[Interest] Interest Digest, Vol 79, Issue 21

Roland Hughes roland at logikalsolutions.com
Mon Apr 30 23:47:40 CEST 2018

> On Thursday, 26 April 2018 01:27:59 PDT Jason H wrote:
>> Nah. Big endian is the way to go. Just because one company made little
>> endian prevelant isn't good enough. Network byte order is big endian, and
>> when looking at hex dumps, the bytes appear in the order that you expect to
>> see them... Which is why I think the big endian deadline was chosen. You're
>> sending data with QDataStream, network byte order is the right order. The
>> bits keep getting more significant, right to left, rather than zig zagging
>> right to left then over to the right for the next byte.
You can be as sarcastic as you want viewing the world from the throw 
away x86 hobby chip grain of sand on the beach. And no, one little 
company, especially if you mean INTEL didn't make little-endian popular.


When Intel developed the8008 
<https://en.wikipedia.org/wiki/Intel_8008>microprocessor for Datapoint, 
they used little-endian for compatibility. However, as Intel was unable 
to deliver the 8008 in time, Datapoint used amedium scale integration 


The Digital Equipment Corporation PDP-10 (1964-1983) is one of the most 
influential computers in history in more ways than can be listed here. 
It was the foundation of the DECsystem-10 and the DECSYSTEM-20 and ran a 
variety of operating systems including TOPS-10, ITS, WAITS, TYMCOM-X, 
TENEX, and TOPS-20. It was the first widely used timesharing system. It 
was the basis of the ARPANET (now Internet). It was the platform upon 
which many of today's popular applications were first developed 
including EMACS, TeX, ISPELL (the first spell-checker), and Kermit.

The big iron of the day, DEC-10, DEC-20 and IBM all used big-endian to 
achieve massive throughput. When DEC started making smaller PDP-11 
machines they went to mixed-endian and dramatically reduced performance. 
While the 11/44 was a massive seller and still exists under the hoods of 
many other devices today, it was sloooow compared to the big boxes.

VAX hardware went little-endian to be more compatible with all of the 
knock-off midrange computers and and the wanna-be dual floppy computers. 
It suffered profusely from I/O throughput as a result.

We won't discuss what happened when the Alpha chip was being "second 
sourced" at INTEL.

For whatever reason the highest throughput machines have always been and 
continue to be big-endian. Don't confuse "throughput" with some rigged 
count of numerical calculations. Throughput is the ability to 
read-calculate-store complete transactions.


        There are more than 30,000 accelerators in operation around the

Tiny wanna-be computers get built into sensors placed around these 
things and they transfer back to big-endian machines for serious 
throughput. The simple act of gathering sensor data and packetizing it 
can easily exceed the throw-away chip's capability so adding more 
workload to it cannot really be considered.

Roland Hughes, President
Logikal Solutions


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.qt-project.org/pipermail/interest/attachments/20180430/8a90e7d7/attachment.html>

More information about the Interest mailing list