On Mon, 18 Aug 2025 19:06:54 -0500, Jon Perryman <[email protected]> wrote:

>On Mon, 18 Aug 2025 18:35:30 -0500, Paul Edwards <[email protected]> wrote:
>
>>Personally, I don't think people should be relying on 64-bit
>>existing at all.
>>
>>It is funny that some people even complain about me using
>>more than 16 MiB of combined OS, code and data for "my"
>>(gcc 3.2.3) program 
>
>GCC is for MVS 3.8J

It's effectively NOT for MVS 3.8J. It realistically needs about
20 MiB, so MVS 3.8J is not suitable, so I created a different
environment (which I called MVS/380).

> which has a 16M (24-bit) optionally with a single 2GB (31 bit)
> address space. 16MB is a requirement for this environment.

>z/OS with IBM XL C does not have these restrictions.

And I didn't want those restrictions either, so I targeted an
environment that didn't have a hard limit of 16 MiB.

And yet people insisted that I should instead change "my"
code instead of seeking an alternative environment, because
"my" code was inherently "wrong" for being "so" memory-hungry,
where memory-hungry is defined as 16 MiB for combined
OS+app code+data.

>I suspect that z/OS DB2 is a heavy user of above the bar memory these days.

That's probably optional data caching. They probably only
have a few MB of actual executable code. Writing and
maintaining code requires actual work by a skilled engineer,
so there is a limit to how big it can be.

BFN. Paul.

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

Reply via email to