Ollama, Llama.cpp, and LMStudio : LLM Showdown in Windows: i9-13900kf Benchmarks скачать в хорошем качестве

Ollama, Llama.cpp, and LMStudio : LLM Showdown in Windows: i9-13900kf Benchmarks

Not everyone has $3000 for a high-end gpu. In this video we hope to show that even a high end office computer cpu can run a LLM locally with adequate performance. Using gpt-oss:20b, we highlight the effectiveness of this BEAST cpu running local LLM. We pushed the limits of the powerful i9-13900KF to see how it handles local LLM workloads across three major platforms: Ollama, Llama.cpp, and LM Studio. This benchmark reveals which tool delivers the best CPU utilization and inference speed, helping you unlock the true AI potential of your high-end Windows desktop. No gpu required. Build: Intel i9-13900k - https://newegg.io/nca343b709 Intel i9-113900kf - https://newegg.io/nc3fa7de10 Motherboard MSI z690 Carbon - https://newegg.io/nc7c96f574 MSI z790 MEG ACE - https://newegg.io/nc07fdfc0c RAM G.Skill Trident 6000 / cl30 - https://newegg.io/nc671c78cc G.Skill Trident 8000 / cl38 - https://newegg.io/nc0125fbd2 (if your mobo supports that high) 0:00 - Intro 3:07 - Ollama 16:21 - Llama.cpp 29:21 - LMStudio 41:14 - Closing