发明名称 SYSTEM AND METHOD FOR GLOBALLY ADDRESSABLE GPU MEMORY
摘要 A system and method for efficient memory access. The method includes receiving a request to access a portion of memory. The request comprises a first address. The method further includes determining whether the first address corresponds to a thread local portion of memory and in response to the first address corresponding to the thread local portion of memory, translating the first address to a second address. The method further includes accessing the thread local portion of memory based on the second address. The second address corresponds to an offset in a region of memory reserved for storing thread local data and allocations into the region are contiguous for a plurality of threads at each thread local offset.
申请公布号 US2014310484(A1) 申请公布日期 2014.10.16
申请号 US201313864182 申请日期 2013.04.16
申请人 NVIDIA Corporation 发明人 GIROUX Olivier
分类号 G06F3/06 主分类号 G06F3/06
代理机构 代理人
主权项 1. A method for configuring memory for access, said method comprising: accessing a portion of an executable program; generating a group of threads comprising a plurality of threads based on said portion of said executable program; assigning each thread of said plurality of threads a respective unique identifier; allocating a respective portion of local memory to each of said plurality of threads, wherein said respective unique identifier is operable for determining a respective base address of said respective portion of local memory corresponding to a respective thread and wherein said respective portion of local memory is operable to be accessed by each of said plurality of threads and wherein each respective portion of local memory comprises a respective contiguous portion corresponding to an offset for each thread of said plurality of threads.
地址 Santa Clara CA US