>All true Tom, but as far as I understand it the Vector Decimal instructions= > still do not provide any more digits of precision than the older, non-vect= >or ones. I believe the OP was asking about more digits of precision, not b= >etter CPU usage. COBOL's ARITH(EXTEND) option still provides only up to 31= > digits of precision. > >And we still do not have COBOL access to the Decimal Float capabilities of = >the hardware. Extended-format Decimal Float provides (if I am reading PoOP= > correctly) up to 34 significant decimal digits of precision.
The number of digits allowed in COBOL programs is not determined by the hardware, it is determined by the COBOL language definition (currently the COBOL 2014 standard) unlike assembler where it is defined by what instructions are available. The current standard says that 31 digits is the max. No IBM clients have requested the abiity to use more than 31 digits. As for DFP, users have been taking advanatge of DFP in COBOL programs since COBOL 5.1 in 2013, Basd on ARCH the compiler decides whether to use packed- decimal, DFP, or Vector Packed Decimal instructions to do arithmetic. Cheers, TomR >> COBOL is the Language of the Future! << ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
