↝ OWC / MacSales.com... ↜
↝ diglloyd Deal Finder... ↜
Buy other stuff at Amazon.com...
Photoshop CS6 Performance — diglloydSpeed1
Related: CPU cores, iMac, Mac Pro, MacBook, MacBook Pro, Macs, Photoshop
The diglloydSpeed1 benchmark is an in-memory test that easily fits in 8GB, so it’s a pure speed test which does not involve disk I/O. Hence the results depend only on raw CPU processing speed.
vs Mac Pro 6 or 12 core at 3.33 GHz
Apple Mac Pro 12-core 3.33 GHz is my customized system.
The late 2012 iMac 3.4 GHz stomps the fastest Mac Pro systems available.
The results are not unexpected; Photoshop CS6 does not use even two CPU cores well so the latest generation Intel Core i7 running at 3.4 GHz easily outpaces a 3-year-old Xeon running at a slower clock speed.
Some of the speed difference might be Photoshop itself: the slow 12-core performance is not so much a criticism of the 12-core Mac Pro as of the performance bugs with Photoshop CS6, which suffers from internal implementation inefficiencies (performance bugs) on systems with more than 6 CPU cores.
When Photoshop runs with only 4 cores (8 virtual cores) instead of 6 (12) or 12 (24) cores, it is simply more efficient (less overhead, fewer bottlenecks).
At much lower cost than Apple, with more options.
Lloyd recommends 64GB for iMac or Mac Pro for photography/videography.
vs MacBook Pro Retina
The iMac runs at 3.4 GHz with turbo boost to 3.9 GHz, and the MacBook Pro with Retina Display runs at 2.7 GHz with turbo boost to 3.7 GHz.
Thus, the base clock speed on the iMac is 25% faster than the MacBook Pro, but its turbo boost speed is only 5% faster.
The actual real-world 15% difference is positioned between the 5% turbo / 25% base clock speed values; this is probably due to the number of CPU cores used at various parts of the test (the amount of turbo boost is limited by the number of CPU cores in use).
*OpenGL (GPU) was disabled in Photoshop CS6 13.0.3, because it slows things down by a few percent (a result seen repeatedly with many Macs). OpenGL can be helpful for more obscure Photoshop operations, but these are not tested here.