> We have been asked not to keep this disussion up on the list. There's hardly any point to discussing it off-list. > > We keep saying the same things over and over so it is hardly productive. > > But one more time: > > 1. There are certaily pathological situations where swap files work > > better; and the disk allocation you use in you note is a pathalogical > > as I have seen in a long time. For most people your situation doesnot > > exist. My situation is actually typical. The fact there are other partitions between simply highlights the separation between swap and data. If their were just two partitions, much of the Linux data would be far distant from the swap partition. > > 2. Disks are slow so having to access the disk and move the head four times > > for each page retrieval of write has got to be slower than only > > ond disk acess and no head moving in the swap partition. I won't even ask The heads will not move only of they're physically over the right cylinder. The drive I used in the example has probably got two real platters; four heads. The Seagate 8 Gbyte drive I have here has two platters. > > why you didn't put the Linux swap partition next to the Linux partition. The swap partition is actually planned for OS/2 - didn't you see its size? > > Conclusion: Some times (rarely) swap files are what you want. But for > > most people (certainly the people on this list) swap partitions > > are what they want. > > However, I really don't think more dialogue on this will be productive > > since I can perceive that we are already at the place where we each think I alone have actually produced reasoned argument. It can only be overturned with reasoned argument, not by reiterating unsupported assertions. If you want to disprove me and you have the hardware available, benchmark it. I assume from James' tagline he IS in a position to do so. -- Cheers John Summerfield http://os2.ami.com.au/os2/ for OS/2 support. Configuration, networking, combined IBM ftpsites index. -- To unsubscribe: mail -s unsubscribe [EMAIL PROTECTED] < /dev/null
