# Overview

Intel® software for general-purpose GPU capabilities enables Intel GPUs to function effectively on Linux systems. It ensures that the GPUs are recognized and used efficiently for a wide range of tasks, including graphics-intensive applications and general-purpose computing.

Intel® General-Purpose GPUs are versatile graphics processing units designed to handle a variety of tasks across different computing environments, from personal computers to large data centers. Data Center GPUs boost performance for complex tasks like AI, scientific simulations, and video processing. Client GPUs enhance gaming, multimedia, and general computing experiences by delivering smoother graphics and faster processing. Intel GPUs are built to be efficient, powerful, and compatible with a wide range of software, making them a solid choice for both professional and personal use.

## Data Center GPUs

Intel offers the following Data Center GPUs:

- [Intel® Data Center GPU Max Series](http://intel.com/maxseriesgpu): The highest-performing, highest-density, general-purpose discrete GPU, which is Intel’s foundational GPU compute building block. It has over 100 billion transistors and contains up to 128 Xe cores.

- [Intel® Data Center GPU Flex Series](http://intel.com/flexseriesgpu): A general-purpose GPU optimized for media stream density and quality. It ensures high reliability, availability, and scalability. Its hardware includes Xe cores and media acceleration.

Solutions built using Intel GPU products rely on several software components working together. The following diagram presents the GPU architecture. 

![Software Stack](../assets/generic-stack.svg)

The Intel GPU Base provides standard interfaces to higher-level use cases.

The Linux\* kernel drivers provide the software connection to the Intel GPU hardware.
The drivers are provided as Dynamic Kernel Module Support ([DKMS](https://github.com/dell/dkms))
drivers "out-of-tree" from the Linux kernel. See [Kernel Driver Types](kernel-driver-types)
for more information.

The Compute Runtime and Media usermode library stacks provide higher-level software support of
industry-standard APIs. Compute Runtime supports OpenCL and Level Zero, which are a foundation for oneAPI GPU acceleration. The Media stack provides oneVPL with integrations for [GStreamer](https://gstreamer.freedesktop.org/) and [ffmpeg](https://ffmpeg.org/).

Intel® [XPU Manager (XPUM)](https://github.com/intel/xpumanager) is designed for managing the physical hardware, including firmware updates, monitoring, diagnostics, and configuration. XPUM uses
the oneAPI Level Zero Sysman API and exposes standard interfaces for telemetry, including [Prometheus](https://prometheus.io/).

## Client GPUs

Intel offers the [Intel® Arc series GPUs](https://www.intel.com/content/www/us/en/products/details/discrete-gpus/arc.html), which are designed for gamers, creators, and everyday users. These GPUs deliver high-performance graphics for gaming, fast video editing, and smooth streaming. They are compatible with a wide range of devices and applications, providing a versatile solution for both professional and personal use, ensuring a rich visual experience across all platforms.