SHOOTOUT: best 'mid 2014'
Retina MacBook Pro quad-core
versus best 'late 2013' counterpart
Posted Monday, August 11th, 2014 by rob-ART morgan, mad scientist
Updated Tuesday, August 12th, with "cooler" results*
How does the best 15" Retina MacBook Pro for 2014 compare to the best from 2013? We ran our typical artificial benchmarks to see get a quick answer.
GRAPH LEGEND
mid 2014 rMBP = 'mid 2014' Retina MacBook Pro 2.8GHz Quad Core i7 with GeForce GT 750M GPU, Intel Iris Pro GPU, 16G of RAM, and 1TB flash storage
late 2013 rMBP = 'late 2013' Retina MacBook Pro 2.6GHz Quad Core i7 with GeForce GT 750M GPU, Intel Iris Pro GPU, 16G of RAM, and 512G flash storage
(Both were running Mavericks OS X 10.9.4)
(* We suspected that some confusing results were the result of thermal throttling so we re-ran the tests with both rMBPs sitting on a laptop cooling stand. Some results changed for the better. Some did not.)
Cinebench R15 -- Render 3D Model first with Single Core then with all Cores with Hyper-threading. (RED graph bar indicates fastest CB Rating)
Geekbench 3 -- Here is the breakout of the 64-bit multi-core performance ratings for Integer, Floating Point, and Memory (RED graph bar indicates the fastest SCORE.)
GPUtest is a collection of OpenGL stress tests for single GPUs. FurMark is called the "GPU burner." TessMark measures tessellation processing speed. Julia is a new double precision tester. We unchecked "Automatic graphics switching" to force use of the high performance graphics of the GeForce GT 750M. (RED graph bar means FASTEST in Frames Per Second.)
LuxMark is an OpenCL benchmark. We rendered the Sala scene three different ways. (RED graph bar means FASTEST in Thousands of Samples per Second.)
AJA System Test measures file level, large sequential storage transfer speed using a 16GB test document that simulates capture and playback of uncompressed HD video with 4096x2160 frame size. (RED means fastest in Megabytes per Second.)
We ran 5 cycles of the Standard 1K to 1024K Random test in QuickBench to predict boot volume performance and to simulate an application that executes multiple small random transfers. (RED means fastest in Megabytes per Second.)
INSIGHTS
CPU UNEVEN BUMP
The best CTO 'mid 2014' MacBook Pro quad-core is clocked 7.7% faster than the best 'late 2013' MacBook Pro quad-core CTO, but as you can see from the CPU heavy benchmarks above, that does not translate in all situations. In fact, in the "CPU only" LuxMark OpenCL test, last year's rMBP tested faster.
GPU DISSONANCE
It's not clear how the 'mid 2014' beats the 'late 2013' running the OpenGL tests (FurMark, TessMark, Julia) since the GeForce GT 750M is supposed to be identical in specs. Nor is it clear why in the OpenCL test (LuxMark Sala), the older rMBP beats the newer one rendering with the integrated Intel Iris Pro. (* We suspected that some confusing results were the result of thermal throttling so we re-ran the tests with both rMBPs sitting on a laptop cooling stand. Some results changed for the better but the ranking remained the same.)
STORAGE SURGE
Our storage test wasn't exactly "apples to apples" since the newest rMBP had the 1TB flash blade while last years' rMBP had the 512G flash blade. You would think they would be almost identical in speed, but that was not the case. The 1TB tested much faster. I guess you could say that you *can* have both speed and capacity -- for a price.
We have yet to test with 'real world' apps, but if you already own a 'late 2013' rMBP, you may want to wait for the Broadwell version of the Retina MacBook Pro rumored to be coming next year. If you own a 2012 or earlier MacBook Pro, the newest rMBP should provide a significant boost across the board in performance.
Comments? Suggestions? Feel free to email me,
Follow me on Twitter @barefeats
WHERE TO BUY A MAC PRO or other APPLE products:
USA readers can help us earn a commission by using this Apple Store USA link or by clicking on any Apple display ad. Live outside the USA? No problem. We are affiliated with these Apple Stores:
