TY - JOUR
T1 - Task allocation with algorithm transformation for reducing data-transfer bottlenecks in heterogeneous multi-core processors
T2 - A case study of HOG descriptor computation
AU - Waidyasooriya, Hasitha Muthumala
AU - Okumura, Daisuke
AU - Hariyama, Masanori
AU - Kameyama, Michitaka
PY - 2010/12
Y1 - 2010/12
N2 - Heterogeneous multi-core processors are attracted by the media processing applications due to their capability of drawing strengths of different cores to improve the overall performance. However, the data transfer bottlenecks and limitations in the task allocation due to the accelerator-incompatible operations prevents us from gaining full potential of the heterogeneous multi-core processors. This paper presents a task allocation method based on algorithm transformation to increase the freedom of task allocation. We use approximation methods such as CORDIC algorithms to map the accelerator- incompatible operations to accelerator cores. According to the experimental results using HOG descriptor computation, the proposed task allocation method reduces the data transfer time by more than 82% and the total processing time by more than 79% compared to the conventional task allocation method.
AB - Heterogeneous multi-core processors are attracted by the media processing applications due to their capability of drawing strengths of different cores to improve the overall performance. However, the data transfer bottlenecks and limitations in the task allocation due to the accelerator-incompatible operations prevents us from gaining full potential of the heterogeneous multi-core processors. This paper presents a task allocation method based on algorithm transformation to increase the freedom of task allocation. We use approximation methods such as CORDIC algorithms to map the accelerator- incompatible operations to accelerator cores. According to the experimental results using HOG descriptor computation, the proposed task allocation method reduces the data transfer time by more than 82% and the total processing time by more than 79% compared to the conventional task allocation method.
KW - Heterogeneous multi-core processor
KW - Systemon-chip
KW - Task-allocation
UR - http://www.scopus.com/inward/record.url?scp=78649993686&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=78649993686&partnerID=8YFLogxK
U2 - 10.1587/transfun.E93.A.2570
DO - 10.1587/transfun.E93.A.2570
M3 - Article
AN - SCOPUS:78649993686
SN - 0916-8508
VL - E93-A
SP - 2570
EP - 2580
JO - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
JF - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
IS - 12
ER -