[Info-vax] RealWorldTech on Poulson

[Johnny Billquist]

On 2011-07-04 17.54, Michael Kraemer wrote:
> In article<iusece$760$1 at Iltempo.Update.UU.SE>, Johnny Billquist
> <bqt at softjar.se>  writes:
>>
>> Well, the Alpha was not really a latecomer.
>
> It was.
> The first Alpha computers you could buy as late as early 1993.
> The first Power's arrived in 1990, the first HP "Snakes"
> in 1991, iirc, just to name a few competitors,
> not to mention Sun and SGI who had their stuff out much earlier.
> These few years made the difference between live or die.

True, if you talk about RISC in general, then yes, Alpha was definitely 
late.

>> It's also about 20 years old
>> now.
>
> hindsight doesn't matter in this case,
> the early years made the difference.

It definitely didn't help that DEC took their time to commit to a RISC 
architecture, yes.

>> And it was the first high speed CPU out there,
>
> it probably was the highest clocked CPU,
> but this doesn't necessarily translate into
> the fastest computers.
> If you look at DEC's initial lineup (3000 AXP)
> they had to clock it up to 200 MHz just to marginally
> beat an HP 755 (@99Mhz) or an IBM 580 (@60MHz).
> http://tinyurl.com/64j4go8
> And we did not have discussed price yet.
> Not very impressive IMHO for an architecture already three years late.

So, the HP 755 (which shows the better numbers) gives about 80/150 for 
int and fp, while the Alphas best is 130/200. That's more than what I'd 
call "marginally better". But the big problems with the SpecINT and 
SpecFP tests that they are so dependant on compilers that they don't 
really show CPU speeds that you can compare, but show how good your 
compiler is at compiling for those tests. And yes, compilers many times 
had (do they still?) code that specifically detected these tests, and 
popped out specially tuned code for them. So it's questionable how 
relevant information from such tests are.

I remember walking around various fairs at the time the Alpha was new, 
and all competitors machines were pretty slow in comparison at the time. 
It was fun to watch all the vendors trying to show off, and it was 
painfully obvious that compared to the Alpha they were slow.

>> as well as the first
>> clean 64 bit CPU.
>
> That was the Mips R4000.
> And given the RAM constraints of that time,
> 64bit addressing was almost useless.

You're right. The R4000 was introduced in 1991, while the Alpha only 
came in 1992. Got me. :-)
But I wasn't just talking about addressing, even though that too is 
nice. But also, all computations and other stuff that requires more than 
32 bits will also make a big difference.

>> It had a lot going for it. However, it was tied to a
>> company that was sinking, which is a part of what made it fail.
>
> Alpha was just too heavy for relatively small company like DEC.

DEC was the second largest computer company in the world only a few 
years earlier than the Alpha. To call it a "small" company is kindof 
misstating facts. It was probably still, by the time the Alpha was 
introduced, larger than SUN. However, DEC was definitely not going in 
the right direction. Mismanagement at many different places.

	Johnny