LLM inference through the lens of a GPU: every number measured on real hardware, every claim sourced, every result reproducible from a public repo.https://tgarg01.github.io/https://tgarg01.github.io/sillybench-site/posts/01-a-loop-of-902-kernels/https://tgarg01.github.io/sillybench-site/posts/01-a-loop-of-902-kernels/What one decode step of Qwen3.8-27B looks like on 1x NVIDIA H100 GPU: 902 kernels in 20.55 ms, 92.6% of the GPU time in matrix multiplies. Measured with vLLM 0.31 and Nsight Systems.Sun, 18 Oct 2026 00:00:00 GMTLLMInferenceGPUCUDAvLLMNVIDIA