On 26 February 2018 at 13:42, Paul Gilmartin <
[email protected]> wrote:
> On Mon, 26 Feb 2018 18:03:58 +0000, Seymour J Metz wrote:
>> ...
>>Why should the descriptions of a software construct agree with the
hardware used to support it. A TU represents the effective interval for
changing bit 30, not bit 31.
>>
> I feel badly misled by the ply that said:
> Timer units ... come from the long-gone S/360 Interval Timer,
>
> ... followed by a long description of that Interval timer. Factually
correct,
> but I allowed myself to misbelieve they were identical, not just "come
from".
My apologies. Clearly I also misled myself by allowing memory to override
logic. Though I think I did accomplish my point in pointing out that Timer
Units of either size are not TOD Clock units, or anything that can be
derived therefrom by some simple bit shifting.
> It would have been clearer to have stated:
> Timer units ... have twice the value of units of the long-gone
> S/360 Interval Timer,
>
> The macro description, properly, does not mention the Interval Timer.
So is there a definition of these Timer Units beyond that given in the
various macros?
And FYI, I found several SRA Rn,1 instructions in OS/360 that indeed
convert the value in the timer to OS/360 Timer Units, e.g.
L XN2,TIMER .GET TIMER VALUE
SRA XN2,1 .DEV 26 USEC UNITS
Tony H.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
