Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

ForeverBegins

macrumors newbie
Original poster
Jul 15, 2013
1
0
I work between windows 7 and OS X on identical machines (dual boot) and it always strikes just how annoyingly slow windows is. This made me a bit curious, so I timed a few actions such as booting up applications, both photoshop and various web browsers.

In every case on the exact same 15" rMBP with max specs every application took nearly twice as long to boot in Windows 7 as it did OS X.

For example:

Photoshop 120% longer in W7
Chrome 22 75% longer in W7
Opera 12 115% longer in W7
Firefox 15 100% longer in W7

I couldn't reliably duplicate a good way of switching between applications and timing it accuratly but OS X also seems to do this around twice as quickly.

Can anyone explain why there is such a significant difference?
 

roadbloc

macrumors G3
Aug 24, 2009
8,784
215
UK
Interesting you say that because for me OS X has always been slower than Windows.

So, I have literally no idea. :)
 

throAU

macrumors G3
Feb 13, 2012
8,817
6,985
Perth, Western Australia
It depends how you measure - what software you are testing and how.

OpenGL performance is generally faster on Windows, for example.


On reason could be that OS X is a "cleaner" platform than Windows. Due to the "dll hell", windows apps typically include the specific version of a DLL they have been tested to work with, in the application directory.

This means the application will load its own version of a shared library, rather than referencing the already loaded system version.

Mac apps are better behaved than that.


ALSO:

If you're testing this on teh same hardware and are running bootcamp, then if you are using a spinning hard drive (not an SSD), windows will be installed on the slower part of the disk.

Hard drive performance slows down the further from the start of the disk you get, as the angular velocity of each track is slower towards the middle.

The difference in speed between the outermost track (start of first partition) and the innermost track (end of last partition) is in the other of 50%. I.e. the end of the disk is half the speed of the start of the disk.

In that case, it's not Windows being slow - it's your disk. If you were to install 2 identical copies of OS X you'd also see different performance.
 

Nermal

Moderator
Staff member
Dec 7, 2002
20,632
3,987
New Zealand
Interesting you say that because for me OS X has always been slower than Windows.

My "lowly" 16 gig quad i5 with OS X is faster for most tasks than my 24 gig octo Xeon at work. While I think the work PC is faster in absolute terms, OS X seems to optimise the most common tasks a lot better.
 

Comeagain?

macrumors 68020
Feb 17, 2011
2,190
46
Spokane, WA
If you're testing this on teh same hardware and are running bootcamp, then if you are using a spinning hard drive (not an SSD), windows will be installed on the slower part of the disk.

Hard drive performance slows down the further from the start of the disk you get, as the angular velocity of each track is slower towards the middle.

The difference in speed between the outermost track (start of first partition) and the innermost track (end of last partition) is in the other of 50%. I.e. the end of the disk is half the speed of the start of the disk.

In that case, it's not Windows being slow - it's your disk. If you were to install 2 identical copies of OS X you'd also see different performance.

As the OP specifically mentions using a rMBP, this is a moot point.
 

throAU

macrumors G3
Feb 13, 2012
8,817
6,985
Perth, Western Australia
As the OP specifically mentions using a rMBP, this is a moot point.

Fair point.

Not everyone else is however, and others may be seeing similar for the reasons I mention above.

----------

My "lowly" 16 gig quad i5 with OS X is faster for most tasks than my 24 gig octo Xeon at work. While I think the work PC is faster in absolute terms, OS X seems to optimise the most common tasks a lot better.

What year Xeon?

Xeons have been around since about 1998 or earlier, just because it's a Xeon doesn't mean its necessarily quick. For most users SSD is the biggest imapct anyway. The way a workstation is configured may also not be for outright maximum performance either - things like RAID, etc can be configured more for reliability.

Depending on what you're doing too, Xeons lag the desktop CPUs by a generation or so and may not have the most recent CPU instructions for media decoding, etc.
 

sjinsjca

macrumors 68020
Oct 30, 2008
2,238
555
Both platforms have their strengths. One of OS X's is no need for 'round-the-clock antivirus activity. Scanning every file and process for bad stuff takes an astonishing amount of processing power, and then there's the antivirus update process, which can take many minutes and (in the case of my Kaspersky and Avast installations) noticeably slow down the machine while they're downloading and installing.

Try (carefully and temporarily) turning off your antivirus utilities on Windows to give what some might call a fairer test. On the other hand, constant real-time antivirus is a necessity with Windows, so it's hard to say if it's fair to omit it when users should definitely always enable it.

Windows also suffers from disk fragmentation much more than OS X does, but since you're running on an SSD that might not be as much of an issue. On the other hand, Windows gives software more direct access to hardware than OS X does, which can yield a speed boost in some applications such as hardcore gaming.

Lastly, your Windows installation's system-level device drivers (including for the SSD) may not be as optimized for your Apple hardware as OS X is. On the other hand, some subsystems may be more mature and optimized for Windows, such as the OpenGL mentioned by a previous poster. Since your benchmarks are heavy on disk I/O, I'd cast a skeptical eye at the disk device drivers... I'd wager they're generic vanilla device drivers that aren't set up to squeeze all the possible oomph from your nice SSD.
 

throAU

macrumors G3
Feb 13, 2012
8,817
6,985
Perth, Western Australia
W3550, whatever year that is.

That would make it early 2010, but based on a 2008 design. It is based on Nehalem, which is first gen Core i series in regular desktop CPU land.

I don't think Xeons of that age do turbo boost - it will be locked at 3Ghz and is only quad core so you will have 2 sockets. It will also be limited to SATA2 (or equivalent SCSI/SAS).
 
Last edited:

Dalton63841

macrumors 65816
Nov 27, 2010
1,449
8
SEMO, USA
Windows is designed for the greatest common denominator. The software is written in a very general(read: not taking full advantage of what it's given) way to allow it to be compatible with the widest possible range of hardware.

On the other hand, Mac is optimized for the limited set of hardware it runs on.
 

Nermal

Moderator
Staff member
Dec 7, 2002
20,632
3,987
New Zealand
That would make it early 2010, but based on a 2008 design. It is based on Nehalem, which is first gen Core i series in regular desktop CPU land.

I don't think Xeons of that age do turbo boost - it will be locked at 3Ghz and is only quad core so you will have 2 sockets. It will also be limited to SATA2 (or equivalent SCSI/SAS).

Ah, I haven't really poked around inside this machine since it's a work one and we just take what we're given :p

It does have an SSD though!
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.