This disclosure describes techniques to predict power consumption of a computing device under design. Per techniques of this disclosure, a machine learning model is trained based on parameterized hardware attributes of existing devices. A training dataset is obtained based on device parameters and power consumption of existing computing devices. The generated training dataset is utilized to train a machine learning (ML) model which is tested using a variety of hardware combinations. The ML model is verified using a test scenario and computing device configurations that were excluded from the training dataset. Upon successful verification, the ML model can be used for estimating power consumption of computing devices under design based on their component combinations.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.