EAST-DNN: Expediting architectural SimulaTions using deep neural networks: work-in-progress

A rapid and accurate architectural simulator is a cornerstone for an efficient design-space exploration of computing systems. In this paper, we introduce EAST-DNN, a feed-forward deep neural network, to accelerate architectural simulations. EAST-DNN achieves > 106X speedup with an average prediction error of 4.3% over the baseline simulator. It also achieves an average of 2X better accuracy with at least 2.3X speedup compared to state-of-the-art.

[1]  Tor M. Aamodt,et al.  Dynamic Warp Formation and Scheduling for Efficient GPU Control Flow , 2007, 40th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO 2007).

[2]  Sally A. McKee,et al.  Efficiently exploring architectural design spaces via predictive modeling , 2006, ASPLOS XII.

[3]  J. Friedman.  Greedy function approximation: A gradient boosting machine. , 2001 .

[4]  Xian-He Sun,et al.  Efficient design space exploration via statistical sampling and AdaBoost learning , 2016, 2016 53nd ACM/EDAC/IEEE Design Automation Conference (DAC).

[5]  Yoshua Bengio,et al.  Understanding the difficulty of training deep feedforward neural networks , 2010, AISTATS.

[6]  Michael F. P. O'Boyle,et al.  Microarchitectural Design Space Exploration Using an Architecture-Centric Approach , 2007, 40th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO 2007).

[7]  Gu-Yeon Wei,et al.  Aladdin: A pre-RTL, power-performance accelerator simulator enabling large design space exploration of customized architectures , 2014, 2014 ACM/IEEE 41st International Symposium on Computer Architecture (ISCA).

[8]  Matthew Mattina,et al.  SCALE-Sim: Systolic CNN Accelerator , 2018, ArXiv.

[9]  Tianqi Chen,et al.  XGBoost: A Scalable Tree Boosting System , 2016, KDD.

[10]  Mark Horowitz,et al.  1.1 Computing's energy problem (and what we can do about it) , 2014, 2014 IEEE International Solid-State Circuits Conference Digest of Technical Papers (ISSCC).