Source code transformations and optimizations

Cardoso, João M. P.; Coutinho, J.G.F.; Diniz, Pedro C.

doi:10.1016/b978-0-12-804189-5.00005-3

Cited by 9 publications

(4 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The first step of the proposed method includes the modification of the application source code using the loop fusing technique 33 . Listing 5 presents the modified code for a single time step of the solidification application.…”

Section: Optimizations Of Solidification Application With Dynamic Intensity Of Computations For Multi‐/manycore Architecturesmentioning

confidence: 99%

Dynamic workload prediction and distribution in numerical modeling of solidification on multi‐/manycore architectures

Halbiniak

Olas

Szustak

et al. 2020

Concurrency and Computation

View full text Add to dashboard Cite

Summary This work is a part of the global tendency to use modern computing systems for modeling the phase‐field phenomena. The main goal of this article is to improve the performance of a parallel application for the solidification modeling, assuming the dynamic intensity of computations in successive time steps when calculations are performed using a carefully selected group of nodes in the grid. A two‐step method is proposed to optimize the application for multi‐/manycore architectures. In the first step, the loop fusion is used to execute all kernels in a single nested loop and reduce the number of conditional operators. These modifications are vital to implementing the second step, which includes an algorithm for the dynamic workload prediction and load balancing across cores of a computing platform. Two versions of the algorithm are proposed—with the 1D and 2D maps used for predicting the computational domain within the grid. The proposed optimizations allow increasing the application performance significantly for all tested configurations of computing resources. The highest performance gain is achieved for two Intel Xeon Platinum 8180 CPUs, where the new code based on the 2D map yields the speedup of up to 2.74 times, while the usage of the proposed method with the 2D map for a single KNL accelerator permits reducing the execution time up to 1.91 times.

show abstract

Section: Optimizations Of Solidification Application With Dynamic Intensity Of Computations For Multi‐/manycore Architecturesmentioning

confidence: 99%

Dynamic workload prediction and distribution in numerical modeling of solidification on multi‐/manycore architectures

Halbiniak

Olas

Szustak

et al. 2020

Concurrency and Computation

View full text Add to dashboard Cite

show abstract

“…It is a systematic and efficient method that enables the study of the relationship between multiple input and output variables. 28 The design parameters of the H-rotor Darrieus VAWT were optimized using the DOE techniques. Based on the Box-Behnken (DOE) technique, a design matrix was formulated and datasets retrieved from the Q-Blade open-source software were employed in the DOE to determine the optimal configurations.…”

Section: Introductionmentioning

confidence: 99%

Determination of efficient configurations of vertical axis wind turbine using design of experiments

Ahmad

Shahzad

Akram

et al. 2022

Proceedings of the Institution of Mechanical Engineers, Part A:

View full text Add to dashboard Cite

This study aims to optimize the design parameters of a vertical axis wind turbine using design of experiments (DOE). Response surface methodology combined with desirability optimization was used to optimize multiple parameters such as the chord length, number of blades, aspect ratio, and pitch angle. The range of contributing parameters was selected using one factor at a time (OFAT) approach, and the performance parameters were evaluated by applying the double multiple streamtube theory. Datasets retrieved from the Q-Blade open-source software were employed in the DOE to determine the optimal configurations. Based on response surface investigation, a quadratic model was established and the accuracy of the model was determined by applying analysis of variance, goodness of fit, normal plot of residuals, and R-squared values. The optimized chord length, number of blades, pitch angle, and aspect ratio were 0.546 m, 03, −2.82°, and 0.808, respectively. The maximum power coefficient of 0.45 was obtained at a tip speed ratio of 3 from the optimized design parameters. The analysis revealed an increase of approximately 8.43% in the maximum power coefficient of the proposed wind turbine. A 3D unsteady computational fluid dynamics model along with Reynolds-Averaged Navier-Stokes equations were utilized to verify the obtained results, and a maximum difference of less than 1.5% was found. A standard H-rotor Darrieus configuration obtained from the OFAT approach was also tested at different wind speeds for comparison. The analysis revealed that when the wind speed was less than 3.85 m/sec, the standard Darrieus produced no power; however, the self-starting speed of the optimized VAWT was as low as 3.22 m/sec. This novel design methodology provides guidelines for obtaining optimum design configurations and can help in achieving high-fidelity analysis in the early design phase.

show abstract

“…It is often used in combination with loop unrolling and this combination together is a better optimization technique. Loop unrolling is a common loop transformation technique in which the loop body is repeated K times and the loop iteration space is minimised by unrolling a loop by a factor of K [13]. Fig.…”

Section: Power Optimization Techniques and Their Limitationsmentioning

confidence: 99%

“…12 shows a loop unrolled by a factor of 2. Figure 12: loop unrolling by a factor of 2, Source: [13] The importance of software pipelining is highlighted by the wide range of architectures that benefit from it. There are a variety of algorithms for achieving software pipelining, and a comparison is presented in [3].…”

Section: Power Optimization Techniques and Their Limitationsmentioning

confidence: 99%

Multi-Layer Approximate Computing

Brar¹

View full text Add to dashboard Cite

Source code transformations and optimizations

Cited by 9 publications

References 13 publications

Dynamic workload prediction and distribution in numerical modeling of solidification on multi‐/manycore architectures

Dynamic workload prediction and distribution in numerical modeling of solidification on multi‐/manycore architectures

Determination of efficient configurations of vertical axis wind turbine using design of experiments

Multi-Layer Approximate Computing

Contact Info

Product

Resources

About