A cache-friendly sampling strategy for texture-based volume rendering on GPU
The texture-based volume rendering is a memory-intensive algorithm. Its performance relies heavily on the performance of the texture cache. However, most existing texture-based volume rendering methods blindly map computational resources to texture memory and result in incoherent memory access patte...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2017-06-01
|
| Series: | Visual Informatics |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2468502X1730027X |
| _version_ | 1818542321954717696 |
|---|---|
| author | Junpeng Wang Fei Yang Yong Cao |
| author_facet | Junpeng Wang Fei Yang Yong Cao |
| author_sort | Junpeng Wang |
| collection | DOAJ |
| description | The texture-based volume rendering is a memory-intensive algorithm. Its performance relies heavily on the performance of the texture cache. However, most existing texture-based volume rendering methods blindly map computational resources to texture memory and result in incoherent memory access patterns, causing low cache hit rates in certain cases. The distance between samples taken by threads of an atomic scheduling unit (e.g. a warp of 32 threads in CUDA) of the GPU is a crucial factor that affects the texture cache performance. Based on this fact, we present a new sampling strategy, called Warp Marching, for the ray-casting algorithm of texture-based volume rendering. The effects of different sample organizations and different thread-pixel mappings in the ray-casting algorithm are thoroughly analyzed. Also, a pipeline manner color blending approach is introduced and the power of warp-level GPU operations is leveraged to improve the efficiency of parallel executions on the GPU. In addition, the rendering performance of the Warp Marching is view-independent, and it outperforms existing empty space skipping techniques in scenarios that need to render large dynamic volumes in a low resolution image. Through a series of micro-benchmarking and real-life data experiments, we rigorously analyze our sampling strategies and demonstrate significant performance enhancements over existing sampling methods. |
| first_indexed | 2024-12-11T22:20:33Z |
| format | Article |
| id | doaj.art-46f8540dca534a4a85b20fdc78073a05 |
| institution | Directory Open Access Journal |
| issn | 2468-502X |
| language | English |
| last_indexed | 2024-12-11T22:20:33Z |
| publishDate | 2017-06-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Visual Informatics |
| spelling | doaj.art-46f8540dca534a4a85b20fdc78073a052022-12-22T00:48:27ZengElsevierVisual Informatics2468-502X2017-06-01129210510.1016/j.visinf.2017.08.001A cache-friendly sampling strategy for texture-based volume rendering on GPUJunpeng Wang0Fei Yang1Yong Cao2The Ohio State University, Columbus, OH 43210, USANVIDIA Semiconductor Technology (Shanghai) Co., Ltd., Shanghai, ChinaThe Boeing Company, North Charleston, SC 29418, USAThe texture-based volume rendering is a memory-intensive algorithm. Its performance relies heavily on the performance of the texture cache. However, most existing texture-based volume rendering methods blindly map computational resources to texture memory and result in incoherent memory access patterns, causing low cache hit rates in certain cases. The distance between samples taken by threads of an atomic scheduling unit (e.g. a warp of 32 threads in CUDA) of the GPU is a crucial factor that affects the texture cache performance. Based on this fact, we present a new sampling strategy, called Warp Marching, for the ray-casting algorithm of texture-based volume rendering. The effects of different sample organizations and different thread-pixel mappings in the ray-casting algorithm are thoroughly analyzed. Also, a pipeline manner color blending approach is introduced and the power of warp-level GPU operations is leveraged to improve the efficiency of parallel executions on the GPU. In addition, the rendering performance of the Warp Marching is view-independent, and it outperforms existing empty space skipping techniques in scenarios that need to render large dynamic volumes in a low resolution image. Through a series of micro-benchmarking and real-life data experiments, we rigorously analyze our sampling strategies and demonstrate significant performance enhancements over existing sampling methods.http://www.sciencedirect.com/science/article/pii/S2468502X1730027XWarp marchingTexture cache hit rateGPUVolume rendering |
| spellingShingle | Junpeng Wang Fei Yang Yong Cao A cache-friendly sampling strategy for texture-based volume rendering on GPU Visual Informatics Warp marching Texture cache hit rate GPU Volume rendering |
| title | A cache-friendly sampling strategy for texture-based volume rendering on GPU |
| title_full | A cache-friendly sampling strategy for texture-based volume rendering on GPU |
| title_fullStr | A cache-friendly sampling strategy for texture-based volume rendering on GPU |
| title_full_unstemmed | A cache-friendly sampling strategy for texture-based volume rendering on GPU |
| title_short | A cache-friendly sampling strategy for texture-based volume rendering on GPU |
| title_sort | cache friendly sampling strategy for texture based volume rendering on gpu |
| topic | Warp marching Texture cache hit rate GPU Volume rendering |
| url | http://www.sciencedirect.com/science/article/pii/S2468502X1730027X |
| work_keys_str_mv | AT junpengwang acachefriendlysamplingstrategyfortexturebasedvolumerenderingongpu AT feiyang acachefriendlysamplingstrategyfortexturebasedvolumerenderingongpu AT yongcao acachefriendlysamplingstrategyfortexturebasedvolumerenderingongpu AT junpengwang cachefriendlysamplingstrategyfortexturebasedvolumerenderingongpu AT feiyang cachefriendlysamplingstrategyfortexturebasedvolumerenderingongpu AT yongcao cachefriendlysamplingstrategyfortexturebasedvolumerenderingongpu |