A Multi-Node GPGPU Implementation of Non-Linear Anisotropic Diffusion Filter

Pallipuram, Vivek K.; Raut, Nimisha; Ren, Xiaodan; Smith, Melinda; Naik, Sumedh

doi:10.1109/saahpc.2012.11

Cited by 3 publications

(6 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Because these operations are data‐parallel, they are performed on the GPGPU devices using two separate GPGPU kernels, namely the median_kernel and PDE_kernel , whereas the CPU‐host processors perform communication operations such as row exchange and serial processing such as image padding. This task distribution between the CPU‐cores and GPGPU devices enables optimal scaling of the ADF algorithm on multi‐GPGPU systems .…”

Section: Experimental Set‐up Synchronous Iterative Algorithms Mappinmentioning

confidence: 99%

“…In the fourth and final stage, the master process (MPI rank 0) gathers the output image tiles from all other processes and constructs the final output image. A detailed description of the multi‐GPGPU orchestration for the ADF algorithm is provided in .…”

Section: Experimental Set‐up Synchronous Iterative Algorithms Mappinmentioning

confidence: 99%

“…The regression‐based framework primarily targets deterministic synchronous iterative algorithms (SIAs) using our synchronous iterative GPGPU execution (SIGE) model, which serves as the backbone for the proposed framework. SIAs are found throughout scientific computing including but not limited to neural network simulations , stencil‐based image processing operations , 2D PDF estimation , and bio‐molecular dynamics . The regression‐based framework is broken into two primary components: computation and communication.…”

Section: Introductionmentioning

confidence: 99%

“…The SIA case studies used in this research to validate the modeling framework are large‐scale spiking neural network (SNN) simulations and anisotropic diffusion filter (ADF) for massive images . These applications offer a range of communication and computation requirements enabling validation of a large design space.…”

Section: Introductionmentioning

confidence: 99%

“…The communication component of the regressionbased framework models the throughput performance of the following two communication sub-components: (i) inter-processor communication over Infiniband and (ii) CPU-host/GPGPUdevice (host-device) communication over PCI-Ex bus. The regression models for these subcomponents are developed using micro-benchmarks that measure transaction throughput and employ data transfer size and processor count as predictor variables.The SIA case studies used in this research to validate the modeling framework are large-scale spiking neural network (SNN) simulations [9] and anisotropic diffusion filter (ADF) for massive images [10]. These applications offer a range of communication and computation requirements enabling validation of a large design space.…”

mentioning

confidence: 99%

See 4 more Smart Citations

A regression‐based performance prediction framework for synchronous iterative algorithms on general purpose graphical processing unit clusters

Pallipuram

Smith

Raut

et al. 2013

Concurrency and Computation

Self Cite

View full text Add to dashboard Cite

SUMMARYHeterogeneous performance prediction models are valuable tools to accurately predict application runtime, allowing for efficient design space exploration and application mapping. The existing performance models require intricate system architecture knowledge, making the modeling task difficult. In this research, we propose a regression‐based performance prediction framework for general purpose graphical processing unit (GPGPU) clusters that statistically abstracts the system architecture characteristics, enabling performance prediction without detailed system architecture knowledge. The regression‐based framework targets deterministic synchronous iterative algorithms using our synchronous iterative GPGPU execution model and is broken into two components: the computation component that models the GPGPU device and host computations and the communication component that models the network‐level communications. The computation component regression models use algorithm characteristics such as the number of floating‐point operations and total bytes as predictor variables and are trained using several small, instrumented executions of synchronous iterative algorithms that include a range of floating‐point operations‐to‐byte requirements. The regression models for network‐level communications are developed using micro‐benchmarks and employ data transfer size and processor count as predictor variables. Our performance prediction framework achieves prediction accuracy over 90% compared with the actual implementations for several tested GPGPU cluster configurations. The end goal of this research is to offer the scientific computing community, an accurate and easy‐to‐use performance prediction framework that empowers users to optimally utilize the heterogeneous resources. Copyright © 2013 John Wiley & Sons, Ltd.

show abstract

Section: Experimental Set‐up Synchronous Iterative Algorithms Mappinmentioning

confidence: 99%

Section: Experimental Set‐up Synchronous Iterative Algorithms Mappinmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

mentioning

confidence: 99%

See 3 more Smart Citations

A regression‐based performance prediction framework for synchronous iterative algorithms on general purpose graphical processing unit clusters

Pallipuram

Smith

Raut

et al. 2013

Concurrency and Computation

Self Cite

View full text Add to dashboard Cite

show abstract

Subjective versus objective: classifying analytical models for productive heterogeneous performance prediction

et al. 2014

View full text Add to dashboard Cite

Optimal Piecewise Linear Function Approximation for GPU-Based Applications

Berjón

Gallego

Cuevas

et al. 2016

IEEE Trans. Cybern.

View full text Add to dashboard Cite

Many computer vision and human-computer interaction applications developed in recent years need evaluating complex and continuous mathematical functions as an essential step toward proper operation. However, rigorous evaluation of these kind of functions often implies a very high computational cost, unacceptable in real-time applications. To alleviate this problem, functions are commonly approximated by simpler piecewise-polynomial representations. Following this idea, we propose a novel, efficient, and practical technique to evaluate complex and continuous functions using a nearly optimal design of two types of piecewise linear approximations in the case of a large budget of evaluation subintervals. To this end, we develop a thorough error analysis that yields asymptotically tight bounds to accurately quantify the approximation performance of both representations. It provides an improvement upon previous error estimates and allows the user to control the tradeoff between the approximation error and the number of evaluation subintervals. To guarantee real-time operation, the method is suitable for, but not limited to, an efficient implementation in modern graphics processing units, where it outperforms previous alternative approaches by exploiting the fixed-function interpolation routines present in their texture units. The proposed technique is a perfect match for any application requiring the evaluation of continuous functions; we have measured in detail its quality and efficiency on several functions, and, in particular, the Gaussian function because it is extensively used in many areas of computer vision and cybernetics, and it is expensive to evaluate.

show abstract

A Multi-Node GPGPU Implementation of Non-Linear Anisotropic Diffusion Filter

Cited by 3 publications

References 13 publications

A regression‐based performance prediction framework for synchronous iterative algorithms on general purpose graphical processing unit clusters

A regression‐based performance prediction framework for synchronous iterative algorithms on general purpose graphical processing unit clusters

Subjective versus objective: classifying analytical models for productive heterogeneous performance prediction

Optimal Piecewise Linear Function Approximation for GPU-Based Applications

Contact Info

Product

Resources

About