On Tue, Feb 14, 2017 at 7:30 AM, scott Ford <[email protected]> wrote:

> All,
>
> The STC is runnIng in a MVS address space, pretty standard. I am looking
> for a performance boost.
> We currently issue a call and the HLASM code performs the extract and
> creates a file that the STC reads.
> My question is what gives us the best performance, my experience says
> memory ..and not disk storage because of the I/O.
>
> Is my assumption correct ?
>

​I think so. This is assuming that your address space is not memory
constrained. ​Using a PIPE, as in a previous message on this thread,
basically uses memory, but does require "intervention" by the UNIX kernel
services. This may (I don't know) require a cross memory move from your
address space to the kernel buffer (on the write) then back to your address
space (on the read). Personally, if I were doing this "from scratch" and
not trying to adapt existing code, I would simply have the COBOL routine do
a standard CALL of the HLASM routine. The HLASM routine would use "static
variables" to main its place between calls and do the RACF extraction "as
needed", and just pass a buffer to the HLASM into which it places the next
"record". I might even consider replacing the HLASM with a COBOL subroutine
or, at an extreme, a COBOL nested program. But I don't know what the HLASM
routine is doing, so that might not be possible. Nor do I know what version
of COBOL you're using. COBOL 6.1 is far superior to the 3.4.1 that I have
available.



>
> Thanks for the help much appreciated.
>
> Scott
>


-- 
"Irrigation of the land with sewater desalinated by fusion power is
ancient. It's called 'rain'." -- Michael McClary, in alt.fusion

Maranatha! <><
John McKown

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to