The Rise and Fall of Dark Silicon

Industry experts predict that transistor counts will continue to grow exponentially for at least another decade. Historically, we were able to harness all of these transistors to deliver exponential increases in computational power by capitalizing on both technological improvements and micro-architectural innovation. However, after decades of reaping Moore’s bounty, processors eventually hit a power wall. Technological limitations will soon prevent us from powering all transistors simultaneously, leaving a large fraction of the chip powered off (or dark). Short of a technological miracle, we head towards an era of “dark silicon,” able to build dense devices we cannot afford to power. Without the ability to use more transistors, or run them faster, performance improvements are likely to stagnate, unless we change course.

[1]  Tyler Clemons,et al.  Elastic Fidelity: Trading-off Computational Accuracy for Energy Reduction , 2011, ArXiv.

[2]  Aamer Jaleel,et al.  High performance cache replacement using re-reference interval prediction (RRIP) , 2010, ISCA.

[3]  Babak Falsafi,et al.  Near-Optimal Cache Block Placement with Reactive Nonuniform Cache Architectures , 2010, IEEE Micro.

[4]  Babak Falsafi,et al.  Chip multiprocessors for server workloads , 2009 .

[5]  Dan Grossman,et al.  EnerJ: approximate data types for safe and general low-power computation , 2011, PLDI '11.

[6]  Guy E. Blelloch,et al.  Scheduling threads for constructive cache sharing on CMPs , 2007, SPAA '07.

[7]  Wolf-Dietrich Weber,et al.  Power provisioning for a warehouse-sized computer , 2007, ISCA '07.

[8]  Nikolaos Hardavellas,et al.  Dynamic Directories: Reducing On-Chip Interconnect Power in Multicores , 2012, DATE 2012.

[9]  Gokhan Memik,et al.  A Case for Clumsy Packet Processors , 2004, 37th International Symposium on Microarchitecture (MICRO-37'04).

[10]  Ippokratis Pandis,et al.  Data-oriented transaction execution , 2010, Proc. VLDB Endow..

[11]  Christoforos E. Kozyrakis,et al.  Understanding sources of inefficiency in general-purpose chips , 2010, ISCA.

[12]  James C. Hoe,et al.  Single-Chip Heterogeneous Computing: Does the Future Include Custom Logic, FPGAs, and GPGPUs? , 2010, 2010 43rd Annual IEEE/ACM International Symposium on Microarchitecture.

[13]  A.B. Kahng,et al.  Impact of Guardband Reduction On Design Outcomes: A Quantitative Approach , 2009, IEEE Transactions on Semiconductor Manufacturing.

[14]  Karthikeyan Sankaralingam,et al.  Dark Silicon and the End of Multicore Scaling , 2012, IEEE Micro.

[15]  Babak Falsafi,et al.  Toward Dark Silicon in Servers , 2011, IEEE Micro.