[TerryJ]

Hey everybody,

I have had a MacBook Pro for a couple months now and I enjoy it very much. My wife needed a new 'Book to replace her aging TiBook so I bought her a MacBook and on a whim, just made it two... thinking I would sell my Pro if I liked my own MacBook better as the size for travel was appealing. Well, now here I am with a MacBook and MacBook Pro and having to decide which one of these guys gets put on the market.

Anyway, I thought a good tie-breaker for me would be some performance numbers. With the integrated graphics of the MacBook I figured the Pro would trounce it. Surprisingly, the Pro lagged in pretty much every single test... with a HUGE difference in Open GL which is where I expected the opposite. To add insult to injury, the Pro is running 2GB or memory whereas the MacBook is running the stock 512MB. I ran XBench several times and the numbers were fairly consistent each time on both machines. Needless to say, I am quite surprised at this and now more at odds over which machine I like best! Just a brief summary is below. Wondering if anyone else has been in a similar position and what your opinions of these benchmarks mean in real-world use. Can a mere 170MHz make this much difference? Even with the rather large memory difference? I would think that the dedicated 128MB memory in the Pro plus 1.5GB more memory would offset the processor speed easily.

MacBook 2.0/512MB/60

MacBook Pro 1.83/2GB/80

Both running OS X.4.6, all software updates applied.

MacBook Results

Results - 49.63
CPU Test - 72.59
Thread Test - 190.16
Memory Test - 116.84
Quartz Graphics Test - 54.46
OpenGL Graphics Test - 215.06 / 272.82 frames/sec !!!!!!!
User Interface Test - 15.48
Disk test - Sequential - 53.14 / Random 30.44


MacBook Pro Results

Results - 47.33
CPU Test - 59.45
Thread Test - 180.12
Memory Test - 111.63
Quartz Graphics Test - 50.78
OpenGL Graphics Test - 123.11 / 156.17 frames/sec
User Interface Test - 21.95
Disk test - Sequential - 20.48 / Random 26.63

[mduell]

The reason is simple: Xbench is poo.

Aside from graphics, the real-world performance of the two should be nearly identical; the MacBook Pro will be faster for 3D games and such, regardless of what some useless benchmark says.

[TerryJ]

Thanks Mark,

Yeah, I had heard that XBench isn't all that great but still, I felt that consistency counted for something. Even if the numbers were flawed, there was still a distinct and stable difference between the two.

[Tomchu]

Make sure that there's nothing running in the background on the MBP that is eating CPU cycles, and that it is plugged in to AC.

Here are the results from my MBP 2 GHz/2 GB of RAM:

Results: 64.48
CPU Test - 70.96
Thread Test - 194.97
Memory Test - 115.84
Quartz Graphics Test - 56.84
OpenGL Graphics Test - 140.73
User Interface Test - 37.37
Disk Test - 43.06
Random - 28.35

Sooo ... it seems that we can conclude that Xbench is useless. MB beating an MBP in OpenGL? My MBP beating your MB and MBP in UI tests, even though the specs are nearly identical? It all seems a bit odd.

In any case, your CPU test score is still somewhat concerning. Being in the 70-73 range seems to be normal for 2 GHz.

[harrisjamieh]

Cinebench 4D is much better than crappy XBench

[TerryJ]

Hey Tomchu,

Sorry, I forgot to mention that everything was identical on both machines and I made sure there were no unusual processes running, no other apps open, etc. Both were done on a fresh restart with only XBench running, both plugged into a/c power and processor set to Best Performance.

Weird, huh? I think I'll give Cinebench a try and see what happens. I'll report back with the scoop.

[TerryJ]

Ok, upon someone's suggestion, I ran a Cinebench comparison and interestingly, the numbers still come out with the MacBook besting the Pro in some areas. The CPU is the most surprising.



