The Mac Pro Review (Late 2013)
by Anand Lal Shimpi on December 31, 2013 3:18 PM ESTMac Pro vs. Consumer Macs
For my final set of CPU performance charts I put the new Mac Pro through the same set of tests I do all new Macs. There are definitely multithreaded components to these tests (some are indeed highly threaded), but the suite also values good single threaded performance. Here we'll get an idea of how the new Mac Pro, in its most expensive configuration, fares as a normal Mac.
I've already gone through Cinebench 11.5 results, but the following graphs should put in perspective the Mac Pro's performance relative to all consumer Macs:
If there's one graph that tells the story of why Intel's workstation roadmap is ridiculous, it's this one. The Mac Pro follows Intel's workstation roadmap, which ends up being cut down versions of Intel's server silicon, which happens to be a generation behind what you can get on the desktop. So while the latest iMac and MacBook Pro ship with Intel's latest Haswell cores, the Mac Pro uses what those machines had a year ago: Ivy Bridge. Granted everything else around the CPU cores is beefed up (there's more cache, many more PCIe lanes, etc...), but single threaded performance does suffer as a result.
Now part of this is exaggerated by the fact that I'm reviewing the 2.7GHz 12-core Mac Pro configuration. Single core turbo tops out at 3.5GHz vs. 3.9GHz for the rest of the parts. I suspect if you had one of the 8-core models you'd see peak single threaded performance similar to what the 2012 27-inch iMac delivers. The 2013 27-inch iMac with its fastest CPU should still be quicker though. We're not talking about huge margins of victory here, a matter of a handful of percent, but as a much more expensive machine it's frustrating to not see huge performance leadership in all areas.
The Mac Pro is designed to offer competitive single threaded performance, but really deliver for everyone who depends on great multithreaded performance:
If you need more cores, the Mac Pro is literally the only solution Apple offers that can deliver. We're talking about multiple times the performance offered by anything else in Apple's lineup with a Pro suffix.
I'm slowly but surely amassing Cinebench 15 results. The story doesn't really change here, I just thought I'd publish the numbers in case anyone wants data using this new test:
The latest versions of iPhoto and iMovie break comparisons to my older benchmarks so I've had to drop them here. I still have our Photoshop CS5 and Lightroom 3 tests though:
As I mentioned earlier, threading seems to have improved on newer versions of Photoshop. In CS5 our benchmark looks more like a lightly threaded test by comparison. Out of curiosity I ran the test under Photoshop CS6 and came away with a completion time of around 6 seconds.
Our Lightroom 3 export test tells a very similar story. Anyone with lighter workloads looking for a huge performance increase thanks to the Mac Pro will have to look elsewhere. The Mac Pro is at least performance competitive, but in these lightly threaded workloads you won't see a huge uplift.
267 Comments
View All Comments
stevesup - Wednesday, January 1, 2014 - link
Great review, per usual. Even Leo Laporte couldn't dig out a negative nugget to bash Apple with.milkod2001 - Wednesday, January 1, 2014 - link
In a few months a market will probably be flooded with similar cases. Something like semitransparent case in this shape with decent led lighting could actually look quite nice.Back in Windows Vista times I was working with Mac Pro and iMac as graphic designer. It was a pleasure to work with compared to crappy, slow Vista based Pc.
Now with w7 I can't think about single reason I'd want to spend almost twice for Mac Pro compared to W7 Pc(obviously nobody is forcing me to).
Good job with review Anald.
milkod2001 - Wednesday, January 1, 2014 - link
sorry about misspelled name, can't find edit option for postMat9912 - Wednesday, January 1, 2014 - link
Can someone comment on the power consumption of the new Mac Pro when in standby/sleep mode?knweiss - Thursday, January 2, 2014 - link
You'll find the info in the Mac Pro Environmental Report:http://images.apple.com/environment/reports/docs/M...
nomorespam - Wednesday, January 1, 2014 - link
Any idea why the three networking ports couldn't have been combined into a single PCIe 2.0 lane with a switch/bridge?By my math this comes to at most 3.3Gbps (412MB/s) unidirectional with all three ports saturated.
Is this not a possibility or are there other considerations that make this impractical or undesirable?
I'm thinking it would be really nice to get another 2 (ideally 3 for no bottleneck) of the PCIe 2.0 lanes to the USB 3.0 ports and this seems like a valid way to triple available USB 3.0 bandwidth.
The only other place I can see to steal an additional PCIe 2.0 lane back to get to 4 port bottleneck free USB 3.0 I/O is to take one lane away from the SSD controller?
Surely doing so reduces the bus width and resulting net performance even though the theoretical 1.5GB/s that three PCIe 2.0 lanes provide is still faster than the Mac Pro's shipping SSD's?
Like Neil above, I would also like to know more about IP over thunderbolt 1 and 2 and how it works in the real world today - I would suspect the network stack is not in any way optimized for it at this point.
OreoCookie - Wednesday, January 1, 2014 - link
@AnandAny reason why you haven't posted results on your SSD consistency tests for the Mac Pro?
wozwoz - Wednesday, January 1, 2014 - link
Nice review - though rather too long. Sometimes, less is more :)Unfortunately, what is not answered, and remains 'unknown' is:
* Is a D500 graphics card actually any faster than a D300 in real-world tests? [Note that the D500 has a lower clock speed]
* Since the D700 apparently consumes vastly more power than the D300, how does the graphics card effect noise levels and thermal performance of the entire machine?
I liked your chart of CPU turbo boost, as the number of cores in play changes ... first decent explanation of that issue.
bizarrefish - Wednesday, January 1, 2014 - link
Excellent review Anand, very comprehensive and just what I needed to help seal the decision to buy. I believe some of your benchmarks and stress testing was under Boot Camp Windows correct? I too want to use the system as my main Workstation/PC at home but for occasional gaming also. The system will spend most of its time in OSX but Is the driver support good enough to perform well in Boot Camp for modern gaming needs? Thanks.jackbobevolved - Saturday, April 19, 2014 - link
I got the 12 core D700 model and it works great for gaming. The latest Catalyst drivers installed without issue and performance has been amazing. What really blew me away though was comparing the render and export speed on this machine against my old 3,1 8 core with a Radeon 5870. Several hour exports from FCPX were cut to just minutes.