Cryogenic circuit design requires compact transistor models that reproduce measured low-temperature behavior and run in standard circuit simulators. We fit BSIM4 parameters to measured I-V curves of SkyWater 130 nm (SKY130) transistors at 77 K using a neural surrogate of NGSpice. The surrogate maps compact-model parameters and device geometry to the complete set of output and transfer characteristics. Extraction is a bounded least-squares search: a bank of NGSpice simulations supplies initial candidates, finite differences of the surrogate supply the local Jacobian, and NGSpice re-simulates the resulting cards before selection. On 18 measured geometries, we compare a search over seven card parameters with a search over 43 parameters selected by NGSpice sensitivity. Relative to NGSpice re-simulations of the previously released cards, the two searches reduce the in-sample mean relative root-mean-square error (RRMS) by 17% and 26%, respectively. The 43-parameter search has the lower overall error, but the seven-parameter search has the lower mean on high-bias curves, and some individual curves have higher error than with the released cards under both searches.
Figures & tables
Figure 1: Extraction pipeline. NGSpice generates the simulation bank, the frozen forward surrogate is used for local refinement, and NGSpice re-simulates the selected candidates for comparison with measurement.
Figure 2: Mean device RRMS from NGSpice re-simulations of the released, seven-parameter, and 43-parameter cards. “High bias” averages, within each device, the output curves at fixed ∣VGS∣≥1.48\text{,}\mathrm{V} and the transfer curves at fixed $|V_{DS}|\geq$1.48\text{\,}\mathrm{V} .
Figure 3: Output characteristic of the nMOS L=1\text{,}\mathrm{\SIUnitSymbolMicro m} , $W=$3\text{\,}\mathrm{\SIUnitSymbolMicro m} device at VGS=1.85\text{,}\mathrm{V} . Left: full $V_{DS}$ sweep. Right: $V_{DS}\leq$0.37\text{\,}\mathrm{V} . Lines are NGSpice re-simulations; markers are measurements.
Figure 4: Per-curve RRMS, released card vs. fitted card. Each panel shows the 156 of 175 scored curves with released and fitted RRMS at most 0.5; the counts in the panel titles include all 175. Red markers and shading mark curves whose fitted RRMS exceeds the released-card RRMS by more than 0.01 and by more than 10%. The flag uses a fixed threshold; no significance test is applied. One flagged seven-parameter curve lies outside the plotted range; all 38 flagged 43-parameter curves are shown.
Appendix figures & tables12 assets
Supplementary material from the paper’s appendix.
Appendix
Item
Value
Devices
8 nMOS, 10 pMOS; 18 measured geometries
Curves per device
5 output, 6 transfer; 2,046 points
Synthetic data
10,000 LHC rows per geometry and parameter set
Surrogate inputs
parameters, polarity, scaled logL , logW
Hidden layers
4×1024 ; GELU
Split / batch size
80%/10%/10% per device; 512 rows
Appendix
Table 1: Surrogate and extraction configuration.
Type
L ( \SIUnitSymbolMicrom )
W ( \SIUnitSymbolMicrom )
Released
7 param.
43 param.
nMOS
0.15
1.6
0.0595
0.0351
0.0380
nMOS
0.19
7
0.0967
0.0436
0.0346
nMOS
0.25
1.6
0.0983
0.0487
0.0438
nMOS
1
1.6
0.1283
0.0637
0.0648
nMOS
1
3
0.1599
0.1102
0.0657
nMOS
8
1.6
0.1474
0.0809
0.0766
Appendix
Table 2: Per-device RRMS from NGSpice re-simulations. “Released” is the released 77K parameter vector re-simulated with the same simulator and scorer.
Figure 5: RRMS for the nMOS L=0.15\text{,}\mathrm{\SIUnitSymbolMicro m} , $W=$1.6\text{\,}\mathrm{\SIUnitSymbolMicro m} device at fixed ∣VGS∣=1.48\text{,}\mathrm{V}$$ for the main and alternate methods.
Figure 6: Complete I–V overlays for nMOS devices (1 of 4).
Figure 7: Complete I–V overlays for nMOS devices (2 of 4).
Figure 8: Complete I–V overlays for nMOS devices (3 of 4).
Figure 9: Complete I–V overlays for nMOS devices (4 of 4).
