This is a necessary phase prior to the normal EAR utilization and is a kind of hardware characterization of the nodes. During the phase a matrix of coefficients are calculated and stored. These coefficients will be used to predict the energy consumption and performance of each application.
Please, visit the learning phase wiki page to read the manual and the repository to get the scripts and the kernels.
The following table lists tools provided with EAR package to work with coefficients computed during the learning phase.
Name | Description | Basic arguments |
---|---|---|
coeffs_compute | Computes the learning coefficients. | <save path> <min_freq> <nodename> |
coeffs_default | Computes the default coefficients file. | |
coeffs_null | Creates a dummy configuration file to be used by EARD. | <coeff_path>, <max_freq> <min_freq> |
coeffs_show | Shows the computed coefficients file in text format. | <file_path> |
Use the argument
--help
to expand the application information and list the admitted flags.
Compute the coefficients for the node node1001
in which the minimum frequency set during the learning phase was 1900000 KHz
./coeffs_compute /etc/coeffs 1900000 node1001