Convolutional networks (ConvNets) have become a popular approach to computer vision. It is important to accelerate ConvNet training, which is computationally costly. We propose a novel parallel algorithm based on deco...
详细信息
ISBN:
(纸本)9781509021413
Convolutional networks (ConvNets) have become a popular approach to computer vision. It is important to accelerate ConvNet training, which is computationally costly. We propose a novel parallel algorithm based on decomposition into a set of tasks, most of which are convolutions or FFTs. Applying Brent's theorem to the task dependency graph implies that linear speedup with the number of processors is attainable within the PRAM model of parallel computation, for wide network architectures. To attain such performance on real shared-memory machines, our algorithm computes convolutions converging on the same node of the network with temporal locality to reduce cache misses, and sums the convergent convolution outputs via an almost wait-free concurrent method to reduce time spent in critical sections. We implement the algorithm with a publicly available software package called ZNN. Benchmarking with multi-core CPUs shows that ZNN can attain speedup roughly equal to the number of physical cores. We also show that ZNN can attain over 90× speedup on a many-core CPU (Xeon Phi™ Knights Corner). These speedups are achieved for network architectures with widths that are in common use. The task parallelism of the ZNN algorithm is suited to CPUs, while the SIMD parallelism of previous algorithms is compatible with GPUs. Through examples, we show that ZNN can be either faster or slower than certain GPU implementations depending on specifics of the network architecture, kernel sizes, and density and size of the output patch. ZNN may be less costly to develop and maintain, due to the relative ease of general-purpose CPU programming.
Handedness of the director twist in cholesteric liquid crystals is commonly assumed to be the same throughout the medium, determined solely by the chirality of constituent molecules or chiral additives, albeit distort...
详细信息
Controlling functionalities, such as magnetism or ferroelectricity, by means of oxygen vacancies (VO) is a key issue for the future development of transition-metal oxides. Progress in this field is currently addressed...
详细信息
Controlling functionalities, such as magnetism or ferroelectricity, by means of oxygen vacancies (VO) is a key issue for the future development of transition-metal oxides. Progress in this field is currently addressed through VO variations and their impact on mainly one order parameter. Here we reveal a mechanism for tuning both magnetism and ferroelectricity simultaneously by using VO. Combining experimental and density-functional theory studies of Eu0.5Ba0.5TiO3−δ, we demonstrate that oxygen vacancies create Ti3+3d1 defect states, mediating the ferromagnetic coupling between the localized Eu 4f7 spins, and increase an off-center displacement of Ti ions, enhancing the ferroelectric Curie temperature. The dual function of Ti sites also promises a magnetoelectric coupling in the Eu0.5Ba0.5TiO3−δ.
Topologically nontrivial field configurations called "baby skyrmions" behave like particles and give origins to the field of skyrmionics that promises racetrack memory and other technological applications. U...
详细信息
Recent advances in cancer research largely rely on new developments in microscopic or molecular profiling techniques offering high level of detail with respect to either spatial or molecular features, but usually not ...
详细信息
We report results of an investigation of the temperature dependence of the magnon and phonon frequencies in NiO. A combination of Brillouin-Mandelstam and Raman spectroscopies allowed us to elucidate the evolution of ...
详细信息
Currently, there exist a lot of challenges in the transportation scope that researcher are trying to resolve and one of them can be focused on transportation planning. The main contribution of this paper was the desig...
详细信息
ISBN:
(纸本)9781509024377
Currently, there exist a lot of challenges in the transportation scope that researcher are trying to resolve and one of them can be focused on transportation planning. The main contribution of this paper was the design and implementation of an ITS smart sensor prototype that incorporates and combine the Internet of Things (IoT) and Bigdata approaches in order to produce ITS cloud services for helping transportation planning for Bus Rapid Transit (BRT) systems. The ITS smart sensor prototype is capable of detecting several Bluetooth signals belonging to several devices (for instance from mobile phones) that people uses into the BRT (for instance, in Bogota city) system. As from that information, the ITS smart sensor prototype can create the O/D (origin/Destiny) Matrix for several BRT routes and this information can be used by the Administrator Authorities (AA) in order to produce a suitable transportation planning for the BRT systems. In addition, that information can be used by AA as from cloud services.
The learners' needs are an important factor in designing syllabus and materials design, this research deals with the syllabus and material design based on the professional's needs. It is expected that the syll...
The learners' needs are an important factor in designing syllabus and materials design, this research deals with the syllabus and material design based on the professional's needs. It is expected that the syllabus and materials designes are communicatively applicable to the professional academy. Descriptive method is applied in this research. The sample of this research is 30 students of ATII Immanuel Academy Medan. They were selected by random sampling to get the data, the questioners were administered to students. the questioners consisted of 54 items and the semi structured interview consisted of 5 questions, the finding indicated that learners' needs were focused on reading and speaking skills. With reference appropriately and proportionally derived for students of the Professional Academy. Further on the basis of the syllabus, materials are designed in which the skills of using language become a priority. The results of this research will be disseminated using the website
Our quantitative understanding of how scientists choose and shift their research focus over time is highly consequential, because it affects the ways in which scientists are trained, science is funded, knowledge is or...
暂无评论