In recent years, GPUs have become essential for handling high-performance computations, particularly in fields such as artificial intelligence (AI), where complex algorithms require significant processing power. To stay competitive in this rapidly evolving landscape, Intel has focused on developing more powerful discrete GPUs. These GPUs are managed by the i915 device driver, an integral part of the open-source Linux kernel. Intel dedicates substantial resources to continuously update and maintain the driver, ensuring it meets both the industry’s technological demands and the open-source community’s requirements. In this thesis, I examine the general architecture and development process of the i915 driver within the community, with a focus on the implementation of partial memory mapping and compute engine static load balancing.