On Wed, 10 Jul 2024 at 11:25, Steve Thompson <[email protected]> wrote:
> I seem to remember something said about an extension to packed > decimal and new instructions. I've been looking at a recent > z/Arch PoOP in pdf (IBM web site), and I can't seem to find what > I hazily remember from a few years ago. Perhaps it is in a > different publication. > > Could someone point me in the right direction, even if that is, > you must have been dreaming. > Decimal Floating Point. Chapter 20 in the -13 edition of SA22-7832. If you want to get fancier, there are also Vector Decimal instructions in Chapter 25. The issue is, how can COBOL handle larger packed decimal numbers > than PACK/UNPACK can handle? > The question of how/whether COBOL exploits this facility is a different one, of course, but I gather it does. > I see this being a question that is going to get asked on a > project I'm on.... And I can see special macros having to be > developed for this.... > Well perhaps... Certainly using the DFP instructions is quite different both practically and conceptually from the traditional packed decimal. But it will allow you to handle larger (and smaller) numbers, and has all kinds of other benefits, including some consistency with, and easier conversion to/from, the binary FP formats. Tony H. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
