A faster Mac is always a better Mac, and there are many things you can do to get the most performance from your computer. But what really works? Here are some common myths about Macs and what does--and does not--affect performance.
1. More processing cores always means better performance.'
To test this theory, we ran benchmarks on two 2012 Mac Pros, one with 12 processing cores running at 2.4GHz and one with a quad-core processor running at 3.2GHz. With a MathematicaMark score of 5.70, the 12-core Mac Pro's result was twice that of the quad-core Mac Pro. The 12-core Mac Pro also finished the Cinebench CPU test in the half the time of the quad-core Mac Pro.
But despite all of those extra cores, the 12-core Mac Pro posted slower times than the quad-core system in our iTunes encode, Aperture test, and file compression tests.
While some professional applications can benefit greatly from multiple processors, most applications aren't designed to take advantage of more than four cores. For the majority of applications, fewer but faster processors are preferred.
2. Having an external monitor plugged into your MacBook will slow down performance.'
We tested a late 2013 11-inch MacBook Air, with and without a 27-inch Apple Cinema Display attached, and found almost no performance differences between the two configurations in the 14 tests in our Speedmark 9 benchmarking suite.
We thought switching to an older, slower, MacBook Air might show more or a difference, but we were wrong. The only test to show any real difference was our iMovie test, which was less than 4 percent faster on the 2013 MacBook Air without the external monitor and just over 2 percent faster on a 2011 Air without the external monitor. Differences like that are hardly worth mentioning, much less unplugging a monitor over.
3. Lower capacity SSDs are slower than high capacity SSDs.'
When it comes to solid-state drive (SSD) performance, capacity matters. We took a pair of Toshiba Q series Pro drives and two Samsung EVO 840 drives and ran our performance tests on them.
The 512GB Samsung EVO 840 was 39 percent faster than a 256GB EVO 840 in our 10GB large file write test and 26 percent faster than the lower capacity drive in our 10GB files and folders test. Read speeds, however, were unaffected by the capacity. Black Magic and AJA tests both showed the 512GB drive's write speed to be about 32 percent higher than the 256GB model, with read speeds again showing little change.
The difference in write speeds were even more pronounced in our Toshiba tests. Testing two Q series Pro drives--one at 128GB, the other at 512GB--the 512GB Toshiba was 2.5 times as fast as the 128GB SSD in our large file write tests and 2.3 times as fast in our files and folders read test. Read times were within a percent point of each other.
It's also worth noting that the smaller drives were wildly erratic in their write times. Occasionally they would spike to the speeds found in the larger capacity drives, and other times they dropped way below the average. On the other hand, the larger capacity drives were highly consistent in their read and write speeds throughout our testing.
4. Keeping lots of free space on your startup drive will improve your Mac's performance.'
Our tests on a late 2012 27-inch iMac with 2.9GHz quad-core Core i5 processor, 8GB of RAM and a 7200-rpm 1TB hard drive showed some serious performance degradation as the drive filled up. The two tests that showed the biggest change in performance were in our 6GB files and folders copy test and unzipping a 6GB compressed file. Our baseline tests, with the disk about 5 percent full, showed the iMac taking 93 seconds in the copy test and 84 seconds in the unzip test. When we filled the drive to 50 percent of its capacity, the results slowed down by 4.3 percent on the copy test and just under 8 percent on the unzip test. Filling the drive to 80 percent capacity, the baseline results were more than 11 percent faster than the almost-full iMac in the copy test and 17.6 percent faster in the unzip test. Pushing it even further, we ran the tests again at 97 percent of capacity. This time the baseline results were nearly 21 percent faster in the copy test and almost 35 percent faster on the unzip test.
With SSDs, it was a different story. Only at the 97 percent full capacity did we see any difference in our SSD results. The baseline result for SSD in the MacBook Pro was 35 percent faster, but only in the unzip test.
5. Adding RAM always improves performance.'
The lab has done quite a bit of testing on this subject over the years; our most recent coverage was last May with Mountain Lion and older versions of apps. This time out, we took a mid 2012 15-inch MacBook Pro with quad-core 2.3GHz Core i7 processor and a 512GB hard drive and ran it with 4, 8, and 16GB of RAM on loan from Crucial.
The tasks in our Photoshop tests showed the greatest benefit of increased RAM. Using our standard Speedmark 9 action script with a 100MB test file, the 8GB setup was about 14 percent faster than the 4GB configuration. Upping the RAM to 16GB shaved another couple of seconds off of the time and was 15.5 percent faster than the 4GB baseline configuration. We ran a more intensive test, one that uses more hardware acelerated tasks, and found an even greater benefit using increased RAM. In this test, the 4GB configuration took almost exactly 10 minutes to complete, upgrading the RAM to 8GB brought down the time to 7 minutes 18 seconds, and the 16GB configuration finished the test in just under five minutes.
Many other tests, however, were unaffected by the addition of RAM. These tests included Cinebench CPU and Open GL tests, HandBrake, iMovie, Heaven and Valley graphics benchmarks, and PCMark 8's Office application tests. Some tests actually ran slower with more RAM. Our iPhoto import test took 112 seconds with 4GB of RAM, 117 seconds with 8GB of RAM, and 138 seconds with 16GB of RAM. Similarly, our Aperture import and process test showed the 4GB configuration taking just over 121 seconds to complete, 8GB took an extra 10 seconds and the 16GB configuration added another twenty seconds to the time. Copy, zip and unzip tests were also slower with 8GB and even slower with 16GB of RAM installed.
From our tests, 8GB would probably be the sweet spot for most users. It offers a performance boost in applications like Photoshop, but with fewer performance penalties in apps like iPhoto and Aperture.
6.Faster graphics cards only improve gaming performance.'
While faster graphics cards certainly can pump up 3D gaming frame rates, more and more applications are using OpenCL to take advantage of those powerful GPUs. Two such applications are Photoshop and Final Cut Pro X. Much of time, the GPU acceleration makes for a smoother interface, faster previewing and other UI enhancements. Photoshop has a handful of effects, filters, and manipulations that are GPU accelerated.
We took a 2012 Mac Pro--the most recent Mac to offer easy swapping of graphics cards--and ran a Photoshop action script made up of these GPU accelerated tasks on the stock AMD Radeon HD 5770 with 1GB of VRAM and a Sapphire HD 7950 with 3GB of VRAM. The Sapphire finished the test in 239 seconds, 5 percent faster than the stock card.
In the Heaven and Valley graphics benchmarks, the biggest differences showed up in high 2560 by 1600 resolution tests, where the Sapphire was able to push 14.2 frames per second in the Heaven benchmark versus the 5770's unplayable 1.15 frames per second. Valley results at that high resolution were similar, with the Sapphire achieving 18.3 fps versus the Radeon's 1.25fps.