<It states: "The VAX 6500 processor delivered approximately 13 times the
<power of a VAX-11/780 system, per processor."
Your confusing the price of bricks with the weight of concrete.
The MIPS metric is not accurate, I don't know if there is one. Further
VAX doesn't spend all its cycles doing drystones. In the real world
you do some math then do a LOT of housekeeping.
<To me, that means "13 MIPS". 13 MIPS is about 2 to almost 3 times slower
<than a 486dx2/66.
Until you fill memory. then the contest reverses. The 486s of the time
were doing good at maybe 5meg bytes/sec and thats SLOW. The real test
then and now is what the system does with relational database of say
100mb is size.
< CPU MIPS MIPS
< System OS CPU (MHz) V1.1 V2.1
< ---------------------- ------------ ----------- ----- ------ ------
< VAX 8650 4.3 BSD ----------- 18 6.3 6.2
< cc -non_shared -DUNIX -O5 -ifo
It's a non compare as an 8650 is not a 6000 anything. there are differnet
memory bandwidths and IO bandwiths in effect and the slowest is much faster
in action than an 8650.
<And, back of the envelope, the 6500 was 2x the 8650.
And it was 1/5 the volume.
<I don't know what it is about collectors that somehow confuses their
<memories of the past; maybe their internal core memories have suffered som
<bit flips? ;-) (I have some old junk, too, so I consider myself in the
<same camp...)
Well, first of I'm not your average retrorevionist PC collector. I worked
for DEC for 10 years (83-93) spanning the era in quesiton.
<Fact is, these old machines were slow, noisy, hot, power-guzzling behemouth
<compared with what we have today.
Yes. Compared to the "hot" 486dx/50 in 1992 we were using the little
(they are 10wx4Hx16D) VS3100s (same size and lower power than your power
guzzling PC) as the PC crusher and those were typically 2.4-3.5VUP range.
Oh, and I've used a VS3100 recently as a server at work for test purposes
and the 3100/m10E clobbered the P166mmx/scsi box running NT.
<As you know, the vax would run the X -client- which, of course, is not
<much of a load. And as for running 50 users, heck, at the VCF 3.0 there
<was a guy showing an 8080 running timesharing on a bunch of terminals!
Yes, I know and have done. I've tried to run a p133 as a unix host (not
a web server) as a timeshare client and it doesn't load as well as the
VS2000 I have. The 8080 timeshare FYI had terrible latency!
<Therefore, to me, you have to measure the performance in some repeatable
<way. Dhrystone is not the perfect benchmark (which is close to an oxymoro
<anyway), but it is -a- benchmark for integer CPU performance.
Yes so... When I'm running a relational database that number means absolutly
nothing. Running a word processor it's means nothing.
<It's too bad that you had to run w3.11; I'm assuming you're running a real
<OS on the dx2/66.
In 1993 what real OS would have a PC running? DOS? Concurrent dos, SCO
unix? Venix? The hot IO interface was 10b2 and AHA5142 SCSI. The really
hot box was the 4cpu (486dx/50) box NCR made, there was no real OS yet and
w95 was still beta testing on that box.
<Now, let's talk about busses. Just how fast -was- this CI? Let's compare
<that with 66 MHz 64-bit PCI, which has 66e6 x 64 = 4,224,000,000 bits/sec
<peak throughput. What was CI's throughput?
CI, cluster innerconnect was was fast when PCs were running ISA-16. If
memory serves it was a 24-32mbytes/s rate. PCI-66 is now, and slow Alphas
use it. The difference is you could have multiple CI busses and sorry but,
only one PCI. Also PCI is not that fast. It's bursty and can run peak
rates that fast but what is the CPU doing when PCI is honking at that rate?
Answer? WAITING, if your real lucky running from the cache!
But then when VAX was running CI what was the PC running? OK ISA-1 at
some 8meg bytes/sec and memory interfaces were typically 70ns 32bit wide
non-interleaved with caches. SCSI was maybe 10mb/s. the less than
popular MCA bussed IBM hardware were much fast than the ISA and VL bussed
counterparts.
Running VMS in the early 90s we had SMP that was scaleable using LAVC
(cluster over eithernet) and CI. There were also the BI and other
interconnects.
The point being your trying to save your arguement with _now_ hardware
against _then_ hardware. PCs with then hardware were the industry joke
for uptime, reliablility and performance, the AS400s, VAX, Prime, and
others were the systems choice for getting work done on the larger scale.
PCs in many ways are still behind the "big iron" of the early 90s, as
they still lack a really good OS (linux, freebsd are contenders though).
The hardware is not mature, they keep creating new standards that barely
get debugged before they are pass`e. It's always amazing to see older
systems bumping along getting real work done, usually while the PC user
is rebooting for the third time today.
Allison
Received on Sun Oct 24 1999 - 17:42:04 BST