公开(公告)号：US08854380B2

公开(公告)日：2014-10-07

申请号：US14026901

申请日：2013-09-13

Applicant: NVIDIA Corporation

Inventor： Franck R. Diard ,  Ian M. Williams ,  Eric Boucher

IPC: G06F15/16 ,  G06T1/00 ,  G06T15/00 ,  G09G5/00 ,  G06T3/40 ,  G06T5/00 ,  G06T1/20

CPC classification number: G06T5/00 ,  G06F3/1438 ,  G06F3/1446 ,  G06F9/452 ,  G06T1/20 ,  G06T3/4038 ,  G09G5/00 ,  G09G5/399

Abstract: One embodiment of the present invention sets forth a technique for displaying high-resolution images using multiple graphics processing units (GPUs). The graphics driver is configured to present one virtual display device, simulating a high-resolution mosaic display surface, to the operating system and the application programs. The graphics driver is also configured to partition the display surface amongst the GPUs and transmit commands and data to the local memory associated with the first GPU. A video bridge automatically broadcasts this information to the local memories associated with the remaining GPUs. Each GPU renders and displays only the partition of the display surface assigned to that particular GPU, and the GPUs are synchronized to ensure the continuity of the displayed images. This technique allows the system to display higher resolution images than the system hardware would otherwise support, transparently to the operating system and the application programs.

Abstract translation: 本发明的一个实施例提出了使用多个图形处理单元（GPU）显示高分辨率图像的技术。 图形驱动器被配置为向操作系统和应用程序呈现一个虚拟显示设备，模拟高分辨率马赛克显示表面。 图形驱动器还被配置为在GPU之间划分显示表面，并将命令和数据传送到与第一GPU相关联的本地存储器。 视频桥自动将该信息广播到与剩余GPU相关联的本地存储器。 每个GPU渲染并显示分配给该特定GPU的显示表面的分区，并且GPU被同步以确保所显示的图像的连续性。 该技术允许系统显示比系统硬件以其他方式支持的更高分辨率的图像，对操作系统和应用程序是透明的。