Page 26 - Fister jr., Iztok, and Andrej Brodnik (eds.). StuCoSReC. Proceedings of the 2016 3rd Student Computer Science Research Conference. Koper: University of Primorska Press, 2016
P. 26
10 7 ATI Radeon r9 280x The next step is to run the tests on hardware configurations
10 6 AMD FX-6300 of the same price point. Additionally we are working on
upstreaming our implementation to the Jmetal framework.
Runtime in miliseconds 10 5
7. REFERENCES
10 4
[1] J. Durillo, A. Nebro, and E. Alba. The jmetal
10 3 framework for multi-objective optimization: Design
and architecture. In CEC 2010, pages 4138–4325,
10 2 Barcelona, Spain, July 2010.
10 1 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 [2] J. J. Durillo and A. J. Nebro. jmetal: A java
0 Problem Size framework for multi-objective optimization. Advances
in Engineering Software, 42:760–771, 2011.
Figure 2: Runtime comparison on logarithmic scale
[3] B. Gaster, L. Howes, D. R. Kaeli, P. Mistry, and
not directly comparable we show that the speedups obtained D. Schaa. Heterogeneous Computing with OpenCL.
were in a reasonable margin. In best case our implementa- 2012.
tion is over 800 times faster then the CPU implementation.
We did not notice any improvement in case of smaller prob- [4] N. Hansen. The cma evolution strategy: a comparing
lems mainly due the time penalty of transferring problem review. In Towards a new evolutionary computation,
data from RAM to vRAM. pages 75–102. Springer, 2006.
5. ACKNOWLEDGMENT [5] N. Hansen. The CMA evolution strategy: A tutorial.
Vu le, 102(2006):1–34, 2011.
The authors would like to thank dr. Peter Koroˇsec for ad-
vising and guiding the research. [6] M. Harris. Gpgpu: General-purpose computation on
gpus. SIGGRAPH 2005 GPGPU COURSE, 2005.
6. FUTURE WORK
[7] M. Macedonia. The gpu enters computing’s
The hardware used to perform tests was not directly com- mainstream. Computer, 36(10):106–108, 2003.
parable. The Intel configuration had a much better CPU
while the AMD configuration had a state of the art GPU. [8] A. Munshi. The opencl specification. In 2009 IEEE
Hot Chips 21 Symposium (HCS), pages 1–314. IEEE,
2009.
[9] C. Nvidia. Programming guide, 2008.
[10] M. Pharr and R. Fernando. Gpu gems 2: programming
techniques for high-performance graphics and
general-purpose computation. Addison-Wesley
Professional, 2005.
[11] J. Shirazi. Tool report: Jprofiler. Java Performance
Tuning, 2002.
[12] S. Tsutsui and P. Collet. Massively parallel
evolutionary computation on GPGPUs. Springer, 2013.
[13] W. H. Wen-Mei. GPU Computing Gems Emerald
Edition. Elsevier, 2011.
StuCoSReC Proceedings of the 2016 3rd Student Computer Science Research Conference 26
Ljubljana, Slovenia, 12 October
10 6 AMD FX-6300 of the same price point. Additionally we are working on
upstreaming our implementation to the Jmetal framework.
Runtime in miliseconds 10 5
7. REFERENCES
10 4
[1] J. Durillo, A. Nebro, and E. Alba. The jmetal
10 3 framework for multi-objective optimization: Design
and architecture. In CEC 2010, pages 4138–4325,
10 2 Barcelona, Spain, July 2010.
10 1 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 [2] J. J. Durillo and A. J. Nebro. jmetal: A java
0 Problem Size framework for multi-objective optimization. Advances
in Engineering Software, 42:760–771, 2011.
Figure 2: Runtime comparison on logarithmic scale
[3] B. Gaster, L. Howes, D. R. Kaeli, P. Mistry, and
not directly comparable we show that the speedups obtained D. Schaa. Heterogeneous Computing with OpenCL.
were in a reasonable margin. In best case our implementa- 2012.
tion is over 800 times faster then the CPU implementation.
We did not notice any improvement in case of smaller prob- [4] N. Hansen. The cma evolution strategy: a comparing
lems mainly due the time penalty of transferring problem review. In Towards a new evolutionary computation,
data from RAM to vRAM. pages 75–102. Springer, 2006.
5. ACKNOWLEDGMENT [5] N. Hansen. The CMA evolution strategy: A tutorial.
Vu le, 102(2006):1–34, 2011.
The authors would like to thank dr. Peter Koroˇsec for ad-
vising and guiding the research. [6] M. Harris. Gpgpu: General-purpose computation on
gpus. SIGGRAPH 2005 GPGPU COURSE, 2005.
6. FUTURE WORK
[7] M. Macedonia. The gpu enters computing’s
The hardware used to perform tests was not directly com- mainstream. Computer, 36(10):106–108, 2003.
parable. The Intel configuration had a much better CPU
while the AMD configuration had a state of the art GPU. [8] A. Munshi. The opencl specification. In 2009 IEEE
Hot Chips 21 Symposium (HCS), pages 1–314. IEEE,
2009.
[9] C. Nvidia. Programming guide, 2008.
[10] M. Pharr and R. Fernando. Gpu gems 2: programming
techniques for high-performance graphics and
general-purpose computation. Addison-Wesley
Professional, 2005.
[11] J. Shirazi. Tool report: Jprofiler. Java Performance
Tuning, 2002.
[12] S. Tsutsui and P. Collet. Massively parallel
evolutionary computation on GPGPUs. Springer, 2013.
[13] W. H. Wen-Mei. GPU Computing Gems Emerald
Edition. Elsevier, 2011.
StuCoSReC Proceedings of the 2016 3rd Student Computer Science Research Conference 26
Ljubljana, Slovenia, 12 October
