{"type": "article", "title": "When Prefill Stops Being Compute-Bound — The KV Cache Memory Wall and What It Does to Your Scheduler", "publisher": "Web Pulse", "url": "https://wpnews.pro/news/when-prefill-stops-being-compute-bound-the-kv-cache-memory-wall-and-what-it-does", "original_source": "https://kraghavan.ca/llm-infrastructure/inference/2026/09/07/kv-cache-memory-wall.html", "published": "2026-09-07T00:00:00+00:00", "accessed": "2026-09-28", "id": "when-prefill-stops-being-compute-bound-the-kv-cache-memory-wall-and-what-it-does"}