[vc_empty_space][vc_empty_space]
Processor Selection for OpenCL Kernels using KNN Algorithm
Rahmawan H.a,b, Kuspriyantoa, Mutijarsa K.a
a School of Electrical Engineering and Informatics, Institut Teknologi Bandung, Bandung, 40132, Indonesia
b Department of Computer Science, Faculty of Mathematics and Natural Science, Bogor Agricultural University, Bogor, 16680, Indonesia
[vc_row][vc_column][vc_row_inner][vc_column_inner][vc_separator css=”.vc_custom_1624529070653{padding-top: 30px !important;padding-bottom: 30px !important;}”][/vc_column_inner][/vc_row_inner][vc_row_inner layout=”boxed”][vc_column_inner width=”3/4″ css=”.vc_custom_1624695412187{border-right-width: 1px !important;border-right-color: #dddddd !important;border-right-style: solid !important;border-radius: 1px !important;}”][vc_empty_space][megatron_heading title=”Abstract” size=”size-sm” text_align=”text-left”][vc_column_text]© 2018 IEEE.Heterogeneous computing has been one of the mainstreams in the area of high-performance computing. One type of heterogeneous computer is a computer with CPUs, GPUs, and is programmed using OpenCL software framework. To achieve its best performance, an OpenCL kernel should be executed on the right processor. However, the OpenCL framework requires programmers to select processors for their workload. Thus, programmers need to understand which processor is best for their OpenCL kernels. If programmers did not aware about their kernels and the capabilities available processors, then the kernels are potentially not executed on the best processor. To overcome the problem, this research proposes an approach that selects the best processor to execute an OpenCL kernel. The processor selection is accomplished by k-nearest neighbor (KNN) algorithm. In order to achieve acceptable accuracy, the features to be used with KNN are selected beforehand. The selection of features involves two models, namely filter model, and wrapper model. This research uses n-fold cross-validation to evaluate the performance of the proposed method. According to the experiments, on selecting best processors for workloads on four processors, the approach achieved the accuracy of 91.5% to 100%. The result of the experiments also concludes that the features with the most significant contribution are vector integer operations and vector global memory access, or vector integer operations and vector global memory access.[/vc_column_text][vc_empty_space][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][vc_empty_space][megatron_heading title=”Author keywords” size=”size-sm” text_align=”text-left”][vc_column_text]Heterogeneous computers,Heterogeneous computing,High performance computing,Integer operations,K nearest neighbor algorithm,OpenCL,Processor selection,Software frameworks[/vc_column_text][vc_empty_space][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][vc_empty_space][megatron_heading title=”Indexed keywords” size=”size-sm” text_align=”text-left”][vc_column_text]Classification,Heterogeneous computing,KNN,OpenCL,Processor selection[/vc_column_text][vc_empty_space][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][vc_empty_space][megatron_heading title=”Funding details” size=”size-sm” text_align=”text-left”][vc_column_text][/vc_column_text][vc_empty_space][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][vc_empty_space][megatron_heading title=”DOI” size=”size-sm” text_align=”text-left”][vc_column_text]https://doi.org/10.1109/ICITSI.2018.8696079[/vc_column_text][/vc_column_inner][vc_column_inner width=”1/4″][vc_column_text]Widget Plumx[/vc_column_text][/vc_column_inner][/vc_row_inner][/vc_column][/vc_row][vc_row][vc_column][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][/vc_column][/vc_row]