发明名称 Asymmetrical scaling multiple GPU graphics system for implementing cooperative graphics instruction execution
摘要 An asymmetrically scaling multiple GPU graphics system wherein the multiple GPUs are asymmetric, meaning that their rendering capabilities and/or rendering power is not equal. The asymmetric scaling multiple GPU graphics system includes a plurality of GPUs configured to execute graphics instructions from a computer system. A GPU output multiplexer and a controller unit are coupled to the GPUs. The controller unit is configured to control the GPUs and the output multiplexer such that the GPUs cooperatively execute the graphics instructions from the computer system.
申请公布号 US9087161(B1) 申请公布日期 2015.07.21
申请号 US200511093890 申请日期 2005.03.29
申请人 NVIDIA CORPORATION 发明人 Diamond Michael B.
分类号 G06F3/038;G09G5/00;G06F13/40 主分类号 G06F3/038
代理机构 代理人
主权项 1. A multiple GPU (graphics processor unit) graphics system comprising: a plurality of GPUs configured to execute graphics instructions from a computer system, wherein each GPU of the plurality of GPUs comprises a plurality of graphics rendering pipelines; a serial bus connector coupled to at least one GPU of the plurality of GPUs, wherein the serial bus connector is configured to removeably couple the at least one GPU to the computing system, wherein the serial bus is operable to hot-plug the at least one GPU to the computer system and enable the at least one GPU to communicate with the computer system through a serial bus; a GPU output multiplexer; and a controller unit coupled to the GPUs and the output multiplexer, wherein the controller unit is configured to control the GPUs and the output multiplexer such that the GPUs cooperatively execute the graphics instructions from the computer system, and wherein the GPUs have asymmetric graphics rendering capabilities, and wherein the controller unit is operable to allocate a first workload to a first GPU of the plurality of GPUs and operable to allocate a second workload to a second GPU of the plurality of GPUs, wherein the first workload is larger than the second workload, wherein the GPU output multiplexer is operable to multiplex output from the GPUs having asymmetric graphics rendering capabilities.
地址 Santa Clara CA US