Abstract: Average power analysis plays a crucial role in the design of large-scale digital integrated circuits (ICs). The integration of data-driven machine learning (ML) methods into the electronic design automation (EDA) fields has increased the demand for extensive datasets. To address this need, we propose a novel pseudo-circuit generation algorithm rooted in graph topology. This algorithm efficiently produces a multitude of power analysis examples by converting randomly generated directed acyclic graphs (DAGs) into gate-level Verilog pseudo-combinational circuit netlists. The subsequent introduction of register units transforms pseudo-combinational netlists into pseudo-sequential circuit netlists. Hyperparameters facilitate the control of circuit topology, while appropriate sequential constraints are applied during synthesis to yield a pseudo-circuit dataset. We evaluate our approach using the mainstream power analysis software, conducting pre-layout average power tests on the generated circuits, comparing their performance against benchmark datasets, and verifying the results through circuit topology complexity analysis and static timing analysis (STA). The results confirm the effectiveness of the dataset, and demonstrate the operational efficiency and robustness of the algorithm, underscoring its research value.

[1]Ajayi T, Blaauw D, 2019. OpenROAD: toward a self-driving, open-source digital layout implementation tool chain. Proc Government Microcircuit Applications and Critical Technology Conf, p.1105-1110.

[2]Amarú L, Gaillardon PE, De Micheli G, 2015. The EPFL combinational benchmark suite. Proc 24^th Int Workshop on Logic & Synthesis.

[3]Amid A, Biancolin D, Gonzalez A, et al., 2020. Chipyard: integrated design, simulation, and implementation framework for custom SoCs. IEEE Micro, 40(4):10-21.

[4]Brglez F, Bryan D, Kozminski K, 1989. Combinational profiles of sequential benchmark circuits. IEEE Int Symp on Circuits and Systems, p.1929-1934.

[5]Burch R, Najm F, Yang P, et al., 1992. McPOWER: a Monte Carlo approach to power estimation. IEEE/ACM Int Conf on Computer-Aided Design, p.90-97.

[6]Chai ZM, Zhao YX, Lin YB, et al., 2022. CircuitNet: an open-source dataset for machine learning applications in electronic design automation (EDA). Sci China Inform Sci, 65(12):227401.

[7]Chowdhury AB, Tan B, Karri R, et al., 2021. OpenABC-D: a large-scale dataset for machine learning guided integrated circuit synthesis.

[8]Corno F, Reorda M, Squillero G, 2000. RT-level ITC’99 benchmarks and first ATPG results. IEEE Des Test Comput, 17(3):44-53.

[9]Fang WJ, Lu Y, Liu S, et al., 2023. MasterRTL: a pre-synthesis PPA estimation framework for any RTL design. IEEE/ACM Int Conf on Computer-Aided Design, p.1-9.

[10]Gautier Q, Althoff A, Meng PF, et al., 2016. Spector: an OpenCL FPGA benchmark suite. Int Conf on Field-Programmable Technology, p.141-148.

[11]Hansen MC, Yalcin H, Hayes JP, 1999. Unveiling the ISCAS-85 benchmarks: a case study in reverse engineering. IEEE Des Test Comput, 16(3):72-80.

[12]Kaeslin H, 2008. Digital Integrated Circuit Design: from VLSI Architectures to CMOS Fabrication. Cambridge University Press, New York, USA, p.386-458.

[13]Khan S, Shi ZY, Li M, et al., 2024. DeepSeq: deep sequential circuit learning. Design, Automation & Test in Europe Conf & Exhibition, p.1-2.

[14]Kumar AKA, Gerstlauer A, 2019. Learning-based CPU power modeling. ACM/IEEE 1^st Workshop on Machine Learning for CAD, p.1-6.

[15]Kumar AKA, Al-Salamin S, Amrouch H, et al., 2023. Machine learning-based microarchitecture-level power modeling of CPUs. IEEE Trans Comput, 72(4):941-956.

[16]Li M, Khan S, Shi Z, et al., 2022. DeepGate: learning neural representations of logic gates. Proc 59^th ACM/IEEE Design Automation Conf, p.667-672.

[17]Najm FN, 1993. Transition density: a new measure of activity in digital circuits. IEEE Trans Comput-Aided Des Integr Circ Syst, 12(2):310-323.

[18]Nasser Y, Lorandel J, Prévotet JC, et al., 2021. RTL to transistor level power modeling and estimation techniques for FPGA and ASIC: a survey. IEEE Trans Comput-Aided Des Integr Circ Syst, 40(3):479-493.

[19]OnchipUIS, 2016. MRISCV. https://github.com/onchipuis/mriscv [Accessed on July 26, 2024].

[20]OnchipUIS, 2021. VexRiscv. https://github.com/SpinalHDL/VexRiscv [Accessed on July 26, 2024].

[21]OpenCores Team, 2024. OpenCores. https://opencores.org [Accessed on July 26, 2024].

[22]Parulkar I, Wood A, Hoe JC, et al., 2008. OpenSPARC: an open platform for hardware reliability experimentation. 4^th Workshop on Silicon Errors in Logic-System Effects, p.1-6.

[23]Rakesh MB, Das P, Sai Pranav KR, et al., 2023. GRILAPE: graph representation inductive learning-based average power estimation for frontend ASIC RTL designs. 36^th Int Conf on VLSI Design and 22nd Int Conf on Embedded Systems, p.1-6.

[24]Shi ZY, Pan HY, Khan S, et al., 2023. DeepGate2: functionality-aware circuit representation learning. IEEE/ACM Int Conf on Computer-Aided Design, p.1-6.

[25]Wei ZG, Arora A, Li RH, et al., 2023. HLSDataset: open-source dataset for ML-assisted FPGA design using high level synthesis. IEEE 34^th Int Conf on Application-Specific Systems, Architectures and Processors, p.197-204.

[26]Xie ZY, 2023. Efficient runtime power modeling with on-chip power meters. Proc Int Symp on Physical Design, p.168-174.

[27]Xie ZY, Pan JY, Chang CC, et al., 2023a. The dark side: security and reliability concerns in machine learning for EDA. IEEE Trans Comput-Aided Des Integr Circ Syst, 42(4):1171-1184.

[28]Xie ZY, Zhang T, Peng YF, 2023b. Security and reliability challenges in machine learning for EDA: latest advances. 24^th Int Symp on Quality Electronic Design, p.1-6.

[29]YosysHQ, 2021. PicoRV32—a Size-Optimized RISC-V CPU. https://github.com/YosysHQ/picorv32 [Accessed on July 26, 2024].

[30]Zhang YQ, Ren HX, Khailany B, 2020. GRANNITE: graph neural network inference for transferable power estimation. 57^th ACM/IEEE Design Automation Conf, p.1-6.

[31]Zhou GF, Zhou JY, Lin HJ, 2018. Research on NVIDIA deep learning accelerator. 12^th IEEE Int Conf on Anti-counterfeiting, Security, and Identification, p.192-195.

[32]Zhou Y, Ren HX, Zhang YQ, et al., 2019. PRIMAL: power inference using machine learning. Proc 56^th Annual Design Automation Conf, Article 39.

[33]Zou SN, Zhang JX, Shi BZ, et al., 2024. PowerSyn: a logic synthesis framework with early power optimization. IEEE Trans Comput-Aided Des Integr Circ Syst, 43(1):203-216.