The batch loader combines the basic editing and loading services (that can also be provided by the linkage editor and program fetch) into one job step. The batch loader accepts object modules and load modules, and loads them into virtual storage for execution. Unlike the binder and linkage editor, the batch loader does not produce load modules that can be stored in program libraries. The batch loader prepares the executable program in storage and passes control to it directly. The batch loader cannot accept program objects, GOFF object modules, or control statements as input.

The batch loader provides high performance link-loading of programs that require only basic linking and loading, and can be used if the program only requires listing control or other processing options. Because of its limited options and ability to process a job in one job step, the batch loader only requires about half the combined linking and loading time of the linkage editor and program fetch.

Batch loader processing is performed in a load step, which is equivalent to the link-edit and go steps of the binder or linkage editor. The batch loader can be used for both compile-load and load jobs. It can include modules from a call library (SYSLIB), the link pack area (LPA), or both. The batch loader resolves external references between program modules and deletes duplicate copies of program modules. It also relocates all address constants so that control can be passed directly to the assigned entry point in virtual storage.

Like the other program management components, the batch loader supports addressing and residence mode attributes in 24-bit and 31-bit bit addressing ranges. The batch loader program is reenterable and therefore can reside in the resident link pack area.

Except for the processing of in-storage object modules, all of the services of the batch loader can be performed by the binder.