Multilevel Algebraic Approach for Performance Analysis of Parallel Algorithms

Self Cite

Summary Large‐scale problems in engineering and science often require the solution of sparse linear algebra problems and the Krylov subspace iteration methods (KM) have led to a major change in how users deal with them. But, for these solvers to use extreme‐scale hardware efficiently a lot of work was spent to redesign both the KM algorithms and their implementations to address challenges like extreme concurrency, complex memory hierarchies, costly data movement, and heterogeneous node architectures. All the redesign approaches bases the KM algorithm on block‐based strategies which lead to the Block‐KM (BKM) algorithm which has high granularity (i.e., the ratio of computation time to communication time). The work proposes novel parallel revisitation of the modules used in BKM which are based on the overlapping of communication and computation. Such revisitation is evaluated by a model of their granularity and verified on the basis of a case study related to a classical problem from numerical linear algebra.

Section: Discussionmentioning

confidence: 99%

About the granularity portability of block‐based Krylov methods in heterogeneous computing environments

Carracciuolo

Mele

Szustak

2020

Self Cite

“…In this context, there is a need for a more formal analysis of the expected performances, that can help to understand the choices, identify bottlenecks, and drive directions of optimizations. Examples of relevant references concerning the performance analysis and prediction to support the design process can be found in articles 29‐32 …”

Section: Related Workmentioning

confidence: 99%

Exploration of OpenCL Heterogeneous Programming for Porting Solidification Modeling to CPU‐GPU Platforms

Halbiniak

Szustak

Olas

et al. 2020

Summary This article provides a comprehensive study of OpenCL heterogeneous programming for porting applications to CPU–GPU computing platforms, with a real‐life application for the solidification modeling. The aim is to achieve a flexible workload distribution between available CPU–GPU resources and optimize application performance. Considering the solidification application as a use case, we explore the necessary steps required for (i) adaptation of an application to CPU–GPU platforms, and (ii) mapping the application workload onto the OpenCL programming model. The adaptation is based on a reformulation of steps developed previously for CPU–MIC architectures. The mapping process allows us to utilize OpenCL for harnessing CPU and GPU cores using data parallelism, as well as for the management of available compute devices with task parallelism. The resulting OpenCL code's performance and energy efficiency is experimentally studied for two platforms with powerful GPUs of various generations (with Kepler and Volta architectures). The experiments confirm the performance advantage of using computing resources of both GPUs and CPUs. The achieved benefit depends on the relationship between the computing power of CPUs and GPUs. Moreover, this gain entails the growth of the average power that increases the energy consumed during the application execution.

“…The authors define the upper limit as the maximum number of independent sub‐algorithms that can be executed simultaneously in the architecture and the lower limit is provided by the dependency degree. The performance prediction model proposed in Reference 13 was employed for the MGRIT (MUlti‐Grid In Time) algorithm 14 and a simplified version of the model was applied for a matrix‐multiply algorithm 15 …”

Section: Longest Common Subsequencementioning

confidence: 99%

“…We used the approach of Reference 13 (section 2.2) partially, with a focus on expressing the dependencies of the LCS problem and determining the complexity of the algorithm we employed to solve it. The LCS dependencies are shown in Equation ) and illustrated for the 4 × 4 case in Figure 3A,B.…”

Section: Proposed Architecturementioning

confidence: 99%

“…More formally, the dependencies can be expressed as 13 : (a) ∀( i , j ) A i , j ← A i − 1, j and (b) ∀( i , j ) A i , j ← A i , j − 1 . Using the notation of Reference 13, the independent elements are expressed as ∀( i , j , k , l ) A i , j ↮ A k , l if ( i + j ) = ( k + l ). Please note that this last condition states that the anti‐diagonal elements are independent.…”

Section: Proposed Architecturementioning

confidence: 99%

See 1 more Smart Citation

A CPU‐FPGA heterogeneous approach for biological sequence comparison using high‐level synthesis

Jorge

Nery

Melo

et al. 2020

Summary This article presents a high‐level synthesis implementation of the longest common subsequence (LCS) algorithm combined with a weighted‐based scheduler for comparing biological sequences prioritizing energy consumption or execution time. The LCS algorithm has been thoroughly tailored using Vivado High‐Level Synthesis tool, which is able to synthesize register transfer level (RTL) from high‐level language descriptions, such as C/C++. Performance and energy consumption results were obtained with a CPU Intel Core i7‐3770 CPU and an Alpha‐Data ADM‐PCIE‐KU3 board that has a Xilinx Kintex UltraScale XCKU060 FPGA chip. We executed a batch of 20 comparisons of sequences on 10k, 20k, and 50k sizes. Our experiments showed that the energy consumption on the combined approach was significantly lower when compared to the CPU, achieving 75% energy reduction on 50k comparisons. We also used the tool proposed in this article to do a case study on Covid‐19, with real SARS‐CoV‐2 sequences, comparing their LCS scores.