Leakage-Aware Speculative Branch Target Buffer
Modern processors are highly dependent on the speculation of the control flow of the application by means of branch prediction. To provide accurate speculation of the direction of each branch, processors need a complex branch predictor with a large Branch Target Buffer. The branch target
buffer is a memory area in which the target addresses of the taken branches are stored. As the size of the buffer grows it can provide more accurate speculation but it will need more power for the branch target buffer memory area. Most of the branch target buffer, however, is not needed at
any certain point of the working period of the processor. In this work, we introduce Leakage-Aware Speculative Branch Target Buffer to reduce the static power dissipation of the branch target buffer while trying to keep up with the performance of the conventional branch target buffer architecture.
This can be done by introducing a predictive technique in a way that the buffer speculatively activates the next needed entry in the memory while putting the rest of the cells in the low power or “Drowsy” state. The results show that on average our method dissipates about one third
of the static power compared to a conventional branch target buffer while maintaining performance.
Keywords: BRANCH PREDICTION; BRANCH TARGET BUFFER; LOW POWER DESIGN
Document Type: Research Article
Publication date: 01 December 2012
- The electronic systems that can operate with very low power are of great technological interest. The growing research activity in the field of low power electronics requires a forum for rapid dissemination of important results: Journal of Low Power Electronics (JOLPE) is that international forum which offers scientists and engineers timely, peer-reviewed research in this field.
- Editorial Board
- Information for Authors
- Subscribe to this Title
- Terms & Conditions
- Ingenta Connect is not responsible for the content or availability of external websites
- Access Key
- Free content
- Partial Free content
- New content
- Open access content
- Partial Open access content
- Subscribed content
- Partial Subscribed content
- Free trial content