DNNLibGen: Deep Neural Network Based Fast Library Generator

Naswali, Eunice; Quiros, Adalberto Claudio; Chandran, Pravin

doi:10.1109/icecs46596.2019.8965191

Cited by 6 publications

(7 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…State-of-the-art variation models are based on deep learning or machine learning [4], [29]- [32]. These models claim accurate variation estimations across various operating con-ditions.…”

Section: Cross-corner Variation Modelsmentioning

confidence: 99%

“…There are mainly two types of variation models: separate variation models which induce modifications to STAe.g., [30], [32] OCV, AOCV and POCV [1]-and library generation methods-e.g., [4], [29]-which generate generic libraries with estimated timings. In contrast, our proposed model may fall into both categories.…”

Section: F Compatibility With Conventional Delay Modelsmentioning

confidence: 99%

“…Our model can also generate NLDM libraries at non-base corners by applying the scaling to readily available T P D and T tran in the lookup tables. Furthermore, the concept of decomposition itself may find its place in state-of-the-art variation models based on machine learning or deep learning such as [4], [29]- [32], although the impact is to be assessed in future works. Decomposition of T P D yields monotonic and predictable timing components; this process is equivalent to pre-processing input data in machine learning.…”

Section: F Compatibility With Conventional Delay Modelsmentioning

confidence: 99%

“…In this part, we compare our proposed delay variation models to the state-of-the-art variation models [4], [29]- [31]. First, we compare various aspects of these works as summarized in Table 2.…”

Section: Comparison To State-of-the-art Approachesmentioning

confidence: 99%

“…The model in [29] uses a convolutional neural network (CNN) to estimate timing in both inter-corner and intracorner manner. This 40-layer fully connected network has the highest computational overheads of network configuration, training, and inference.…”

Section: Comparison To State-of-the-art Approachesmentioning

confidence: 99%

See 4 more Smart Citations

Cross-Corner Delay Variation Model for Standard Cell Libraries

Kim

Joo²,

Park

et al. 2021

IEEE Access

View full text Add to dashboard Cite

For timing closure of logic circuits, circuit designers must perform sign-offs on a variety of process, voltage, and temperature (PVT) conditions. Designs of advanced logic circuits involve a multitude of voltage islands and operating modes, each of which requires delay characterizations at nearby PVT corners. Furthermore, advanced technologies nodes suffer from corner explosion: while the impact of PVT variations is being exacerbated, process variations are also diversifying, increasing the number of operating conditions exponentially. This paper revisits the importance of cross-corner timing estimations and proposes a delay variation model to mitigate such corner explosion. Our objective is to reduce PVT corner characterization effort for timely static timing analysis on an exploding number of operating conditions. Our proposed Decomposed Propagation Vector Variation Model (DPVVM) decomposes propagation delay and timing constraints into the driving by the receiver cell and its driver cell; by scaling them separately, delay characterization effort is reduced to a fraction of time while realizing accurate timing estimations. We also propose Multi-Dimension Recomposition (MDR) scheme, which exploits a multitude of pre-characterized corners to further improve the consistency of cross-corner timing estimations. As a result, with only 8.0% of a corner characterization effort compared to the full-characterization, DPVVM combined with MDR achieves overall cross-corner timing estimation errors of 4.8% and 5.6% for single cells and complex logic circuits-or improvements of 69% and 61% over the conventional derating method, respectively. Our proposed method's characterization overhead is 11% over the conventional derating method; the overhead is marginal, accounting for only 0.76% of a full-characterization time.

show abstract

“…State-of-the-art variation models are based on deep learning or machine learning [4], [29]- [32]. These models claim accurate variation estimations across various operating con-ditions.…”

Section: Cross-corner Variation Modelsmentioning

confidence: 99%