Home > Administer Data Miner > Workflow Throughput
Your DBA can help optimize throughput. The following have an impact on Workflow throughput, that is, the time and resources required to execute Data Miner workflows:
Most data mining algorithms can take advantage of parallel execution when it is enabled in the database. Parameters in INIT.ORA
control the behavior of parallel execution.
The Data Miner Repository setting MAX_NUM_THREADS
controls how many parallel model builds are allowed across the entire database. This setting only applies to models being created by workflows.
The default setting is 10 models. You can increase or decrease this depending on your machine capacity. Change Number of Concurrent Model Builds explains how to change this value.
Data Miner workflow execution is scheduled using Oracle Scheduler. There are several ways to manage schedule jobs:
Oracle Scheduler does automatic resource management that limits the number of workflows allowed to run based on current system utilization.
You can create a resource plan to prioritize which Oracle Scheduler jobs (in Data Miner's case, jobs are called workflows) run first.
You can set up specific Scheduler windows that run workflows during specific time periods.
Data Miner workflows use the default Oracle Scheduler job class DEFAULT_JOB_CLASS
. You can change the job class to a job class that you create as described in Change Oracle Scheduler Job Class.
For information about Oracle Scheduler, see the Oracle Database Administrator's Guide.