[email protected] (Bill Hitefield) writes: > Perhaps I did not accurately read the introductory post for this > topic, but if we are considering 360 and its descendants, I am > surprised no one has mentioned either SVS or VS1 (though I did see > where MVS was mentioned). One of my first jobs out of college (70s) > involved a conversion from SVS to VS2. A later job involved bringing a > site from VS1 to MVS/370.
VS1 was MFT layed out in 4mbyte virtual address space. VS2 started out MVT layed out in 16mbyte virtual address space (SVS). VS2 was then enhanced to multiple 16mbyte virtual address spaces (MVS) and VS2 designation was droped, just being refereed to as MVS. The initial VS2 development work was done on 360/67 ... a little bit of code to build a 16mbyte virtual memory table at startup .. and then a little bit of code to handle page faults. Not a whole lot of difference from running MVT in a 16mbyte virtual machine. The biggest effort was because applications built channel programs (now with virtual addresses and CCW execution required real addresses) and passed the channel program with EXCP/SVC0 for execution. To do this they borrowed the CP/67 CCWTRANS channel program translator that created a copy of the channel program, substituting real addresses for virtual addresses (CCWTRANS was the largest amount of code moving from MVT to SVS). Discussed in this old post about motivation for moving all of 370 to virtual memory http://www.garlic.com/~lynn/2011d.html#73 Multiple Virtual Memory the above references that VS2 release 1 was SVS, VS2 release 2 was MVS and VS2 release 3 was suppose to be the first Future System release http://www.garlic.com/~lynn/submain.html#futuresys with the failure of FS, 303x (repackaged 158 & 168 as 3031 & 3032, and 3033 was 168 logic remapped to 20% faster chips) and 370XA (3081) were kicked off in parallel ... some additional discussion http://www.jfsowa.com/computer/memo125.htm Endicott did "E-architecture" (as alternative to POK's XA architecture). This was primarily moving (single) virtual memory address space management into microcode ... which then was VS/E and DOS/VSE. The first endicott machines after FS was 138/148 ... which had VM/370 microcode assest ... and Endicott tried to have VM/370 shipped on every machine from the factory (sort of like LPAR currently ships). However POK managed to get corporate to overrule that ... since POK was in the process of getting VM/370 product killed and all the development people moved to POK and assigned to MVS/XA (Endicott eventually saved the VM/370 product mission, but had to reconstitute a development group from scratch). Old post about Endicott suckers me into working on ECPS, the VM microcode assist. http://www.garlic.com/~lynn/94.html#21 370 ECPS VM microcode assist Note: with ECPS and VS1/VM370 handshaking, VS1 would run faster under vm370 than it ran on bare machine. A little garbled here ... https://www-03.ibm.com/systems/z/os/zvse/about/history1970s.html During this period, the practice of charging for IBM software became widespread. ... snip ... Application software started being charged for after the 23Jun1969 unbundling announcement, but they managed to make the case that system software would still be free. http://www.garlic.com/~lynn/submain.html#unbundle It was in the wake of Future System failure along with the clone processor makers getting market foothold (because of lack of 370 products during the FS period) ... that the decision was made to transition to start charging for system software. -- virtualization experience starting Jan1968, online at home since Mar1970 ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
