My understanding of it is that it doesnt necessarily have to be the fastest drive but a drive that does not contain the winblows partition and is not accessed much by other applications. Of course it helps if that drive happens to be the fastest one.
My understanding of it is that it doesnt necessarily have to be the fastest drive but a drive that does not contain the winblows partition and is not accessed much by other applications. Of course it helps if that drive happens to be the fastest one.
If your pagefile is on the same HD or partition as the OS or the applications, the HD must stop what it is doing and move the read head over to the pagefile, do its thing, and return to what it was doing then repeat. It is not slow but it is not very effecient.
Placing your pagefile on a seperate partion on the same HD has the same effect of forcing the read head to seek and read the data then return and repeat.
Placing the pagefile on another HD attached to the mobo as a slave drive on the primary IDE channel (both HDs attached to the same cable) does not help either because the IDE controller must interrupt the system bus in order to ask for access on the slave side. This essentially stops the I/O from the master drive because it must wait for it's turn on the IDE primary bus.
In order to avoid these slower options, I have placed my pagefile on its own HD using it's own channel. That is the only way to ensure that data can be read simultaneously on both drives. Each channel, such at the Primary or Secondary, has its own bus so it does not compete for I/O access.
The same applies to mobos with more than two IDE connectors. I have 4 connectors, therefore I have 4 seperate paths for data to flow without competition over the IDE bus.
If your pagefile is on the same HD or partition as the OS or the applications, the HD must stop what it is doing and move the read head over to the pagefile, do its thing, and return to what it was doing then repeat. It is not slow but it is not very effecient.
Placing your pagefile on a seperate partion on the same HD has the same effect of forcing the read head to seek and read the data then return and repeat.
Placing the pagefile on another HD attached to the mobo as a slave drive on the primary IDE channel (both HDs attached to the same cable) does not help either because the IDE controller must interrupt the system bus in order to ask for access on the slave side. This essentially stops the I/O from the master drive because it must wait for it's turn on the IDE primary bus.
In order to avoid these slower options, I have placed my pagefile on its own HD using it's own channel. That is the only way to ensure that data can be read simultaneously on both drives. Each channel, such at the Primary or Secondary, has its own bus so it does not compete for I/O access.
The same applies to mobos with more than two IDE connectors. I have 4 connectors, therefore I have 4 seperate paths for data to flow without competition over the IDE bus.
Comment