Predicting application performance on hardware accelerators
AssigneeInternational Business Machines Corporation
InventorsIoana M. Baldini Soares, Stephen J. Fink
Life of this patent
Estimating the term…
Abstract
Predicting program performance on hardware devices, in one aspect, may comprise obtaining a set of existing applications and observed performance on a target hardware device. The set of existing applications are run on one or more general purpose computer processors and application features are extracted from the existing application. A machine learning technique is employed to train a predictive model based on the extracted application features and the observed performance for predicting application performance on the target hardware device.
Claims20
A method of predicting program performance on hardware devices, comprising:
- obtaining a set of existing applications and observed performance on a target hardware device;
- running the set of existing applications on one or more general purpose computer processors and extracting application features from the existing application;
- inputting the application features and the observed performance on a target hardware device to a machine learning technique;
- executing the machine learning technique;
- and
- training a predictive model by the machine learning technique for predicting application performance on the target hardware device.
A method of predicting a hardware device for best program performance, comprising:
- obtaining a plurality of existing applications and observed performance on a plurality of target hardware devices, each of the plurality of existing applications labeled with one of the plurality of target hardware devices;
- running the plurality of existing applications on one or more general purpose computer processors and extracting application features from the existing application;
- inputting the application features, labels associated with the existing applications, and the observed performance on the plurality of target hardware device to a machine learning technique;
- executing the machine learning technique;
- and
- training a predictive model by the machine learning technique for predicting a target hardware device out of the plurality of target hardware devices for running a given application.
A computer readable storage medium storing a program of instructions executable by a machine to perform a method of predicting program performance on hardware devices, the method comprising:
- obtaining a set of existing applications and observed performance on a target hardware device;
- running the set of existing applications on one or more general purpose computer processors and extracting application features from the existing application;
- inputting the application features and the observed performance on a target hardware device to a machine learning technique;
- executing the machine learning technique;
- and
- training a predictive model by the machine learning technique for predicting application performance on the target hardware device.
A system for predicting program performance on hardware devices, the method comprising:
- a set of existing applications and observed performance on a target hardware device;
- one or more general purpose computer processors, the one or more general purpose computer processors operable to execute the set of existing applications and extract application features from the existing application;
- and
- a machine learning algorithm operable to run on the one or more general purpose computer processors and further operable to train a predictive model for predicting application performance on the target hardware device, based on the application features and the observed performance on the target hardware device.
Dependencies are inferred from the claim text.