IEEE Transactions on Very Large Scale Integration (VLSI) Systems - Special section on low power
Characterizing process variation in nanometer CMOS
Proceedings of the 44th annual Design Automation Conference
Variation-Aware Application Scheduling and Power Management for Chip Multiprocessors
ISCA '08 Proceedings of the 35th Annual International Symposium on Computer Architecture
Chameleon: Application-Level Power Management
IEEE Transactions on Mobile Computing
Evaluation of the Intel® Core i7 Turbo Boost feature
IISWC '09 Proceedings of the 2009 IEEE International Symposium on Workload Characterization (IISWC)
Evaluating the effectiveness of model-based power characterization
USENIXATC'11 Proceedings of the 2011 USENIX conference on USENIX annual technical conference
Slow down or sleep, that is the question
USENIXATC'11 Proceedings of the 2011 USENIX conference on USENIX annual technical conference
Benchmarking modern multiprocessors
Benchmarking modern multiprocessors
System-level application-aware dynamic power management in adaptive pipelined MPSoCs for multimedia
Proceedings of the International Conference on Computer-Aided Design
IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Hi-index | 0.00 |
The variability in performance and power consumption is slated to grow further with continued scaling of process technologies. While this variability has been studied and modeled before, there is lack of empirical data on its extent, as well as the factors affecting it, especially for modern general purpose microprocessors. Using detailed power measurements we show that the part to part variability for modern processors utilizing the Nehalem microarchitecture is indeed significant. We chose six Core i5-540M laptop processors marketed in the same frequency bins - thus presumed to be identical - and characterized their power consumption for a variety of representative single-threaded and multithreaded application workloads. Our data shows processor power variation ranging from 7%-17% across different applications and configuration options such as Hyper-Threading and Turbo Boost. We present our hypotheses on the underlying causes of this observed power variation and discuss its potential implications.