What Makes a PDP-11/35 or 40 Tick?

[Liam Proven]

On Wed, 12 Jun 2019 at 01:42, Fred Cisin via cctalk
<cctalk at classiccmp.org> wrote:
>
> IFF your old post had a solution!  (rather than being an unresolved
> query)

https://xkcd.com/979/

There are a couple of things now where I know that if I try to find
sources, I'll find myself.

The details of how id Software's game /Quake/ killed Cyrix, the CPU
company, because Quake used a very clever method of interleaving FP
ops with integer ops that didn't use the same registers, allowing the
Pentium core to essentially run them at once. The Cyrix 6x86 core was
quicker -- I bought one myself, what more can I say -- but it couldn't
interleave FP and int ops like that, so it ran Quake like a dog.

And as Quake was *the* most demanding game around in 1996 or so, and
had a super convenient built-in "timedemo" feature, all the magazine's
benchmarks used Quake. So 6x86s got hammered, even though on literally
*ALL* OTHER X86-32 CODE THEY WERE QUICKER, they got bad reviews, and
Cyrix ended up getting acquired by Via.

Or the problems with AMD fglrx on a Toshiba Satellite Pro P300A.
That'll be me, too.

The weirdest one, though, was when I was given the honour (?) of
moderating my first item at the World SF Convention. It was a
collection of silly games. I'd never heard of one of them. I Googled
it. I got my own blog as the № 2 or № 3 hit.

_That_ was odd.

-- 
Liam Proven - Profile: https://about.me/liamproven
Email: lproven at cix.co.uk - Google Mail/Hangouts/Plus: lproven at gmail.com
Twitter/Facebook/Flickr: lproven - Skype/LinkedIn: liamproven
UK: +44 7939-087884 - ČR (+ WhatsApp/Telegram/Signal): +420 702 829 053