Subject: RE: Relative performance?
To: 'Chuck McManis' <cmcmanis@mcmanis.com>
From: Carlini, Antonio <Antonio.Carlini@riverstonenet.com>
List: port-vax
Date: 06/30/2001 09:48:43
> Chuck McManis wrote:
>Ok, the HW FAQ however shows the 4000/100A as being NVAX @ 72Mhz
>and shows the 4000/90 is NVAX @ 72Mhz and 40 VUPS. And called a
"Cheetah-Q"
>and the 3100/M90 is called the Cheetah, so presumably its the
3100/M90 with
>the addition of the Qbus interface.
The VAXstation 4000/90 stuff often gets confused - not least by me
:-)
(I've had a VS4000-90 on my desk for quite a while ... only when I
finally
twigged that it claimed to be 83MHz did I realise that it was a
VS4000-90A ... the badge on the front has not indication)
The VAXstation 4000 Performance Summary (EC-N0857-51)
has the VS4000-90 clocking in at 32.7 SPECmarks and
about 32VUPs. It indeed has a 72MHz clock (well 71MHz according
to the doc ...) and a cycle time of 14ns. The backup cache is 256KB.
The VAXstation 4000-96 (the fastest VAXstation ever made I think)
has a 10ns cycle time but I have no information about a VUP
rating ... 40 seems perfectly believeable.
The VAX 4000-100 uses a 72MHz 14ns NVAX (same as the
VS4000-90) but with only 128KB backup cache. The
VAX 4000-100A is (IIRC) the same hardware but with
a duaghtercard on the CPU module that allows access
to a few more SHACs to improve I/O capacity.
The VAX 4000-100 and the MicroVAX 3100-90 are identical.
The motherboard is the same. It can be persuaded to adopt one
or other personality with the appropriate console TEST command.
To actually use the DSSI or Q-bus of the VAX 4000-100 you do
need the right enclosure.
Similarly the VAX 4000-105A is a MicroVAX 3100-95, the
VAX 4000-106A is a MicroVAX 3100-95, the VAX 4000-106A
is a MicroVAX 3100-96 and the VAX 4000-108A is a
MicroVAX 3100-98. (This is the only case I can recall
where changing the CPU personality changes the KA number).
Antonio