MacBook Results
Cinebench v.9.5

CPU Benchmark

Rendering (1 CPU) - 305 CB-CPU
Rendering (x CPU) - 573 CB-CPU
Multiprocessor speedup: 1.88x


Graphics Benchmark

C4D Shading 352 CB-GFX
OpenGL SW-L - 1178 CB-GFX
OpenGL HW-L - 1066 CB-GFX
OpenGL speedup 3.35x


MacBook Pro Results
Cinebench v.9.5

CPU Benchmark

Rendering (1 CPU) - 281 CB-CPU
Rendering (x CPU) - 531 CB-CPU
Multiprocessor speedup: 1.89x


Graphics Benchmark

C4D Shading 328 CB-GFX
OpenGL SW-L - 1263 CB-GFX
OpenGL HW-L - 2485 CB-GFX
OpenGL speedup 7.58x

[harrisjamieh]

Why is the CPU most surprising? Whats to be surprised about - the MB has a faster CPU than the MacBook Pro (the ones you tested). How is it a surprise that the faster CPU got a higher CPU score?

[TerryJ]

I meant GPU, sorry to confuse you on that one harrisjamieh. As you can see, aside from HW-L, the MacBook did pretty darn well against the MBP with it's dedicated card and considerably higher memory and specs.

[Tomchu]

170 MHz faster ... not that much of a percentage increase, yet the score is much higher percentage-wise than the percent increase of the CPU speed.

[snoopy199]

I think testing like configurations would have proved to be more consistant and credible. Regardless of the numbers the Macbook had only 512meg where as the MBP had a 2 gig.

If your going to compare the performance of two machines make sure they are configured similarly.

I myself noticed a marked improvement in performance going from 512meg to a gig.

[Tomchu]

Upgrading the RAM isn't going to improve raw CPU performance.

[snoopy199]

Originally Posted by Tomchu

Upgrading the RAM isn't going to improve raw CPU performance.

No but when your benchmarking the whole computer its wise to compare apples to apples.

[greenamp]

Originally Posted by Tomchu

Upgrading the RAM isn't going to improve raw CPU performance.

Actually, yes it will. In terms of being benchmarked, that is.

[Tomchu]

Originally Posted by greenamp

Actually, yes it will. In terms of being benchmarked, that is.

Actually, no it won't. If your CPU is capable of 2000 MIPS, adding more RAM isn't going to increase or decrease that amount. It's pure and simple computer science.

I can prove this with any number of utilities on any number of platforms, if you don't believe it.

[harrisjamieh]

Originally Posted by TerryJ

I meant GPU, sorry to confuse you on that one harrisjamieh. As you can see, aside from HW-L, the MacBook did pretty darn well against the MBP with it's dedicated card and considerably higher memory and specs.

Thats because the HW-L score is really the only score that tests the graphics. SW-L score is mainly done via software and the CPU. In nearly all past macs (all of which had dedicated video cards), HW-L score is HIGHER than SW-L, not lower like in the MB, showing how weak the 950 actually is

[greenamp]

Originally Posted by Tomchu

Actually, no it won't. If your CPU is capable of 2000 MIPS, adding more RAM isn't going to increase or decrease that amount. It's pure and simple computer science.

I can prove this with any number of utilities on any number of platforms, if you don't believe it.

I'm talking about CPU performance in a computer environment, where the amount of data a CPU can process is 1st dependent on how fast and numerous the RAM can send it.

[rhashem]

Originally Posted by greenamp

I'm talking about CPU performance in a computer environment, where the amount of data a CPU can process is 1st dependent on how fast and numerous the RAM can send it.

The RAM is a couple of orders of magnitude slower than the CPU. If a "CPU benchmark" is streaming data from the RAM, instead of mostly hitting the cache, then its not a CPU benchmark --- its a memory bandwidth benchmark, and won't pick up the difference between a 1.0 GHz and 2.0 GHz CPU, much less a 1.83 GHz and a 2.0 GHz CPU.

[Tomchu]

Thanks, rhashem. Sometimes it's difficult getting through to people who have pre-determined ideas about what goes on inside a computer.

[amazing]

barefeats.com has some interesting iMovie and iDVD results where the MB does indeed perform slightly better than the MBP

http://www.barefeats.com/mbcd4.html

Of course, if you're a gamer the MBP is dramatically better than the MB

http://www.barefeats.com/mbcd3.html