Does having one drive (or logical partition) dedicated to Windows and applications, and another drive or partition dedicated to data still offer a performance benefit?
It used to be that I'd install Windows on its own partition, have another partition for data, setup my page file to be a static size equal to the amount of RAM I had multiplied by some factor that nobody could agree on (I had one computer with a separate physical hard drive dedicated for the page file), go deep into settings and check some box saying I wanted to optimize things for server use, and countless other rituals that are probably stupid now.
I understand that putting data on a separate logical partition, at least, lets you reformat your OS partition without losing data, but I have backup in place that would survive my entire box exploding, so that's not really an issue. And since things are better than they were with Windows Me & 95, I don't find myself reformatting every 3 months, either.
Anyway, we now have defraggers that claim to put applications on the fast part of the disk, so it's not as though your pictures will be hogging the outer edges of the HD. So, I fail to see why a fresh install of Windows would be any faster than a fresh install with 100 gigs of images on the same partition (assuming you don't fill the disk up to 99% or something unreasonable).
So, is any of this data segregation and other old strategies really necessary anymore for performance purposes? If so, why? Are there any specific reasons other than ritual?
With HDDs so cheap nowdays, why even mess with partitions. The best way to go, is to get one drive for windows and programs, and dedicate at least one other drive for everything else.
I find it necessary, not just for performance, but to make sure everything is backed up just in case of HDD failure or any other problems.
Though, if you like putting everything on one disk, go for it. It's not my possible loss!
Rather than just have a seperate partition now for data, I have seperate hard drives.
A good solution for an unreliable PC I find is to just leave 10gb at the end of your boot drive as a sort of 'rescue space' in case your Windows install becomes corrupted and you need to reinstall without risking your data.
i find it easier to manage when using different drives or partitions
i've got 1 drive, which is currently big enough, so no need for different drives,
that's why I'm using partitions.
different drives could be a bit faster while loading large files, from different disks,
so, it really depends on what you want.
but putting OS+software and data on different locations, drives or partitions, is a must.
I find that putting the OS and the data on two separate partitions makes for faster defragmenting when I want to clean up the system. However, don't forget that, due to NTFS being able to 'mount' a partition on a directory (instead of giving it a drive letter) makes you able to have a single files tree, yet if the primary partition gets trashed your data is still safe.
For example, you can mount a secondary partition on c:\Documents and Settings, and always have your app settings physically apart from the rest of the system (making a system partition corruption spare the data).
Since this is a principle used even in consumer-grade Linux (mount a secondary partition on /home is default setting on most distros I have tested) I would wager that even on hyper stable system, storing user data and system binaries is still quite relevant.