michealb...@optonline.net (Micheal Butz) writes: > Who says I'm developing anything > > Dataspaces are more painfully and probably slower then regular storage > dataspaces were probably meant to > Releive storage constriant
multiple address spaces started in "811" (for nov78 date on the documents) architecture (370/xa appeared with 3081). subset was retrofitted to 3033 as "dual-address" space mode. os/360 was heavily pointer passing paradigm. in the initial morph to virtual memory, it was basically mvt laid out in single 16mbyte virtual address space (as SVS). In transition to multiple virtual address sapce (MVS), copy of the kernel image was mapped to (half/8mbytes of) every (16mbyte) application virtual address space. The problem was that there were some number of mvt subsystems that now found themselves in their own virtual address space (different from application). In order to support the pointer passing paradigm (with subsystems accessing application parameters), the common segment area (CSA) was created ... applications could stuff their parameters in the common segment, and generate a subsystem call (that required passing thru the kernel to switch to the subsystem virtual address space). At some larger installations CSA grew to 4-5 mbytes (leaving only 3-4 mbytes for application execution), some installations were facing prospect of CSA increase to 6mbytes (leaving only 2mbytes in every virtual address space for application execution). dual-address space mode allowed for parameter pointer to be passed to a subsystem and the parameters list could be access in the application virtual address space w/o requiring CSA (starting to cap the explosion in CSA size growth). Some of the larger MVS internal shops were chip design with large fortran applications that were seven mbytes that required carefully crafted MVS systems that kept CSA to 1mbyte max. There was a period when some these internal MVS premier shops were facing being forced to vm370/cms ... since they could get nearly the full 16mbyte virtual address space for their application execution (it was rather odd some of the hold-outs since in this period, vast majority of the internal mainframes were vm370). -- virtualization experience starting Jan1968, online at home since Mar1970 ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
