Multi-threaded feeders

Multi-threaded feeders improve the performance of bulk feeder construction and cube feeder updates by leveraging the amount of available CPU cores.

Important:

Multi-threaded feeders are not supported when your Cognos TM1 model uses conditional feeders. When issues are detected, you must disable the multi-threaded loading of individual cubes.

Multi-threaded query capabilities extend to feeders by the parallelization of processing steps. Configuration settings in the tm1s.cfg file enable/disable multi-threaded feeders and feeder construction at start-up. To apply multi-threaded feeder functionality, feeders must be well-defined to arrive at the same result, no matter what sequence of cell changes led to feeder generation.

For example, setting CellA = N and then setting CellB = M should render the same feeder results as setting CellB = M and then setting CellA = N.

While this requirement is not unique to multi-threaded feeders, it can have a more prominent effect when MTFeeders is enabled. When cell values are modified by users or TurboIntegrator, feeders are updated in response to individual cell changes and follow the sequence of cell updates. However, feeder regeneration (for example, during server startup) processes cells in the order they were stored in the cube. This example shows how sequence-dependent feeder definitions can lead to different generated feeder values even in single threaded mode.

For more information on multi-threaded feeder configuration options, refer to Parameters in the tm1s.cfg File in Planning Analytics Local Installation and Configuration.