public repositories matching this topic...
Benchmark: how many agentic sessions one 8x H100 node serves on Qwen3-235B-A22B FP8 with NVIDIA Dynamo, vLLM and TensorRT-LLM, and what moves the limit.
performance-engineering benchmark gpu moe ai-agents inference-optimization mixture-of-experts kv-cache fp8 llm-serving vllm llm-inference h100 tensorrt-llm agentic-ai qwen3 nvidia-dynamo aiperf disaggregated-inference expert-parallelism
-
Updated
Oct 24, 2026 - Python
Load benchmark for OpenAI-compatible LLM servers with character-chat and RAG-shaped traffic: cost per 1M output tokens, per turn and per session at your latency SLO (built on AIPerf)
-
Updated
Oct 20, 2026 - Python
Add this topic to your repo
To associate your repository with the aiperf topic, visit your repo's landing page and select "manage topics."