[HECnet] Another TOPS-10 DECnet problem

[Thomas DeBellis]

You had a 20 that would handle 600 students in 1977/???/  Yikes!!  What 
kind of monster was that and how did you do this without melting?

As I recall, the maximum number of lines that an RSX-20F front end would 
handle with a full complement of DH11's is 128.  I have heard of a site 
(Case Western Reserve) having more than one front end handling lines, 
which I would imagine could get you up to 256.  There was also another 
site that moved the terminal handling code out of section 0, which was 
probably a great idea.

I also recall that the maximum job number that you could get on CFS was 
512 for a maximum of 128 jobs per node.

But I don't think I'm aware of any 20 doing more than about 70 active 
users without some serious load issues.  That would be a mix of text 
editing, email and maybe some iterative program development.  At 
Columbia, our standard 20 configuration was two DF20's (?) for (I think) 
3.5 MW.  We experimented with 'maxing' a 20 out with two DH20's and an 
MA20 for 3.75 MW, but it did not perform substantially better.

I'm not sure what the limiting factor was, whether it was paging or 
CPU.  If it was paging, then assuming you weren't doing too many 
multi-section programs, a 32MW Toad would clearly handle a higher number 
of users (and perhaps far higher), but you are then talking about 
contending for other things, RH20's, disks and at some point, you are 
just out of CPU.

There was a multi-processor version of TENEX, but nothing like that 
exists in any Tops-20 source that I've seen.
> ------------------------------------------------------------------------
> On 11/11/21 2:43 PM, Paul Koning wrote:
>
> :-)  They sure where logged out when the message "plato off" appeared on every terminal...
>
> 	paul
>
>> ------------------------------------------------------------------------
>> On Nov 11, 2021, at 2:29 PM, Steve Davidson<steve at davidson.net>  wrote:
>>
>> Paul,
>>
>> Don’t you mean 600 logged OUT students?
>>
>> -Steve
>>
>> SF:IP2
>>> ------------------------------------------------------------------------
>>> On Nov 11, 2021, at 13:45, Paul Koning<paulkoning at comcast.net>  wrote:
>>>
>>> Indeed.  I still remember the blistering I got (in 1977) when I added a new feature and the result was a system crash (buffer overflow due to falling behind) in the middle of the day, at max load.  Max load was 600 logged-in students.  Ouch.
>>>
>>>    paul
>>>
>>>> ------------------------------------------------------------------------
>>>> On Nov 11, 2021, at 1:25 PM, Thomas DeBellis<tommytimesharing at gmail.com>  wrote:
>>>>
>>>> ...
>>>> 100 students?  That's all?  Heh...  I can't even remember how many we had online with six 20's, but it didn't matter.  If one of them went out, you got your choice of machine room phone color: bright red hot or white hot.   Been there, done that.  Glad I'm not doing it any more.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sonic.net/pipermail/hecnet-list/attachments/20211111/63bee50f/attachment.htm>