Painting on PIacement: Forecasting Routing Congestion using Conditional Generative Adversarial Nets

@article{Yu2019PaintingOP,
  title={Painting on PIacement: Forecasting Routing Congestion using Conditional Generative Adversarial Nets},
  author={Cunxi Yu and Zhiru Zhang},
  journal={2019 56th ACM/IEEE Design Automation Conference (DAC)},
  year={2019},
  pages={1-6}
}
  • Cunxi Yu, Zhiru Zhang
  • Published 2019
  • Computer Science
  • 2019 56th ACM/IEEE Design Automation Conference (DAC)
Physical design process commonly consumes hours to days for large designs, and routing is known as the most critical step. Demands for accurate routing quality prediction raise to a new level to accel-erate hardware innovation with advanced technology nodes. This work presents an approach that forecasts the density of all routing channels over the entire floorplan, with features collected up to placement, using conditional GANs. Specifically, forecasting the routing congestion is constructed as… Expand
High-Definition Routing Congestion Prediction for Large-Scale FPGAs
TLDR
This work proposes a novel approach to predict the routing congestion map for large-scale FPGA designs at the placement stage that can be incorporated into the placement engine to provide congestion prediction resulting in up to 7% reduction in routed wirelength for the most congested design in ISPD 2016 benchmark. Expand
Towards Decrypting the Art of Analog Layout: Placement Quality Prediction via Transfer Learning
  • Mingjie Liu, Keren Zhu, +4 authors D. Pan
  • Computer Science
  • 2020 Design, Automation & Test in Europe Conference & Exhibition (DATE)
  • 2020
TLDR
This paper uses a state-of-the-art automatic analog layout tool and industry-level simulator to generate labeled training data in an automated manner and proposes a 3D convolutional neural network model to predict the relative placement quality using well-crafted placement features. Expand
A Deep-Learning Framework for Predicting Congestion During FPGA Placement
TLDR
Compared to MLCong, a state-of-the-art machine-learning based congestion-estimation model, DLCong achieves an almost 9% improvement in congestion accuracy, while exhibiting inference times of a few milliseconds. Expand
DRC Hotspot Prediction at Sub-10nm Process Nodes Using Customized Convolutional Network
TLDR
A new convolutional neural network technique, J-Net, is introduced for the prediction with mixed resolution features of DRC hotspot prediction, which can improve true positive rate by 37%, 40% and 14% respectively, compared to three recent works, with similar false positive rates. Expand
Generative Learning in VLSI Design for Manufacturability: Current Status and Future Directions
TLDR
The unique features of generative learning are examined that have been leveraged to improve DFM efficiency in an unprecedented way; hence, paving the way to a new data-driven DFM approach. Expand
EM-GAN: Fast Stress Analysis for Multi-Segment Interconnect Using Generative Adversarial Networks
TLDR
A fast transient hydrostatic stress analysis for electromigration (EM) failure assessment for multi-segment interconnects using generative adversarial networks (GANs) using the image synthesis feature of GAN-based generative deep neural networks is proposed. Expand
FlowTune: Practical Multi-armed Bandits in Boolean Optimization
  • Cunxi Yu
  • Computer Science
  • 2020 IEEE/ACM International Conference On Computer Aided Design (ICCAD)
  • 2020
TLDR
A generic end-to-end and high-performance domain-specific, multi-stage multi-armed bandit framework for Boolean logic optimization that outperforms both hand-crafted flows and ML explored flows in quality of results, and is orders of magnitude faster compared to ML-based approaches. Expand
Re-examining VLSI Manufacturing and Yield through the Lens of Deep Learning : (Invited Talk)
TLDR
This work sheds light on the recent Deep Learning (DL) based approaches that have provided a new lens to examine traditional manufacturability and yield challenges, and presents lithography modeling and simulation techniques, leveraging advanced learning paradigms, which have demonstrated unprecedented efficiency. Expand
Decision Making in Synthesis cross Technologies using LSTMs and Transfer Learning
  • Cunxi Yu, Wang Zhou
  • Computer Science
  • 2020 ACM/IEEE 2nd Workshop on Machine Learning for CAD (MLCAD)
  • 2020
TLDR
The main idea is leveraging LSTM-based network to forecast the QoR, where the inputs are synthesis flows represented in novel timed-flow modeling, and QoRs are ground truth, which enables accurate predictions cross different technologies and different IC designs. Expand
Fast IR drop estimation with machine learning
TLDR
This paper provides a review to the latest progress in ML-based IR drop estimation techniques and serves as a vehicle for discussing some general challenges faced by ML applications in electronics design automation (EDA), and demonstrating how to integrate ML models with conventional techniques for the better efficiency of EDA tools. Expand
...
1
2
...

References

SHOWING 1-10 OF 24 REFERENCES
RouteNet: Routability prediction for Mixed-Size Designs Using Convolutional Neural Network
TLDR
The proposed method, called RouteNet, can either evaluate the overall routability of cell placement solutions without global routing or predict the locations of DRC (Design Rule Checking) hotspots, and significantly outperforms other machine learning approaches such as support vector machine and logistic regression. Expand
Clock-aware ultrascale FPGA placement with machine learning routability prediction: (Invited paper)
TLDR
This paper proposes several detailed placement techniques, i.e., two-step clock constraint legalization and chain move, and explores the possibilities to use machine learning-based methods to predict routing congestion in UltraScale FPGAs. Expand
Layout hotspot detection with feature tensor generation and deep biased learning
TLDR
A deep learning framework for high performance and large scale hotspot detection is developed and a biased learning algorithm is proposed to train the convolutional neural network to further improve detection accuracy with small false alarm penalties. Expand
Developing Synthesis Flows Without Human Knowledge
TLDR
This work presents a fully autonomous framework that artificially produces design-specific synthesis flows without human guidance and baseline flows, using Convolutional Neural Network (CNN). Expand
Routability Optimization for Industrial Designs at Sub-14nm Process Nodes Using Machine Learning
TLDR
This paper demonstrates on several layouts of a sub-14nm industrial design that this method predicts the locations of 74% of the detailed-route DRCs and automatically reduces the number of detailed- route DRC violations by up to 5x. Expand
Image-to-Image Translation with Conditional Adversarial Networks
TLDR
Conditional adversarial networks are investigated as a general-purpose solution to image-to-image translation problems and it is demonstrated that this approach is effective at synthesizing photos from label maps, reconstructing objects from edge maps, and colorizing images, among other tasks. Expand
EPIC: Efficient prediction of IC manufacturing hotspots with a unified meta-classification formulation
TLDR
EPIC is an efficient and effective predictor for IC manufacturing hotspots in deep sub-wavelength lithography and proposes a unified framework to combine different hotspot detection methods together, such as machine learning and pattern matching, using mathematical programming/optimization. Expand
Fast and Accurate Estimation of Quality of Results in High-Level Synthesis with Machine Learning
TLDR
This work builds a large collection of C-to-FPGA results from a diverse set of realistic HLS applications and identifies relevant features from HLS reports for estimating post-implementation metrics, and trains and compares a number of promising machine learning models to effectively and efficiently bridge the accuracy gap. Expand
Generative Adversarial Nets
We propose a new framework for estimating generative models via an adversarial process, in which we simultaneously train two models: a generative model G that captures the data distribution, and aExpand
Machine-learning-based hotspot detection using topological classification and critical feature extraction
TLDR
By combing topological classification and critical feature extraction, this hotspot detection framework achieves very high accuracy and outperforms the 2012 CAD Contest at ICCAD winner on accuracy and false alarm. Expand
...
1
2
3
...