9370 is a dog! Was: Re: Re[2]: Firsts

From: Philip.Belben_at_powertech.co.uk <(Philip.Belben_at_powertech.co.uk)>
Date: Mon Feb 16 11:33:27 1998

Sorry to take so long replying - I have been away for a week doing various
things (Job interview, visit parents, fail to mend car, etc.). One of the
advantages of being away is that a couple of dozen messages after a
question I think I could answer, somebody else answers it for me! Far
less typing to do!

> Certainly more _fun_ than some of the films that have been released for
> sure! The wife will not agree though. But if one of us had a large old
> S/360 or S/370 in our basements, the power consumption would financially
> kill us (because of the water chillers for cooling, current consumption
> of older technology electronics, etc.) I'd love to even _see_ a large
> S/360 and all its utilities.

Possibly. But I'm not convinced. I seem to remember the biggest IBM
3090 system drew about 250kW (I'll have to look that up too!). I
imagine a typical system might draw perhaps 50kW. The biggest problem
with this is the supply to your house - few can manage more than 25kW.
At 8c per kWh, 50kW is $4 per hour - cheap enough to run for an evening
a month, say.

> Only one 8 Meg board installed. Maximum, IIRC, was 16 Mb storage. DASD was
> probably used as workstation storage. Philip, can you confirm any of this
> from your sales literature? If you're interested, William, I can give you
> the lineup of PC board modules in the CPU which I briefly talked about
> earlier in the thread.

9373 (9370 model 20) was 4 MB expandable to 16 MB; all the others were
8MB, expandable to 16 MB. However, only the 9377 (model 90) had a
64-bit data path; all the others were 32 bit, so I suspect the 4 meg
card would work in the 9375. But there's no knowing, with IBM.

I see on this page there is a comparison "not fully verified by IBM" by
an outside agency, comparing 370 machines with Vaxen. A 9373 is slower
than a microvax II; a 9377, rather faster than a Vax 8650, reaching
nearly 0.8 Mflop. The 4381 - another low(ish) end 370 - had three
models listed, up to around 1.2 Mflop for the model 13. (At least IBM
didn't try and claim twice this for the model 14, which is two 13s
back-to-back).

Philip
Received on Mon Feb 16 1998 - 11:33:27 GMT