Permissive BSD-style license, and can be used by for-profit entities.įor exact details, see the licenses directory in the SD-VBS distribution Most, but not all, of the benchmarks are available under the (CAREER: Energy-Efficient Parallel Architectures for Computer Vision.)Īll of the benchmarks can be used for academic or non-profit research. Brian Demsky of UC Irvine has Feature Tracking parallelized using his parallelizing compiler: 26x speedup on 62 tiles! Image size for those who are looking for larger working sets and run times.ġ1.21.09 Face Detection will be released at a later date MSER has been substituted for it. (including 800圆x1080) for those benchmarks that scale with
Version 1.2 (coming soon) of the benchmark suite will include larger inputs We give the run times, in cycles, below, on a 2.6 GHz Opteron system,īENCHMARK TEST SIM_FAST SIM SQCIF QCIF CIFĭisparity Map 41.6K 2.1M 6.1M 23.4M 291.2M 11.0Gįeature Tracking 242.8K 2.0M 5.8M 11.4G 12.5G 14.9G This version includes bug fixes and has more input data sets (vga, fullhd-1080p, wuxga). This version excludes wuxga input data set from all the benchmarks since the sizes of wuxga and fullhd were similar (1920x12x1080 respectively). This version includes bug fixes to the timing functions. This version improves memory allocation/deallocation so that SD-VBSĬode can be incorporated into vision applications.Ġ3.11.10 VBS is now being used by over a hundred researchers! This version includes bug fixes to the Image Segmentation benchmark. This version replaces stack allocation with a heap allocation for Image Segmentation benchmark. Look at these slides to see an overview of the algorithms and the benchmarks.Ġ1.15.11 U Penn has parallelized several of the benchmarks for their 2012 HPCA paper! Presented by Sravanthi Kota Venkata on October 6, 2009. Please cite this paper if you find the benchmark suite useful. IEEE Interational Symposium on Workload Characterization, October 2009. Sravanthi Kota Venkata, Ikkjin Ahn, Donghwan Jeon, Anshuman Gupta,Ĭhristopher Louie, Saturnino Garcia, Serge Belongie, and Michael Bedford Taylor. SD-VBS: The San Diego Vision Benchmark Suite
Benchmark computer architecture software#
Vision processors, and parallel software engineering tools for multicore here. You can learn more about Michael Taylor's UC San Diego research in manycore architecture, Researchers to control simulation time, and to understand propertiesĪs inputs increase to leverage better processor performance.ĭisparity Map Motion, Tracking and Stereo Vision Feature Tracking Motion, Tracking and Stereo Vision Image Segmentation Image Analysis Scale Invariant Feature Transform (SIFT) Image Analysis Maximally Stable Regions (MSER) Image Analysis Robot Localization Image Understanding Support Vector Machines (SVM) Image Understanding Image Stitch Image Processing and Formation Texture Synthesis Image Processing and Formation Minimizes pointer usage and employs clean constructs to makeįurthermore, we provide a spectrum of input sets that enable Is the preferred language of vision researchers, while C makes itĮasier to map the applications to research platforms. Eachīenchmark is provided in both MATLAB and C form.
The applications are drawn from the current state-of-the-art inĬomputer vision, in consultation with vision researchers. To have a fair amount of parallelism, which makes them a good candidateįor formulating future multicore and parallel architectures. It is intended to help architects, compiler writers, and systemĭesigners study the construction of future systems that excelĪt vision-oriented applications. Suite of diverse vision applications drawn from the vision domain. The San Diego Vision Benchmark Suite (SD-VBS) is a
Benchmark computer architecture download#
(Note: If you are trying to download from China, and are having problems with the above link, it is because of the Vision and brain benchmark suite for computer architects in existence! Thank you for your interest in the San Diego Vision Benchmark Suite, brought to you by Taylor's Bespoke Silicon Group (now at UW)!Ĭlick here for the new CortexSuite, the largest
0 kommentar(er)
