Abstract:Nowadays, multi-core processor, also called Chip Multiprocessor (CMP) becomes the mainstream that can achieve higher computation capability. However, energy issue is still a crucial problem for design and manufacture of multi-core processor. Tag reduction technique can save energy of the single-core system. This chapter introduces the Tag Reduction on CMP (TRoCMP) that is a novel approach to energy saving for multicore system. We first extend tag reduction from single-core to multi-core processor, including proposing 3 heuristic algorithms to implement TRoCMP. Then the performance overhead is considered, so that Core Degree mechanism and a refined heuristic algorithm are further introduced and designed to find out the trade-off of energy saving and performance overhead of TRoCMP. In particular, we formulate the energy consumption and performance overhead of TRoCMP to analyze and estimate them. In experiments, we modify the Linux kernel and implement new modules to collect the experimental data from benchmarks of SPEC CPU2006 running on a real operating system. In this way, the precision of our experiments is guaranteed, since tag reduction is very sensitive to the usage of physical memory. The experimental results show that our TRoCMP can save total energy up to 83.93% and 76.16% on 8-core and 4-core processors in average respectively, compared to the one that the tag-reduction is not used for. TRoCMP outperforms significantly tag reduction on single-core processor as well. With consideration of performance overhead, when Core Degree is set to 6, the best balance of energy saving and performance overhead can be achieved.

I-Cache Tag Reduction for Low Power Chip Multiprocessor.

Energy Efficiency of a Multi-Core Processor by Tag Reduction

An Improved Approach to Tag Reduction on Low Power Cmp with Trade-Off of Energy and Performance

The Core Degree Based Tag Reduction on Chip Multiprocessor to Balance Energy Saving and Performance Overhead

Low power instruction cache based on adjacent line linking access

A Way Memoization Technique for Reducing Power Consumption of Caches in Application Specific Integrated Processors

Study on the Low Power Cache Design Technique

Minimizing Energy Consumption for Embedded Multicore Systems Using Cache Configuration and Task Mapping

TRoCMP: An Approach to Energy Saving for Multi-Core Systems

Exploring the Limits of Tag Reduction for Energy Saving on a Multi-core Processor

Cache Miss Reduction Through Hardware-Assisted Loop Optimization

Online Cache Modeling for Commodity Multicore Processors

Pre-Visiting Tag and Keeping Way History to Reduce Power in Instruction Cache

Low power design strategy for unified cache based on division access of instruction and data

An Efficient Lightweight Shared Cache Design for Chip Multiprocessors

The energy optimization for architectures with limited addressing modes using scratch-pad memory

Design of a Low-Power Cache with Partial-Tag Comparison

Cache Power Optimization Based on Compare-Based Adaptive Clock Gating and Its 65nm SoC Implementation

Design of L2 CACHE Replacement Arithmetic for Chip Multiprocessor Architecture

L1 Collective Cache: Managing Shared Data for Chip Multiprocessors

Cache Hierarchy Optimization