[comp.lang.asm370] CLOCK

OPRJ38@TREARN.BITNET (OR) (09/06/90)

              I DON'T UNDERSTAND SOME FIELDS.
              WHAT ARE CVTTZ, CVTPTR, CVT DSECT ..
              AS I UNDERSTAND;
              THIS TIME IS 51 BIT MICROSECONDS, AND IT IS UPDATES
              EVERY 1 MICROSECOND. CORRECT?

-ORHAN-

JAF@UNB.CA ("Tony Fitzgerald 453-4573", 506) (09/06/90)

On  Thu, 6 Sep 90 06:09:31 EDT  OR <@OHSTVMA.IRCC.OHIO-STATE.
EDU:OPRJ38@TREARN.BITNET> writes:

>              I DON'T UNDERSTAND SOME FIELDS.
>              WHAT ARE CVTTZ, CVTPTR, CVT DSECT ..

  These are MVS fields.  CVTPTR is at location X'10' decimal 16 in the
PSA (Prefix Storage Area) or the first 1K of real and virtual storage
in MVS.  The CVTPTR is a single word containing the address of the
Communications Vector Table, at one time perhaps the single common
anchor point for most other OS data areas, still very important in MVS.
The CVT dsect is fully described in the MVS/XA Debugging Handbook,
LC28-1165, but which may not be available at your site unless you have
MVS licensed.  It was given in the example, only as one thing that might
have to be done to the TOD clock before using it.  The field CVTTZ
contains an adjustment that must be added to the high order word of the
TOD clock in an MVS environment to convert the TOD clock from GMT
(Coordinated Universal Time) to Local time.  In other systems, other
adjustments probably must be made.  Thus the argument, use the standard
services (the TIME macro in MVS) rather than the TOD clock if you want
to calculate dates or times.

>              AS I UNDERSTAND;
>              THIS TIME IS 51 BIT MICROSECONDS, AND IT IS UPDATES
>              EVERY 1 MICROSECOND. CORRECT?

Partly,  it depends on the speed of the CPU.  A fast CPU may update
a bit to the right of bit 51, a slow CPU to the left.  In any case,
whatever bit actually does the counting, it will be as though it were
bit 51 every micro second.  Note, however, that even the clocks of
multi-million dollar IBM CPUs are not 100% accurate and some drift may
be expected over time.

>
> -ORHAN-

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
-  J. Anthony Fitzgerald                  Phone:     (506) 453-4573
   Computing Services                     NetNorth:  JAF@UNB.ca
   The University of New Brunswick              or   JAF@UNB
   PO Box 4400
   Fredericton, N.B.  Canada
   E3B 5A3