The efficiency of ubiquitous SIMD (Single Instruction Multiple Data) media processors is seriously limited by the bottleneck effect of the scalar kernels in media applications. To solve this problem, a dual-core framework, composed of a micro control unit and an instruction buffer, is proposed. This framework can dynamically decouple the scalar and vector pipelines of the original single-core SIMD architecture into two free-running cores. Thus, the bottleneck effect can be eliminated by effectively exploiting the parallelism between scalar and vector kernels. The dual-core framework achieves the best attributes of both single-core and dual-core SIMD architectures. Experimental results exhibit an average performance improvement of 33%, at an area overhead of 4.26%. What's more, with the increase of the SIMD width, higher performance gain and lower cost can be expected.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Yaohua WANG, Shuming CHEN, Hu CHEN, Jianghua WAN, Kai ZHANG, Sheng LIU, "Dual-Core Framework: Eliminating the Bottleneck Effect of Scalar Kernels on SIMD Architectures" in IEICE TRANSACTIONS on Information,
vol. E96-D, no. 2, pp. 365-369, February 2013, doi: 10.1587/transinf.E96.D.365.
Abstract: The efficiency of ubiquitous SIMD (Single Instruction Multiple Data) media processors is seriously limited by the bottleneck effect of the scalar kernels in media applications. To solve this problem, a dual-core framework, composed of a micro control unit and an instruction buffer, is proposed. This framework can dynamically decouple the scalar and vector pipelines of the original single-core SIMD architecture into two free-running cores. Thus, the bottleneck effect can be eliminated by effectively exploiting the parallelism between scalar and vector kernels. The dual-core framework achieves the best attributes of both single-core and dual-core SIMD architectures. Experimental results exhibit an average performance improvement of 33%, at an area overhead of 4.26%. What's more, with the increase of the SIMD width, higher performance gain and lower cost can be expected.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E96.D.365/_p
Copy
@ARTICLE{e96-d_2_365,
author={Yaohua WANG, Shuming CHEN, Hu CHEN, Jianghua WAN, Kai ZHANG, Sheng LIU, },
journal={IEICE TRANSACTIONS on Information},
title={Dual-Core Framework: Eliminating the Bottleneck Effect of Scalar Kernels on SIMD Architectures},
year={2013},
volume={E96-D},
number={2},
pages={365-369},
abstract={The efficiency of ubiquitous SIMD (Single Instruction Multiple Data) media processors is seriously limited by the bottleneck effect of the scalar kernels in media applications. To solve this problem, a dual-core framework, composed of a micro control unit and an instruction buffer, is proposed. This framework can dynamically decouple the scalar and vector pipelines of the original single-core SIMD architecture into two free-running cores. Thus, the bottleneck effect can be eliminated by effectively exploiting the parallelism between scalar and vector kernels. The dual-core framework achieves the best attributes of both single-core and dual-core SIMD architectures. Experimental results exhibit an average performance improvement of 33%, at an area overhead of 4.26%. What's more, with the increase of the SIMD width, higher performance gain and lower cost can be expected.},
keywords={},
doi={10.1587/transinf.E96.D.365},
ISSN={1745-1361},
month={February},}
Copy
TY - JOUR
TI - Dual-Core Framework: Eliminating the Bottleneck Effect of Scalar Kernels on SIMD Architectures
T2 - IEICE TRANSACTIONS on Information
SP - 365
EP - 369
AU - Yaohua WANG
AU - Shuming CHEN
AU - Hu CHEN
AU - Jianghua WAN
AU - Kai ZHANG
AU - Sheng LIU
PY - 2013
DO - 10.1587/transinf.E96.D.365
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E96-D
IS - 2
JA - IEICE TRANSACTIONS on Information
Y1 - February 2013
AB - The efficiency of ubiquitous SIMD (Single Instruction Multiple Data) media processors is seriously limited by the bottleneck effect of the scalar kernels in media applications. To solve this problem, a dual-core framework, composed of a micro control unit and an instruction buffer, is proposed. This framework can dynamically decouple the scalar and vector pipelines of the original single-core SIMD architecture into two free-running cores. Thus, the bottleneck effect can be eliminated by effectively exploiting the parallelism between scalar and vector kernels. The dual-core framework achieves the best attributes of both single-core and dual-core SIMD architectures. Experimental results exhibit an average performance improvement of 33%, at an area overhead of 4.26%. What's more, with the increase of the SIMD width, higher performance gain and lower cost can be expected.
ER -