Apple Mac OS X vs. Windows XP Performance Comparison

01
of 09

Introduction and Comments

Windows XP on an Intel Based Mac Mini
Windows XP on an Intel Based Mac Mini. ©Mark Kyrnin

Introduction

Last year, Apple announced that they intended to switch from using IBM's PowerPC hardware to Intel processors. This brought a lot of hope that individuals who want to run Windows and Mac operating systems on a single platform. At the release, these hopes were quickly dashed by the realization that Microsoft installers would not work.

Eventually a contest was formed to build a prize for the first person to find a reproducible method for installing Windows XP on the Mac. That challenge was completed and the results were posted to the contest providers at OnMac.net. With this now available, it is possible to compare the two operating systems to one another.

Windows XP on Mac

This article is not going to go into detail about how to get the Windows operating system installed onto an Intel based Mac computer. Those looking for that information should visit the "HOW TO" FAQ found on the OnMac.net web site. Having said that, I will make a few comments about the process and some things users should be aware of.

First, the process detailed will only produce a dual boot system. It is not possible to remove the Mac OS X entirely and solely install Windows XP on the computer system. This is still be investigated by the community. Second, the drivers for the hardware are very kludged together from other hardware vendors. Installing them can be tricky. Some items do not even having working drivers yet.

Hardware and Software

02
of 09

Hardware and Software

Hardware

For the purpose of this article, the Intel based Mac Mini was selected to compare the Windows XP and Mac OS X operating systems. The primary reason for the Mac Mini selection was that it has the best overall driver support of the available Intel based systems available. The system was upgraded to the full system specs available from the Apple web site and are as follows:

  • Intel Core Duo T2300 (1.67GHz) Dual Core Processor
  • 2GB PC2-5300 DDR2 Memory (5-5-5-15 Timing)
  • 120GB 5400rpm 8MB SATA Hard Drive
  • 8x DVD+/-RW Burner
  • Intel GMA 950 Integrated Graphics
  • AirPort Extreme 802.11b/g Wireless and Bluetooth

Software

The software is a very important part of this performance comparison. The two operating systems used in the comparison are Windows XP Professional with Service Pack 2 and the Intel based Mac OS X version 10.4.5. They were installed using the methods detailed by the instructions provided by OnMac.net website.

For the purpose of comparing the two operating systems, several basic computing tasks that users normally perform were selected. Next, the task was to find software that would run on both operating systems that were comparable. This was a difficult task as some may be compiled for both platforms, but many are solely written for one or the other. In cases such as these, two applications with similar functions were selected.

Universal Apps and File Systems

03
of 09

Universal Applications and File Systems

Universal Applications

One of the problems with switching from the PowerPC RISC architecture to Intel meant that applications would need to be rewritten. To help speed up the process of transition, Apple developed Rosetta. This is an application that runs inside of the OS X operating system and dynamically translates code from older PowerPC software to run under the Intel hardware. Newer applications that will run natively under the OS are called Universal Applications.

While this system works seamlessly, there is a performance loss when running non-Universal Applications. Apple notes that programs running under Rosetta on the Intel based Macs will be as fast as the older PowerPC systems. They do not however say how much performance is lost when running under Rosetta compare to a Universal program. Since not all applications have been ported over to the new platform yet, some of my tests had to be done with non-Universal programs. I will make notes to when I used such programs in the individual tests.

File Systems

While the tests are using the same hardware, the software applications are very different. One of these differences that can impact the performance of the hard drive is the file systems that each of the operating systems use. Windows XP uses NTFS while Mac OS X uses HPFS+. Each of these file systems handle data in different ways. So, even with similar applications, the data access could cause fluctuations in the performance.

File System Test

04
of 09

File System Test

Win XP and Mac OS X File Copy Test
Win XP and Mac OS X File Copy Test. ©Mark Kyrnin

File System Test

With the idea that each OS uses a different file system, I figured a simple test for the file system performance could help determine how this could impact other tests. The test involves using the native functions of the operating system to select files from a remote drive, copying them to the local drive and timing how long it takes. Since this uses functions native to both operating systems, there is no emulation on the Mac side.

Test Steps

  1. Attach 250GB USB 2.0 hard drive to Mac Mini
  2. Select directory that contains roughly 8,000 files (9.5GB) in various directories
  3. Copy selected directory onto the native hard drive partition
  4. Time start of copy to completion

Results

  • Mac OS X – 16m 3s
  • Windows XP – 12m 21s

The results of this test show that the Windows NTFS file system appears to be faster at the basic function of writing data to the hard drive when compared to the Mac HPFS+ file system. This is likely due to the fact that the NTFS file system does not have as many features as the HPFS+ system. Of course, this was also a test that featured far more data than a user normally will deal with at once.

Still, users should be aware that disk intensive tasks can be slower on the Mac OS X native file system compared to the Windows native file system. The fact that the Mac Mini uses a notebook hard drive also means that performance will be slower than most desktop computer systems.

File Archive Test

05
of 09

File Archiving Test

Win XP and Mac OS X File Archive Test
Win XP and Mac OS X File Archive Test. ©Mark Kyrnin

File Archive Test

In this day and age, users collect a large amount of data on their computers. Audio files, photographs and music can eat up space. Backing up this data is something that a lot of us should do. This is also a good test of the file system as well as the performance of the processor in compacting the data into an archive.

This test was done using the RAR 3.51 archiving program as it exists for both Windows XP and Mac OS X and can be run from a command line avoiding the graphical interface. The RAR application is not a Universal Application and runs under the Rosetta emulation.

