New Mac, Aperture flies but...

[CrackedButter]

...I wanted to do a comparison test between the two machines. I found however that the newer machine was actually slower during processing but not importing.

These were the times to import 192 raw files (3.3GB) -

To finish copying and importing
Intel i5 took 2:49 minutes
Core Duo 2 took 3:39 minutes

To finish copying, importing and processing
Core Duo 2 7:00 mins
Intel i5 - 7:25

So where the i5 was 50 second faster on the first task, it finishes 25 seconds slower overall, how is this possible? I've probably done something wrong.

The specs were as followed:
13" MacBook Pro with i5, 2.3Ghz dual core, 4GB & 500GB 7200rpm drive (HD Graphics 3000). VS
13" MacBook with Core Duo 2, 2Ghz dual core, 3GB & 320rpm drive (GMA 950).

So in every respect the i5 should win easily, but it didn't. The only difference was that the MacBook Pro already had another 3000 raw files in the database. But I can't see how that would affect performance to that degree. OR can it?

I'm going to fill up the old machine with the same library and return, but if anybody has another idea, I'm interested.

 

[Unpleasant]

GPU Acceleration.

http://en.wikipedia.org/wiki/Core_Image#Supported_Graphics_Processors

Aperture uses Core Image. Core Image supports GPU Acceleration.

 

[OreoCookie]

GPU Acceleration.

Not in this case: both machines don't have discrete gpus, and yes, the new machine should be faster.

 

[Unpleasant]

Not in this case: both machines don't have discrete gpus, and yes, the new machine should be faster.

I'm not sure how accurate Wiki is but his GMA950 is on the list. So it will Accelerate it a little while the HD3000 isn't.

 

[xStep]

Not in this case: both machines don't have discrete gpus, and yes, the new machine should be faster.

I don't know that the GPU is the cause of the difference here, but...

I don't think the fact a GPU isn't discrete means that CoreImage can't take advantage of the processing power. As far as I know, the difference between discrete and non-discrete GPUs is the access to private memory compared to access to the main system memory. Sure, having private memory can give a huge performance boost, but two different GPUs could have the same instruction set but different memory access capability.

 

[CrackedButter]

I haven't had chance to get back on the old machine, sorry for those curious. Will try soon.