Figure 10: Complete I–V overlays for pMOS devices (1 of 5).
Figure 11: Complete I–V overlays for pMOS devices (2 of 5).
Figure 12: Complete I–V overlays for pMOS devices (3 of 5).
Figure 13: Complete I–V overlays for pMOS devices (4 of 5).
Figure 14: Complete I–V overlays for pMOS devices (5 of 5).
Neural network (NN)-based transistor models have proven to be a promising solution to accelerate device modeling. Although these models demonstrated remarkable speedup in circuit simulations, they were not rigorously tested in complex tasks within standard EDA tool flows for chip design, which require very high transistor model precision to provide accurate results and converge properly. To investigate their full potential and promise, we develop NN-based transistor models that accurately predict the drain current and various charges for a 3nm nanosheet, and employ the models for a standard cell library characterization. This enables us to evaluate the accuracy and speedup under a very wide range of SPICE simulations, benchmarking our models against an industry-standard implementation. Since standard cell library characterization involves thousands of SPICE simulations, it serves as a great benchmark for transistor model quality and consistency. To further evaluate model accuracy, we also perform analog- and digital-level simulations for various complex circuits. Our experiments reveal that NN-based models are a valid alternative to existing transistor compact models, capable of producing accurate delay and power estimates with a much higher speed and convergence time compared to the conventional Verilog-A-based industry-standard (BSIM-CMG) transistor compact model. Our NNs achieve sub-0.11% and 0.76% errors for delay and total power evaluation on a wide range of circuits, while taking up to 8 times less time during the standard cell library characterization. Additionally, we provide comprehensive information and guidance on NN accuracy, unveiling the relationship between NN size and its effect on circuit simulation precision.
Rodion Novkin, Hussam Amrouch
Technical University of Munich; TUM School of Computation, Information and Technology; Chair of AI Processor Design; Munich Institute of Robotics and Machine Intelligence, Munich, Germany
Compact transistor models are the mathematical backbone of circuit simulation. However, at advanced nodes such as 3nm, transport physics becomes too complex for traditional hand-derived equations to capture accurately. Purely data-driven neural surrogates, on the other hand, are numerically unstable inside circuit solvers and offer no physical insight into their own predictions. We introduce PRISM-UDE (Physics-Regularized Iterative Symbolic Modeling via Universal Differential Equations), a framework that embeds a small neural network inside a physics-based transistor model, using the network only to learn the transport behavior that the analytical baseline misses, rather than replacing the physics altogether. Once trained, this neural correction is distilled into a single, interpretable closed-form expression via symbolic regression, making the final model fully analytical and simulator-ready. Applied to a 3nm FinFET benchmark dataset, PRISM-UDE reduces prediction error more than sixfold (70.33% to 11.01%) relative to the standard physics-only baseline. The distilled expression preserves this accuracy almost exactly while eliminating the neural network entirely. We further validate the extracted expression directly inside a SPICE circuit simulator, confirming stable, physically consistent behavior under both static bias sweeps and dynamic switching conditions.
Analog circuit design relies heavily on reusing existing intellectual property (IP), yet searching across heterogeneous representations such as SPICE netlists, schematics, and functional descriptions remains challenging. Existing methods are largely limited to exact matching within a single modality, failing to capture cross-modal semantic relationships. To bridge this gap, we present AnalogRetriever, a unified tri-modal retrieval framework for analog circuit search. We first build a high-quality dataset on top of Masala-CHAI through a two-stage repair pipeline that raises the netlist compile rate from 22.0% to 99.97%. Built on this foundation, AnalogRetriever encodes schematics and descriptions with a vision-language model and netlists with a port-aware relational graph convolutional network, mapping all three modalities into a shared embedding space via curriculum contrastive learning. Experiments show that AnalogRetriever achieves an average Recall@1 of 75.2% across all six cross-modal retrieval directions, significantly outperforming existing baselines. When integrated into the AnalogCoder agentic framework as a retrieval-augmented generation module, it consistently improves functional pass rates and enables previously unsolved tasks to be completed.
Yihan Wang, Lei Li, Yao Lai +2
Tsinghua University, Beijing, China · The University of Hong Kong, Hong Kong, China · University of Cambridge, Cambridge, United Kingdom +1