Mesh traversal is a common and essential geometry processing problem in computer graphics. The traversal typically processes each face in a mesh in a systematic and consistent order for different applications such as mesh compression, rendering, curvature tracing, mesh simplification and surface smoothing. While cache-efficient mesh traversal methods where data and computations are reordered for good cache reuse have been well-studied, their run-time performances are limited by implicit(automatic) memory management. In this work we explore explicit optimizations on Explicitly Managed Memory (EMM) systems. Unlike conventional processors, EMM hardware has no automatic caching or prefetching. Instead, programmers decide what, when and where data are placed in the memory hierarchy, and then manually initiate the transfer and move the data by software. Programming such a system, however, is challenging since having programmers understand the hierarchies and manually move the data can be difficult and tedious. We address this challenge by introducing an access analyzer - distance profiler, associated with an interface construct. We demonstrate the effectiveness of the access analyzer and the runtime improvement obtained for several mesh traversal algorithms with different access patterns.
Copyright is held by the author.
The author granted permission for the file to be printed and for the text to be copied and pasted.
Supervisor or Senior Supervisor
Thesis advisor: Zhang, Richard
Member of collection