Comparing Apples to Apples

I've been trying to help a fellow who bought an Apple Mac mini Core Solo and finds it slow, particularly with Camino and Audacity. The Core Solo variant never should have been offered, and no longer is, but he got one, so I wanted to figure out if it was actually slower than its predecessor G4 system and how it stacked up to other systems. Here's what I wrote, for those who enjoy benchmarks.

OK, so here's the full comparison, including my PowerMac G5 Dual Processor @ 2.7 GHz, my Dell Intel Pentium 4 @ 2.66 GHz (Windows XP SP 2), my Mac mini PowerPC G4 @ 1.44 GHz and the previous results. The PowerMac has a ATI Radeon X850 XT with 256 megs of ram; the Dell has an ATI Radeon 9700 with 128 megs of ram.

	
Rendering Single CPU
	My Mac mini (PowerPC G4):		138
	Nathan's Mac mini (Core Solo):		206
	My Pentium 4 @ 2.66 GHz:		259
	MacBook Core Duo 2.0 GHz:		303
	My PowerMac G5 @ 2.7 GHz: 		410

Rendering Two CPUs/Cores
	My Mac mini (PowerPC G4):		--
	Nathan's Mac mini (Core Solo):		--
	My Pentium 4 @ 2.66 GHz:		--
	MacBook Core Duo 2.0 GHz:		564	
	My PowerMac G5 @ 2.7 GHz:		744


Shading (CINEMA 4D)
	My Mac mini (PowerPC G4):		152
	Nathan's Mac mini (Core Solo):		206
	My Pentium 4 @ 2.66 GHz:		267
	MacBook Core Duo 2.0 GHz:		348
	My PowerMac G5 @ 2.7 GHz:		405

Shading (OpenGL S/W)
	My Mac mini (PowerPC G4):		463
	Nathan's Mac mini (Core Solo):		888
	My Pentium 4 @ 2.66 GHz:		1097
	MacBook Core Duo 2.0 GHz:		1141
	My PowerMac G5 @ 2.7 GHz:		1279

Shading (OpenGL H/W)
	My Mac mini (PowerPC G4):		475
	Nathan's Mac mini (Core Solo):		940
	MacBook Core Duo 2.0 GHz:		1031
	My Pentium 4 @ 2.66 GHz:		2336
	My PowerMac G5 @ 2.7 GHz:		2902

First note: this isn't terribly scientific. Not all the systems are configured as closely as possible, but it'll provide an idea. My Mac mini has 256 megs of ram; my PowerMac has 1.5 GB; my Dell has 768. My mini is running Mac OS X 10.4.8 Server. The Mini and the Windows system are on a KVM switch hooked to an LCD panel doing 1280×1024, as is the MacBook (well, it is wide screen, so the resolution is a bit different), I believe, whereas your system was driving a higher resolution 1440×900, which ought to slow it down a bit, and my PowerMac is driving a 1920×1200 screen, which should slow it down even farther.

Big point: the Core Solo performs substantially better in every test over the PowerPC G4 Mac mini — even in the hardware OpenGL test — despite the PowerPC Mac mini's discrete video card with 32 megs of video ram.

There's a number of things that could be going on: first, not all apps are optimized for Intel yet. Your best bet is still to use the apps optimized for the job. Safari should best Camino, GarageBand should do better than Audacity (Audacity isn't even a native Aqua app, and as I've said, I'm not sure how well Apple's X11 is optimized — it is OK, but still…).

But, given that the Core Solo outperforms my Mac mini, I'm scratching my head a bit. I'm not sure why yours isn't performing acceptably, because mine is quite usable even though it is quite a bit slower.

The other thing I can offer from this is that at least in terms of rendering power, you can see that the PowerMac is still a monster compared to the other systems. This might indicate that you should consider buying one. They are getting cheaper on eBay and it still has some good life in it. Or, given that the Mac Pros are showing up as even faster, if you spring for the extra expensive system, it should give you a LOT more performance.

Nevertheless, I think the Core architecture really shines here, once you go to a multi-core system. If you look at the MacBook, it performs significantly better than the Mini, especially when you have two cores going. Part of that is purely a clock difference, but the second core helps a lot. The MacBook @ 2.0 GHz bests the faster clock speed Pentium 4 @ 2.66 GHz on every test, even when using just one core, save for when the test is primarily based on video card performance (the last test), where the Radeon 9700 with 128 MB of ram still shines fairly well.

I think that means that you would see a substantial performance increase with a Core Duo Mac mini, but whether it would be enough for you is hard to say. Waiting for a Core 2 Duo Mac mini might be the best thing. It will give an additional speed boost, but still would be cheaper than a used G5 or a Mac Pro.

So there you have it. If you had a sleepless night last night wondering if a Pentium 4 could beat a MacBook, or whether a Mac mini Core Solo would trounce a PowerMac G5, now you know.