Importance of single-core performance in the multicore era
Toshinori Sato, Hideki Mori, Rikiya Yano, Takanori Hayashida
Abstract
Toshinori Sato, Hideki Mori, Rikiya Yano, Takanori Hayashida
Abstract
This paper first investigates what the best multicore configuration will be in the future, when the number of usable transistors further increases. Comparing five multicore models: single-core, many-core, heterogeneous multicore, scalable homogeneous multicore, and dynamically configurable multicore, surprisingly unveils that single-core performance is a key to improve multicore performance. Based on the findings, this paper secondly proposes a technique to improve single-core performance. It is based on Intel's Turbo Boost technology. From the detailed simulations, it is found that the technique achieves single-core performance improvement.
OpenAlex reports 3 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
This paper first investigates what the best multicore configuration will be in the future, when the number of usable transistors further increases. Comparing five multicore models: single-core, many-core, heterogeneous multicore, scalable homogeneous multicore, and dynamically configurable multicore, surprisingly unveils that single-core performance is a key to improve multicore performance. Based on the findings, this paper secondly proposes a technique to improve single-core performance. It is based on Intel's Turbo Boost technology. From the detailed simulations, it is found that the technique achieves single-core performance improvement.
Key concepts: Multi-core processor, Computer science, Scalability, Core (optical fiber), Single-core, USable, Computer architecture, Parallel computing