Stream Processing on Multi-cores with GPUs: Parallel Programming Models' Challenges

Rockenbach, Dinei; Stein, C. Michael; Griebler, Dalvan; Mencagli, Gabriele; Torquati, Massimo; Danelutto, Marco; Fernandes, Luiz Gustavo

doi:10.1109/ipdpsw.2019.00137

Cited by 9 publications

(8 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In our previous work, 5 we identified the need for creating microbatches of stream items to properly exploit many‐core accelerators like GPUs with SPar. This raised the problem of properly identifying the microbatch size to improve the performance of stream processing applications.…”

Section: Adaptive Microbatching For Stream Processing On Gpusmentioning

confidence: 99%

“…When such specialized coprocessors are not available, graphics processing units (GPUs) still represent interesting candidates for offloading data compression tasks, also because they are becoming popular in IoT embedded systems 4 . However, the efficient exploitation of GPU devices is considered challenging in the context of stream processing, 5 in particular when the target performance metric to optimize is latency. This is because GPUs generally work well when a high volume of data is available at the same time, condition that in stream processing applications can be obtained by properly batching input data before processing them as a whole.…”

Section: Introductionmentioning

confidence: 99%

“…This approach has been pursued in our previous work for the Lempel‐Ziv‐Storer‐Szymanski (LZSS) compression algorithm 8 . We observed that the size of the offloaded batch plays an important role having a significant impact both on the system throughput and on its end‐to‐end latency 5 . The ideal batch size depends on several factors, among them, the current workload and the input data rate.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Latency‐aware adaptive micro‐batching techniques for streamed data compression on graphics processing units

Stein¹,

Rockenbach

Griebler

et al. 2020

Concurrency and Computation

Self Cite

View full text Add to dashboard Cite

Summary Stream processing is a parallel paradigm used in many application domains. With the advance of graphics processing units (GPUs), their usage in stream processing applications has increased as well. The efficient utilization of GPU accelerators in streaming scenarios requires to batch input elements in microbatches, whose computation is offloaded on the GPU leveraging data parallelism within the same batch of data. Since data elements are continuously received based on the input speed, the bigger the microbatch size the higher the latency to completely buffer it and to start the processing on the device. Unfortunately, stream processing applications often have strict latency requirements that need to find the best size of the microbatches and to adapt it dynamically based on the workload conditions as well as according to the characteristics of the underlying device and network. In this work, we aim at implementing latency‐aware adaptive microbatching techniques and algorithms for streaming compression applications targeting GPUs. The evaluation is conducted using the Lempel‐Ziv‐Storer‐Szymanski compression application considering different input workloads. As a general result of our work, we noticed that algorithms with elastic adaptation factors respond better for stable workloads, while algorithms with narrower targets respond better for highly unbalanced workloads.

show abstract

Section: Adaptive Microbatching For Stream Processing On Gpusmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Latency‐aware adaptive micro‐batching techniques for streamed data compression on graphics processing units

Stein¹,

Rockenbach

Griebler

et al. 2020

Concurrency and Computation

Self Cite

View full text Add to dashboard Cite

show abstract

“…In our previous work, we evaluated different parallel programming models when implementing stream and data parallelism combined [19]. One lesson learned is that fine-grained stream processing may not generate enough workload to properly exploit massively parallel architectures such as GPUs.…”

Section: New Attributes For Sparmentioning

confidence: 99%

High-Level Stream Parallelism Abstractions with SPar Targeting GPUs

Griebler

Fernandes

Danelutto

et al. 2020

Parallel Computing: Technology Trends

View full text Add to dashboard Cite

The combined exploitation of stream and data parallelism is demonstrating encouraging performance results in the literature for heterogeneous architectures, which are present on every computer systems today. However, provide parallel software efficiently targeting those architectures requires significant programming effort and expertise. The SPar domain-specific language already represents a solution to this problem providing proven high-level programming abstractions for multi-core architectures. In this paper, we enrich the SPar language adding support for GPUs. New transformation rules are designed for generating parallel code using stream and data parallel patterns. Our experiments revealed that these transformations rules are able to improve performance while the high-level programming abstractions are maintained.

show abstract

“…Under these circumstances, traditional manual parallelization methods pose great challenges to researchers, especially in terms of increasing researchers' practice cycle and learning costs. 9,10 The MIC architecture was designed for HPC by the Intel ® Corporation. Based on platforms of Intel® integrated multicores, MIC has more cores than other commonly used multicore processors.…”

mentioning

confidence: 99%

Preliminary study on the automatic parallelism optimization model for image enhancement algorithms based on Intel's® Xeon Phi

Huang

Yang

Jian

et al. 2021

Concurrency and Computation

View full text Add to dashboard Cite

In unmanned aerial vehicle (UAV) image-processing applications, one needs to implement different parallel image-enhancement algorithms on several high-performance computing platforms utilizing various programming models. To speed up the parallelization procedure and improve its efficiency, the automatic parallel software package, Par4All, is applied in this work. We find that the performance of the original automatic parallelization algorithm produced with Par4All is inefficient. To resolve this problem, we propose different optimization approaches for Par4All based on Intel®'s Xeon Phi high-performance computing platform that are based on the structural features of the image-enhancement algorithms, which can further optimize the original parallel algorithm. These approaches mainly include: (1) Par4All automatic parallel search module optimization, (2) dynamic thread setting optimization, and (3) the collaborative parallelization of both CPU and many integrated core (MIC) processors.According to the results of the comparison experiments involving different algorithms, it is shown that the proposed optimization approaches for these kinds of algorithms can significantly improve the performance of automatic parallel algorithms. The acceleration ratio increases approximately by 30%, 70%, and 80% for the multiscale Retinex, Gaussian-filtering and median-filtering algorithms, respectively. As continuation and deepening of our previous research work, this research has the potential to be beneficial for other researchers in image-processing applications with image-enhancement algorithms. K E Y W O R D Sautomatic parallelism, image-enhancement algorithms, Par4All, Intel® Xeon Phi, unmanned aerial vehicles INTRODUCTIONAs mentioned by References 1, it is known that some image-enhancement algorithms, for example, median filtering, Gaussian filtering, wavelet transform, and the multiscale Retinex (MSR) algorithm, are used when preprocessing large amounts of unmanned aerial vehicle (UAV) remote-sensing (RS) images to solve the problems of poor clarity, insufficient contrast and low adaptability. In addition, many-cores computing platforms, e.g., graphics processing units (GPUs) and many integrate cores (MICs), have been involved in the development of some parallel algorithms to speed up the processing times 2 ; such approaches are expected to become mainstream in the near future. Experimental results have shown that parallelized algorithms can significantly improve computation speed and provide better acceleration efficiency. [3][4][5] Usually, there are different kinds of serial

show abstract

Stream Processing on Multi-cores with GPUs: Parallel Programming Models' Challenges

Cited by 9 publications

References 14 publications

Latency‐aware adaptive micro‐batching techniques for streamed data compression on graphics processing units

Latency‐aware adaptive micro‐batching techniques for streamed data compression on graphics processing units

High-Level Stream Parallelism Abstractions with SPar Targeting GPUs

Preliminary study on the automatic parallelism optimization model for image enhancement algorithms based on Intel's® Xeon Phi

Contact Info

Product

Resources

About