Effective Utilization of CUDA Hyper-Q for Improved Power and Performance Efficiency

Luley, Ryan; Qiu, Qinru

doi:10.1109/ipdpsw.2016.154

Cited by 10 publications

(3 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hyper-Q feature supported by NVIDIA GPUs enables concurrent execution of multiple independent kernels on a single GPU. However, when the execution of multiple kernels are not properly ordered, contention for shared resources can degrade the overall performance (Luley & Qiu, 2016). A model has been developed by (Lázaro-Muñoz et al, 2017) that determines the order of kernel execution so as to increase the possibilities of simultaneously performing the data transfer and kernel execution, and reduce the total execution time.…”

Section: Related Workmentioning

confidence: 99%

Performance Enhancement of Cuda Applications by Overlapping Data Transfer and Kernel Execution

Raju¹,

Chiplunkar²

2021

acs

View full text Add to dashboard Cite

The CPU-GPU combination is a widely used heterogeneous computing system in which the CPU and GPU have different address spaces. Since the GPU cannot directly access the CPU memory, prior to invoking the GPU function the input data must be available on the GPU memory. On completion of GPU function, the results of computation are transferred to CPU memory. The CPU-GPU data transfer happens through PCI-Express bus. The PCI-E bandwidth is much lesser than that of GPU memory. The speed at which the data is transferred is limited by the PCI-E bandwidth. Hence, the PCI-E acts as a performance bottleneck. In this paper two approaches are discussed to minimize the overhead of data transfer, namely, performing the data transfer while the GPU function is being executed and reducing the amount of data to be transferred to GPU. The effectiveness of these approaches on the execution time of a set of CUDA applications is realized using CUDA streams. The results of our experiments show that the execution time of applications can be minimized with the proposed approaches.

show abstract

Section: Related Workmentioning

confidence: 99%

Performance Enhancement of Cuda Applications by Overlapping Data Transfer and Kernel Execution

Raju¹,

Chiplunkar²

2021

acs

View full text Add to dashboard Cite

show abstract

“…The DP structure allows the kernel to run in other kernels [6]. The use of HQ utilizes parallel paths on the host (CPU) in order to run concurrently on the device (GPU) kernel [7]. Finding BMU in the parallel SOM in this study is divided into 3 kernels.…”

Section: Finding Bmumentioning

confidence: 99%

“…The development of nVidia technology invented the Kepler generation GPU with the ability of Dynamic Parallel [6] and Hyper-Q [7]. Both of these capabilities yield a more concurrent computing and efficiency in utilizing GPU.…”

Section: Introductionmentioning

confidence: 99%

Enhancing Performance of Parallel Self-Organizing Map on Large Dataset with Dynamic Parallel and Hyper-Q

Sibero

Sitompul

Nasution

2018

jocai

View full text Add to dashboard Cite

Self-Organizing Map (SOM) is an unsupervised artificial neural network algorithm. Even though this algorithm is known to be an appealing clustering method,many efforts to improve its performance are still pursued in various research works. In order to gain faster computation time, for instance, running SOM in parallel had been focused in many previous research works. Utilization of the Graphics Processing Unit (GPU) as a parallel calculation engine is also continuously improved. However, total computation time in parallel SOM is still not optimal on processing large dataset. In this research, we propose a combination of Dynamic Parallel and Hyper-Q to further improve the performance of parallel SOM in terms of faster computing time. Dynamic Parallel and Hyper-Q are utilized on the process of calculating distance and searching best-matching unit (BMU), while updating weight and its neighbors are performed using Hyper-Q only. Result of this study indicates an increase in SOM parallel performance up to two times faster compared to those without using Dynamic Parallel and Hyper-Q.

show abstract

A Dynamic Acceleration Method for Remote Sensing Image Processing Based on CUDA

Zuo

Zhang

Qiao

et al. 2021

Simulation Tools and Techniques

View full text Add to dashboard Cite

Effective Utilization of CUDA Hyper-Q for Improved Power and Performance Efficiency

Cited by 10 publications

References 9 publications

Performance Enhancement of Cuda Applications by Overlapping Data Transfer and Kernel Execution

Performance Enhancement of Cuda Applications by Overlapping Data Transfer and Kernel Execution

Enhancing Performance of Parallel Self-Organizing Map on Large Dataset with Dynamic Parallel and Hyper-Q

A Dynamic Acceleration Method for Remote Sensing Image Processing Based on CUDA

Contact Info

Product

Resources

About