IMHO, the fault lies in the character stream orientation of UNIX, C, HTML
etc.  The shorted-sighted design was motivated by the limited budgets and
underpowered systems of many early UNIX users.

On record oriented systems, (z/OS and z/VM) common operations are faster,
because the needed information is not embedded in the data.  For example:

   - Read/skip-to the next record.
   - Find/check the length of a string.

On byte stream oriented systems, every single character, including the
otherwise uninteresting ones, must go through the CPU for such operations.
Record oriented systems can efficiently add the record length to the
current record address, or compare a target character position to the
length of the record to avoid string overflow (e.g).

It doesn't help that most character oriented systems use 16 bit characters
whereas most work on zSeries is done with 8-byte characters.  The workload,
all other things being equal, is essentially doubled.  As stated above, all
other things are anything but equal.

Note that UNIX (etc.) systems originally used 4K buffers between pipe
stages;  Most such systems now use 64K buffers to support the heavy load of
characters run through the CPU and to reduce costly redispatching.

It may be related that IBM went from two cache levels to four about the
time USS was added.  More hardware manufacturing and heat generation
resulted.

Unfortunately this mistake is so pervasive that it might never be fixed.
The impact is not just on electricity usage, but also on global warming and
climate change due to extra cooling and extra hardware.

Anyone who understands that global warming and climate change are
existential threats and that it may be too late to avoid catastrophic
impacts would be well advised to keep their record oriented systems and
move away from UNIX, Linux, and Windows where feasible.

Just my "buck three eighty", or two cents if you prefer.

OREXXMan
Q: What do you call the residence of the ungulate with the largest antlers?
A: A moose pad.
:-D
Would you rather pass data in move mode (*nix piping) or locate mode
(Pipes) or via disk (JCL)?  Why do you think you rarely see *nix commands
with more than a dozen filters, while Pipelines specifications are commonly
over 100s of stages, and 1000s of stages are not uncommon.
REXX is the new C.


On Sat, Feb 11, 2023 at 8:37 PM Bill Johnson <
00000047540adefe-dmarc-requ...@listserv.ua.edu> wrote:

> Correct. I copied the article from the NYT & then reposted the paragraph
> in the article which discussed the study.
>
>
>
> Heh - I don't think those are rankings - just (former) links from the
> article in whatever publication Bill copied from.
>
>
> > >    ...
> > >The largest cloud data centers, sometimes the size of football fields,
> > are owned and operated by big tech companies like Google, Microsoft,
> Amazon
> > and Facebook.
> > >    ...
> > >Over the years, data center electricity consumption has been a story of
> > economic incentives and technology advances combining to tackle a
> problem.
> > >
> > Do they use:
> > o IBM z?
> > o IBM supercomputers?
> > o Others, such as overseas-sourced (specify)?
> >
>
> At one time Facebook published detailed specs for its homegrown PC servers,
> in contrast to the likes of Google, Amazon, and Microsoft, for whom it's
> all trade secrets. I've no idea if they've kept the specs current. Lynn
> Wheeler wrote about this stuff a number of times when he was active on
> IBM-MAIN, though mostly from an available-compute-power perspective rather
> than a power efficiency one.
>
> Tony H.
>
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
>
>
>
>
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
>

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Reply via email to