HP3000-L Archives

August 2000, Week 2

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Winston Kriger <[log in to unmask]>
Reply To:
Winston Kriger <[log in to unmask]>
Date:
Tue, 8 Aug 2000 18:18:04 -0600
Content-Type:
text/plain
Parts/Attachments:
text/plain (23 lines)
Gavin Scott wrote in message <[log in to unmask]>...

>One excuse for dropping the power-fail support was that modern CPUs now
>contain too much state (big caches, etc.) to flush out to main memory in the
>time available between when the line voltage drops and the time that the
>3000's power supplies can no longer support operation of the CPU.  Thus the
>requirement for a UPS starting around the time of the 9x8 boxes.


Since the 917/LX has a larger cache (96KB) than a '918/LX (64KB), the term
"excuse" seems appropriate.   I imagine either cache could be flushed to
main memory in about a millisecond or less, which would be no problem
for even a "cheap" power supply to buffer.  Of course, a 997 might need some
pretty big Capacitors to flush it's cache(s). I think the decision to start using
'ordinary' disk drives after HP decided to exit the Disk business might have
been another reason.  Most HP drives were power-fail aware and
could maintain consistency when the power died (I think the HP C2590AM
drive was the last of the Powerfail-proof drives).   Of course, I've seen
installations where the CPU Box had a UPS,  but not a rack of connected
external Seagate drives with no powerfail smarts  ........

Winston K.

ATOM RSS1 RSS2