XPU - Article Directory

Using Local LLMs and Prompt Enhancement on XPU: How Maestro uses llama.cpp’s SYCL runtime for local prompt enhancement on Intel Arc GPUs, including the default 32K context limit.

In the world of natural language processing, Large Language Models (LLMs) have revolutionized the way we interact with machines. One such model is llama.cpp, a…

November 14, 2026

Artificial-intelligence

XPU-Specific Environment Variables and Memory Management: An explanation of variables like MAESTRO_XPU_MEMORY_FRACTION and the default capping of PyTorch’s XPU allocator to 90% of device memory.

When working with XPU-specific environment variables and memory management, it's essential to understand the default settings and how they impact your PyTorch applications. In this…

November 14, 2026

Artificial-intelligence

Configuring Intel Arc GPUs in Pinokio for Maestro: A guide on how to force device detection (MAESTRO_DEVICE=xpu) when Pinokio doesn’t automatically identify an Intel Arc GPU.

Configuring Intel Arc GPUs in Pinokio for Maestro: A Guide on How to Force Device Detection When it comes to leveraging the power of Intel…

November 14, 2026