A measured benchmark on a modded 48GB RTX 4090 shows Strata's n-gram table belongs on SSD while conversation parking delivers the large speedup.
Start with your machine.
Capacity is the reported total for this setup. Multi-GPU and unified-memory layouts can behave differently.
A documented, measured local coding setup: Qwen 3.8 27B on one RTX 4090 with llama.cpp and OpenCode, holding a 128k context and reporting around 92 tokens/s generation with speculative decoding.
The same project's measured run of a ternary-quantised 27B model: about 92 tokens/s generation with no speculative decoding, from 6.7 GB of weights, at a documented cost in agentic coding ability.
Three GPUs spread across machines run Qwen 3.8 27B through a custom stateful agent layer to build a personal finance and portfolio tracker without shipping documents to the cloud.