Here is a look at how my new MacBook Pro stacks up against the PowerMac G5 which it will compliment. Both machines have essentially two CPUs running at the same speed, the same amount of RAM, very similar hard drives, and very similar video cards. It will be interesting to see how each fares in a variety of testing scenarios.
Computers used:
PowerMac G5
|
MacBook Pro
|
Before running the tests, each machine was restarted. In between each test, as little extra activity was done (i.e. grabbing patches if needed). No other applications were run while running each test.
XBench (version: 1.2)I ran the entire quite of XBench tests three times to get an average of the results. The most important/reliable test results are shown in the graph below. Larger numbers are better.
These Xbench results are interesting. The memory test is the only test in which the MacBook Pro beats the G5. Otherwise, the G5 scores are far and away better than the MacBook Pro's. Both machines have very similar specs, so I'm not sure how much confidence I have in these numbers. Let's see what the other benchmarks show before making a conclusion.
GeekBench is a suite of Single Threaded and Multi threaded tests which aim to gauge the processor and memory speed of a computer. In this graph, I averaged the scores for the Integer, Floating Point (scalar), and Memory (non streaming) based tests. Larger number are better.
The results from GeekBench are nearly the exact opposite of those from XBench (except the Memory outcomes). The Integer result is almost even, while the Floating Point and Memory tests ran about 9% faster on the MacBook Pro than the G5. This outcome is a win for the MacBook Pro. Notice: as of Preview 3, the vector tests were not fully supported on the MacBook Pro, so they were omitted from the Floating Point numbers.
CineBench uses various rendering techniques to measure the speed of your processor and video card. I have separated the data for both categories into separate graphs. Larger numbers are better.
The C4D (Cinema4D) test is really separate from the other two. As you can see, the MacBook Pro wins two of the three trials. The G5 wins the hardware OpenGL test by only 3.2%, so it's very close. The biggest thing to note from this test is how optimized the OpenGL software rendering engine is on the Intel based Mac. I wonder if this will translate into better hardware OpenGL numbers in the near future (via driver updates).
On the CPU side of Cinebench, you can see how evenly these computer match up. the Single CPU rendering is almost exactly the same, while the MacBook Pro has about a 5% advantage on the multiple CPU rending. That's good news for this low power, dual core chip.
OpenMark is a purely video test. It has very little to do with with the CPU, so it's a good test to push the graphics processors to their limits. Larger numbers are better.
Similar to the outcome of the hardware OpenGL test under Cinebench, the G5 wins this test by 2.3%. This certainly shows that the ATI Radeon Mobility 1600 is a very good laptop video card.
On the exact opposite spectrum of the last three test suites, AJA KONA evaluates the read and write speeds of your hard drive. I set up this test such that the Read/Write test is run on a 1 GB file. Larger numbers are better.
This test, combined with the Disk test from XBench, concludes that the G5's hard drive is about 18-20% faster than the one in the MacBook Pro. This really surprises me, as both drives are 7200 RPM SATA. Looking up the model number provided by System Profiler confirms that the hard drive is 7200 RPM with an 8 MB cache. I can only speculate on why these numbers have come out this way, and be thankful that I did not go with a 5400 RPM hard drive.
Update: I copied a 915 MB file via firewire 400 to an ATA hard drive also running at 7200 RPM. On the G5, this operation took about 34.5 seconds, while the MacBook Pro needed 35.8 seconds to complete the transfer. Therefore, some hard drive functions will be just as fast on the MacBook Pro as on the G5.
Since I haven't bought any other first person shooters in a few years, Unreal Tournament 2004 is my most taxing game. Therefore, it's the only game result shown in this article. Each of the two test types were run at 1024x768 at Max settings. For the Botmatch, I chose to leave the sound turned on, and used the Inferno map. The Flyby map doesn't have the option to turn sound on, and was run on the default Antalus map. Larger numbers are better.
The results from this test are split. The MacBook Pro won the Botmatch test by a couple frames per second. The Flyby test was a completely different story, as the G5 wiped the floor with the MacBook Pro. The good news here is that you don't actually play the game in flyby, so the Botmatch is a more accurate representation of real world results.
I think it can be said that the MacBook Pro is as fast, or maybe a little bit faster than the G5. The notable exception is the hard drive, where the G5 inexplicably beats the MacBook Pro hands down. Nonetheless, both machines are very powerful, and could suit almost anyone's computing needs (with the exception of extreme gamers).
Updates:
- June 02, 2006 - Updated the GeekBench scores with the Preview 3 version. The Integer test remained about the same, while the FPU and Memory tests now show a 9% advantage on the MacBook Pro instead of a 33-35% advantage.
Mac, the Mac logo, Mac OS, the Apple logo, and other Apple product names are trademarks of Apple, Inc., registered in the U.S. and other countries. The Made on a Mac Badge is a trademark of Apple, Inc., used with permission.
Additional company names, company logos, product names, and product images may be trademarks or registered trademarks and are hereby acknowledged.
Arbitrary Quote - I finally came up with a punishment for the boy. First, no leaving the house, not even for school. Second, no egg nog. In fact, no nog period! And third, absolutely no stealing for three months!