Open Access. Powered by Scholars. Published by Universities.®
Physical Sciences and Mathematics Commons™
Open Access. Powered by Scholars. Published by Universities.®
Articles 1 - 2 of 2
Full-Text Articles in Physical Sciences and Mathematics
Time Series Models For Predicting Application Gpu Utilization And Power Draw Based On Trace Data, Dorothy Xiaoshuang Parry
Time Series Models For Predicting Application Gpu Utilization And Power Draw Based On Trace Data, Dorothy Xiaoshuang Parry
Electrical & Computer Engineering Theses & Dissertations
This work explores collecting performance metrics and leveraging various statistical and machine learning time series predictive models on a memory-intensive application, Inception v3. Trace data collected using nvidia-smi measured GPU utilization and power draw for two runs of Inception3. Experimental results from the statistical and machine learning-based time series predictive algorithms showed that the predictions from statistical-based models were unable to capture the complex changes in the trace data. The Probabilistic TNN model provided the best results for the power draw trace, according to the test evaluation metrics. For the GPU utilization trace, the RNN models produced the most accurate …
Modelling Locally Changing Variance Structured Time Series Data By Using Breakpoints Bootstrap Filtering, Rajan Lamichhane
Modelling Locally Changing Variance Structured Time Series Data By Using Breakpoints Bootstrap Filtering, Rajan Lamichhane
Mathematics & Statistics Theses & Dissertations
Stochastic processes have applications in many areas such as oceanography and engineering. Special classes of such processes deal with time series of sparse data. Studies in such cases focus in the analysis, construction and prediction in parametric models. Here, we assume several non-linear time series with additive noise components, and the model fitting is proposed in two stages. The first stage identifies the density using all the clusters information, without specifying any prior knowledge of the underlying distribution function of the time series. The effect of covariates is controlled by fitting the linear regression model with serially correlated errors. In …