Brian Greul (Exch)
bbraf08 at exch-imc.b-r.com
Tue Oct 7 12:47:46 EDT 1997
I think the reason that we have never seen one is that the way microsoft
allocates physical ram really makes it unnecessary.
Unfortunately there a couple of problems in our case.
We have a conflict of ideals.
NT is designed to be run on a UPS, microsloth makes no accomodations for
The RC5 client is trying to work around that.
Perhaps there is a more graceful way to accomplish the same thing
without terrorizing the hard drive?
I agree with an earlier post, let's make this an optional feature.
Under NT this could cause fragmentation in the NTFS partition. This
leads to really nasty things.
I have several machines which are idle the bulk of the time.
I would much rather see some work go into optimizing the client.
Adding disk processes makes it run slower.
Especially for IDE systems.
Under NT SCSI is a multitasking disk access system.
IDE is a serial command structure.
It is only able to process one command at a time.
This has awful ramifications for a multitasking O/S.
It means that you can't do a read and a write at the same time.
Most people who are participating in this project are probably using IDE
based systems anyway.
Let's try to reduce the number of disk accesses to a minimum.
You might try sub-blocking.
Where the PC receives a block and if the owner so designates the PC
splits the block into 10 or so pieces and crunches them individually.
That way if the PC dies, abends, or jumps out a window only part of the
block was processed.
The sub block that wasn't processed can be redone.
Granted there is some waste here, but I suspect the waste in this
process is smaller across the board than the waste from lost CPU time
and excess HD writes.
>From: Kevin van Haaren [SMTP:KvanHaaren at HNTB.com]
>Sent: Tuesday, October 07, 1997 9:53 AM
>To: 'rc5 at llamas.net'
>Subject: RE: [rc5] v2.005
>I've never been able to get NT to spin down a disk because the OS is
>always hitting the pagefile....
>On the mac side people have been setting up ramdisks to avoid this
>problem. I've never seen a NT ramdisk, but that doesn't mean it doesn't
>> -----Original Message-----
>> From: Jonas Svensson [SMTP:jonass at lysator.liu.se]
>> Sent: Monday, October 06, 1997 5:27 PM
>> To: rc5 at llamas.net
>> Subject: Re: [rc5] v2.005
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>> I would like to see one thing:
>> A client that do not touch the disc at all. I can not use spindown
>> because of the rc5 client is always writing to disc. (windows
>> -----BEGIN PGP SIGNATURE-----
>> Version: PGP for Personal Privacy 5.0
>> Charset: noconv
>> -----END PGP SIGNATURE-----
>> jonass at lysator.liu.se
>> a.k.a. jonsv119 at student.liu.se a.k.a 2:204/204.8 at fidonet
>> To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe
>> rc5' in the body.
>To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.
More information about the rc5