Info-3b2 Digest, Number 8

Info-3b2 Mailing List info-3b2 at lamc
Sun Aug 21 18:27:31 AEST 1988


 
                           Info-3b2 Digest, Number 8
 
                           Saturday, August 20th 1988
 
Today's Topics:
 
                                  Re: 3B2-500s
                                  Re:  3B2-500s
----------------------------------------------------------------------
 
Subject: Re: 3B2-500s
Date: Fri, 19 Aug 88 23:21:47 CDT
From: Rich Andrews <sun!texsun!rich%jolnet>

> 
> --- Forwarded mail from necntc!pyramid!cepu!john 
> 
> >From necntc!pyramid!cepu!john  Thu Aug 18 22:45:01 1988 
> Subject: 3B2-500s
> To: bradley!info-3b2
> Date: 16 Aug 88 15:15:38 CDT (Tue)
> From: sundc!sun!bradley!john at EDDIE.MIT.EDU (John Lengeling)
> Message-Id: <8808161515.AA01762 at bradley.UUCP>
> 
> I am thinking about replacing a 310 with a 500, and I was wondering
> how a 3B2/500 compares to a 310.  
> 
> How much of a performance improvement is a 500 over a 310?
> What kind of performance increase can be achieved with adding a:
> 
> 	VCACHE
> 	22Mhz system board
> 	multiple processors
> 
> The 310 that I am using is currently 13 simultanious users running large
> applications (>350K executables)  The main problem with the machine is that
> it needs more memory...
> 
> ----------------------------------------------------------------------------
> John Lengeling			UUCP: {cepu,att,uiucdcs,noao}!bradley!john
> Bradley University		ARPA: cepu!bradley!john at seas.ucla.edu
> Distributed Systems  		ATTMAIL: attmail!bradley!john
> Bradley Hall Room 6F            PHONE: (309) 677-2337
> Peoria, IL 61625
> 
> --- End of forwarded message from necntc!pyramid!cepu!john 
> 
> -- 
> To submit articles mail to ..!{pacbell,well,netsys,hoptoad}!lamc!info-3b2
> Administravia to info-3b2-request at lamc.
> Co-ordinator: kdavis at lamc (Ken Davis)
> 

upgrading to a 3b2/500 would be an incredible improvement!  It has to be 
experienced!  (right charlie?)  I know for a fact that when killer was 
a 3b2/400 with 4 meg it was 4 days behind in spooling up the news for 
other sites.  It took the 500 about 4 hours to catch up!  Impressive
and awe inspiring!

rich andrews 
SA for jolnet

------------------------------
 
Date: Fri, 19 Aug 88 18:24 CDT
From: sysop at killer.Dallas.TX.US (Charles Boykin-BBS Admin)
Subject: Re:  3B2-500s

   I recently changed killer from a 3B2/400 to a 3B2/500 with 8 MB memory, a
vcache, and a MPB. The 400 was four days behind unpacking/uncompressing the
news batches and was pretty slow. The 500 unpacked the news backlog within
two hours and was completely up to date in 2 1/2 while the CPU was running
at 60% idle during all this. The kernel still has not been tuned for any
change in performance from default and the machine will only go into swapping
when four uncompress processes, two compress processes, and more than five
users on the system - and the most I have seen it swap is ~400 blocks. With
the MPB, the system is rated at 4.0 MIPS with 16ms average disk access time -
I have 2.1 GIG disk space on the system. So, to answer your question, there
is really no comparison between the 310 and the 500 - the 500 is far faster.
The 3B2/400 uses the same system board as the 310 so the machines are virtually
identical except for the ports card cage and the physical capacity of two
hard disks internal. There is an additional enhancement now available for the
500 - a 22 MHZ CPU that will handle two MPB boards and can use the new 16MB
RAM cards - that are the same size as the 4MB cards. This allows 32MB RAM
with the 22MHZ CPU with comparable performance improvements.

                                                   Charlie Boykin
                                                   ..killer!root

 
-------------------------------------
 
To join this group or have your thoughts in the next issue, please
send electronic mail to Ken Davis at the following address;
 
	{pacbell,netsys,hoptoad,well}!lamc!info-3b2-request
 
The views expressed in Info-3b2 Digest are those of the 
individual authors only.
 
*********************
End of Info-3b2 Digest
*********************



More information about the Comp.sys.att mailing list