Test Steps

  1. Open terminal or command window
  2. Use the RAR command to select and compress 3.5GB of data into a single archive file
  3. Time process until completion

Results

  • Mac OS X – 63m 57s
  • Windows XP – 48m 13s

Based on the results here, the process under the Windows operating system is roughly 25% faster than the same task under Mac OS X. While the rar application does run under Rosetta, the performance drop off from this is likely much smaller than the difference in the file systems. After all, the previous file performance test showed a similar 25% performance difference when just writing data to the drive.

Audio Conversion Test

06
of 09

Audio Conversion Test

Win XP and Mac OS X iTunes Audio Test
Win XP and Mac OS X iTunes Audio Test. ©Mark Kyrnin

Audio Conversion Test

With the popularity of the iPod and digital audio on computers, running a test of an audio application is a logical choice. Of course, Apple produces the iTunes application both for Windows XP and natively for the new Intel Mac OS X as a Universal Application. This makes using this application perfect for this test.

Since importing audio to the computer is limited to the speed of the optical drive, I decided instead to test the speed of the programs by converting a 22min long WAV file that was previously imported from a CD to the AAC file format. This would give a better indication of how the applications perform with the processor and file system.

Test Steps

  1. Under iTunes Preferences, select AAC format for Importing
  2. Select WAV file in iTunes Library
  3. Select "Covert Selection to AAC" from the right click menu
  4. Time process to completion

Results

  • Mac OS X – 1m 29s
  • Windows XP – 1m 26s

Unlike previous tests of the file system, this test shows that both the Windows XP and Mac OS X programs are on even footing. Much of this can be attributed to the fact that Apple wrote the code for the application and compiled it natively to use the Intel hardware similarly regardless of the Windows or Mac OS X operating system.

Graphic Editing Test

07
of 09

Graphic Editing Test

Windows XP and Mac OS X Graphic Edit Test
Windows XP and Mac OS X Graphic Edit Test. ©Mark Kyrnin

Graphic Editing Test

For this test I used the GIMP (GNU Image Manipulation Program) version 2.2.10 that is available for both operating systems. This is not a Universal application for Mac and runs with Rosetta. In addition, I downloaded a popular script called warp-sharp to cleanup photographs. This along with the artistic Old Photo script from the GIMP program were used on a single 5 megapixel digital photograph for comparison.

Test Steps

  1. Open photograph file in GIMP
  2. Select Alchemy | Warp-Sharp from the Script-Fu Menu
  3. Press OK to use the default settings
  4. Time script to completion
  5. Select Décor | Old Photo from the Script-Fu Menu
  6. Press OK to use the default settings
  7. Time script to completion

Results

Warp-Sharp Script

  • Mac OS X – 47s
  • Windows XP – 32s

Old Photo Script

  • Mac OS X – 36s
  • Windows XP - 28s

In this test, we are seeing a 22% and 30% faster performance from the application running in Windows XP over Mac OS X. Since the application does not use the hard disk at all during this process, the performance gap is likely attributed to the fact that the code has to be translated via Rosetta.

Digital Video Editing Test

08
of 09

Digital Video Editing Test

Windows XP and Mac OS X Digital Video Test
Windows XP and Mac OS X Digital Video Test. ©Mark Kyrnin

Digital Video Editing Test

I was unable to find a program that was written for both Windows XP and Mac OS X for this test. As a result, I chose two applications that had very similar functions that could convert a AVI file from a DV camcorder into an autoplay DVD. For Windows, I selected the Nero 7 application while the iDVD 6 program was used for Mac OS X. iDVD is a Universal Application written by Apple and does not use the Rosetta emulation.

Test Steps

iDVD 6 Steps

  1. Open iDVD 6
  2. Open "One Step from Movie File"
  3. Select File
  4. Time until DVD burn is completed

Nero 7 Steps

  1. Open Nero StartSmart
  2. Select DVD Video | Photo and Video | Make Your Own DVD-Video
  3. Add File to Project
  4. Select Next
  5. Select "Do not create a menu"
  6. Select Next
  7. Select Next
  8. Select Burn
  9. Time until DVD burn is completed

Results

  • Mac OS X – 23m 32s
  • Windows XP – 15m 30s

In this case, the conversion of the video from the DV file to the DVD is 34% faster under Nero 7 on Windows XP than iDVD 6 on Mac OS X. Now they admittedly are different programs that use different code so the results were expected to be different. The major difference in performance is likely the result of the file system performance though. Still, with all the steps to do this conversion in Nero compared to iDVD, the Apple process is much easier for the consumer.

Conclusions

09
of 09

Conclusions

Based upon the tests and the results, it appears that the Windows XP operating system is actually a better performer when it comes to running the applications compared to the Mac OS X operating system. This performance gap can be as much as 34% faster in two similar applications. Having said that, there are a number of caveats that I wish to point out.

First and foremost is the fact that many of the applications in this test were running under the Rosetta emulation due to lack of Universal Applications. When a Universal Application such as iTunes is used there is no performance difference. This means that the performance gap will likely be closed between the two operating systems as more applications are ported over to the Universal binaries. Because of this, I hope to revisit this test in about 6 months or so when many of the applications have been converted to see what performance difference exists then.

Second, there is the difference in operating systems and usability. While windows does perform better in many of the test, the amount of text and menus that a user needs to go through to accomplish a task is far easier on Mac OS X compared to the Windows XP interface. This may make the performance difference negligible for those who can't figure out how to use the applications.

Finally, the process of installing Windows XP onto a Mac is not an easy process and not recommended at this point for those that aren't very knowledgeable on computers.