32-bit Win XP BootCamp - should the disk be listed as a SCSI device?
Was the previous performance issue virtual IDE did not have buffer block access and SCSI did---does this still matter?
Raw disk access and virtual disks are completely different. Raw disks will always be the same disk type as they are on the system, which in your cases is IDE (or SATA actually). There are no changes you can make to Boot Camp disks for buffered IO, raw disk has different performance metrics and requirements. There are not changes you can make in beta 4 to change raw disk performance.
Pat
Excellent info. I get it now. Thanks.
I don't have facts or figures, but my instinct in working with some disk intensive apps over the past day is BootCamp access is slow. Very slow. System is WinXP (2-cpu, approx. 640MB memory) running on MacPro Quad 2GB.
My assumptions:
Single file VMs are faster than raw disks and bootcamp access.
Single file VMs can be setup using the SCSI driver which is faster than IDE.
Questions:
Do I understand the performance hierarchy correctly? (based on above assumptions)
Will bootcamp access become faster in the future or is there a limitation of the pathway that prohibits a feature found in other disks (like multithreading in a portion of the driver, buffering somewhere, or some other blocking issue)?
Would it help to benchmark my findings or are you aware of the discrepency and difference?
Is this slowdown the price to be paid for Bootcamp vs. single file VM? 😐
Thanks in advance.
I'd like to know this as well. The whole "Bootcamp vs file based performance" has always been a bit of a question for me. I can't seem to find a solid answer.
My assumptions:
Single file VMs are faster than raw disks and bootcamp access.
Single file VMs can be setup using the SCSI driver which is faster than IDE.
Both of your assumptions are actually correct...for further information on the first one take a look at the following topic[/url].
Will bootcamp access become faster in the future or
is there a limitation of the pathway that prohibits a
feature found in other disks (like multithreading in
a portion of the driver, buffering somewhere, or some
other blocking issue)?
You should find the answer in the above referenced topic...actually you will yourself in that topic as well...
Is this slowdown the price to be paid for Bootcamp
vs. single file VM? 😐
Pretty much....although I would consider the slowdown pretty neglible...
Ciao, Andreas
Yes, Andreas, it appears I was asking the same question after answering before. I didn't hit my head though! After reading about the scsi optimization and with the evolved support for Boot Camp I wanted input on the new beta.
I have not measured the disks, but here are my observations:
1) Boot Camp disks are "technically slow", but they are reasonably fast for nearly any type of desktop computing. For any normal usage, most people wouldn't notice at all.
2) Small delays are compounded across multiple files resulting in very noticeable time differences though when doing any serious high-perf computing, i.e. compiling large apps, working with lots of media files, crunching data matrices with inputs from many files. The scsi single VM is an adequate solution though in this scenario. It's fast!
I have not measured the disks, but here are my
observations:
1) Boot Camp disks are "technically slow", but they
are reasonably fast for nearly any type of desktop
computing. For any normal usage, most people
wouldn't notice at all.
2) Small delays are compounded across multiple files
resulting in very noticeable time differences though
when doing any serious high-perf computing, i.e.
compiling large apps, working with lots of media
files, crunching data matrices with inputs from many
files. The scsi single VM is an adequate solution
though in this scenario. It's fast!
I would second that...it is funny though...I am trying to decide whether I need a Boot Camp partition or not...since 4 days or so...still no decision... :smileysilly:
I guess, I am \*really* getting old...
Ciao, Andreas