China
Africa
Explore chapters and articles related to this topic
A Systematic Review of 3D Imaging in Biomedical Applications
Published in K.C. Santosh, Sameer Antani, D.S. Guru, Nilanjan Dey, Medical Imaging, 2019
Darshan D. Ruikar, Dattatray D. Sawat, K.C. Santosh
Recent studies show that researchers prefer hardware accelerators along with software libraries for their visualization experiments. Weinrich et al. [33,65] used CUDA and OpenGL‡, along with two GPUs, namely, GeForce 8800 GTX and Quadro fx 5600, for their visualization experiments. The speed was increased up to 148 times higher than CPU. These were earlier versions of CUDA and OpenGL. At the time of writing this paper, NVIDIA has released a new CUDA version: CUDA 10.0. Recent improvements in CUDA include support for various libraries such as deep learning for medical images [34]. Libraries such as cuDNN, GIE, cuBLAS, cuSPARSE, and NCCL support deep learning in CUDA. The 4D processing of medical images is supported by some of the CUDA libraries (CUFFT and NPP). The present study demonstrated latencies in the visualization pipeline using GPU [35]. In this study, the visualization algorithm is designed using3Dtexture mapping (3DTM), software-based raycasting (SOFTRC), and hardware-accelerated raycasting (HWRC). They used three GPUs to compare the performance gain. Recently, NVIDIA introduced an RTX* feature in its flagship GPU RTX2080, which is beneficial for direct as well as indirect volume rendering [71].
Hardware implementation
Published in Tomoyoshi Shimobaba, Tomoyoshi Ito, Computer Holography, 2019
Tomoyoshi Shimobaba, Tomoyoshi Ito
To the best of our knowledge, the first use of a graphics processing machine (graphics workstation) in a hologram calculation was reported by Ref. [164]. Since then, NVIDIA Geforce 4, Quadro and GeForce6 have been adopted in hologram calculations [165–167].
Proposal and evaluation of adjusting resource amount for automatically offloaded applications
Published in Cogent Engineering, 2022
Regarding GPU, I use two boards of NVIDIA Tesla T4 (CUDA core: 2560, Memory: GDDR6 16GB) and NVIDIA Quadro P4000 (CUDA core: 1792, Memory: GDDR5 8GB). I use CUDA Toolkit 10.1 and PGI compiler 19.10 for GPU control. NVIDIA vGPU virtual Compute Server virtualizes GPU resources. Using vGPU, Tesla T4 resources are divided, and resources of 1 board are divided into 1, 2, and 4 parts. Kernel-based VM (KVM) of RHEL7.9 is used for CPU virtualization. The resource of the VM that becomes 1 standard size is 2 cores and 16GB RAM. Half size (1 core), standard size (2 cores), and double size (4 cores) can be selected. For example, when setting the CPU and GPU resources with standard sizes at a time, our implementation virtualizes CPU and GPU resources and links the 2-core CPU and Tesla T4 1 board. Minimum unit sizes are 1-core CPU and 1/4 GPU board. Figure 3 shows the experimental environment and specifications. Here, the application code used by the user is specified from the client notebook PC, tuned using the bare metal verification machine, and then deployed to the virtual running environment for the actual use.
Network-wide ride-sourcing passenger demand origin-destination matrix prediction with a generative adversarial network
Published in Transportmetrica A: Transport Science, 2022
Changlin Li, Liang Zheng, Ning Jia
All numerical experiments are implemented in a GPU computer workstation equipped with Xeon 4110 central processing unit, 256 GB memory, and a Quadro P4000 graphics processing unit. The number of backtracking cycles T is set as 4 for all deep learning-based methods. Random search is applied to optimise the hyper-parameters of all prediction methods, which outperforms other strategies (e.g. grid search and manual search) (Bergstra and Bengio 2012). The prediction accuracy is evaluated by two well-known measurements: mean square error (MSE) and mean absolute percentage error (MAPE), formulated respectively as where Q is the number of prediction cycles, whose value is 332.
Study and evaluation of automatic GPU offloading method from various language applications
Published in International Journal of Parallel, Emergent and Distributed Systems, 2022
Regarding GPU, we use NVIDIA GeForce RTX 2080 Ti and NVIDIA Quadro K5200. To control the GPU, CUDA Toolkit 10.1 is commonly used. We use PGI compiler 19.10, which is an OpenACC compiler for C language, pyCUDA 2019.1.2 and Cupy 7.8 for Python (which are CUDA interpreter and CUDA calling library), and JCuda 10.1 for Java, which binds Java with the CUDA runtime libraries.