I am trying to come up with a complete mapping of 4-digit machine types to C/C++ ARCH levels. (The universe for my "complete" is z architecture.)
I am in pretty good shape for z900 through z13. I am dealing with future machines by assuming that anything not in my table is of an architecture level higher than those I know about. For my application this is a safe assumption. (I can failsafe in that direction; I do not want to fail in the other direction: report that the hardware is down-level when in fact it is not.) Google tells me that the zPDT is machine type 1090 or 1091. Is there a correlation to architecture level? I kind of get the feeling that IBM ships a software update from time to time that updates the instruction set to *some* level. Is there a way to determine the architecture level of a zPDT instance other than by testing feature bits in the CVT? How does one set the C or COBOL ARCH level if one is targeting zPDT? FSI boxes are 390 only, right? Outside of the z universe? VM does not generally alter the problem instruction set of a guest, right, nor change the machine type number? (I recall that one can change the serial number.) If VM is running on a z196, every guest sees a machine type of 2817, correct? VM neither adds store on condition to a machine that does not support it at the hardware level, nor deletes it from a machine that does? I know that z/OS is not licensed on Hercules but hypothetically if someone were to run z/OS on Hercules, the machine type reported and the instruction set are consistent - is that correct? If someone is running Hercules in for example "z10 mode" then the machine type reported will be 2097/8 and decimal floating point will be present but load on condition will not? Charles ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
