Tags ‘benchmark’

How-To: Benchmark your Mac with these three free downloads


Over the course of writing guides to boosting Mac and hard drive speeds, I’ve discussed the incredible performance improvements Macs can get from simple upgrades — adding RAM, choosing a fast solid state drive (SSD) as an internal or external drive, and even running a simple disk optimizer tool. But there’s a common question that comes up when considering upgrades: how can you tell in advance how big of an improvement you’ll actually see?

The answer: benchmarking tools. Many apps help you measure the speed of various components of your Mac, and with a little help, you can estimate the performance jumps you’ll see after an upgrade. Below, I’ll introduce three of the best free Mac benchmarking tools, and explain how they work…

For Hard Drive Speeds: BlackMagic Disk Speed Test

Measuring the speed of your hard drive is the easiest benchmarking process around, and the best tool I’ve found for that task is the BlackMagic Disk Speed Test by BlackMagic Design. Completely free to download from the Mac App Store, this app has only a single window and very few settings to worry about. If you have only one hard drive, you can just hit the Start button after you’ve quit all of your other apps; otherwise, you can access settings by pressing the gear button between the two speedometer circles, or use the File and Stress menus at the top of the screen. Here, you can choose the right hard drive to test, and the level of stress for the testing (1GB is least, 5GB is most).


BlackMagic designed this app to help video editors determine whether their hard drives could handle various video files, ranging from basic, low-bandwidth NTSC videos to more demanding 1080p videos with higher frame rates and color depths. Unless you’re editing video, those details (summarized in the Will it Work? and How Fast? charts below the speedometers) won’t matter at all to you. You only need to focus on the two big gauges.


The drive’s Read speed is on the right, with the Write speed on the left, respectively giving you a sense of how fast apps and videos will load, and how fast things you create will be written to the drive. Speeds in the 25-30 Megabyte per second (MB/s) range are slow — what you’d expect from an external hard drive connected via USB 2.0. The same drive placed inside an iMac, or connected via USB 3.0, could reach four or five times that speed — around 100-120MB/second.


But pop an SSD like the top-selling Samsung 850 EVO I’ve recommended into the same iMac, and these are the kind of speeds you can see: around 500MB/second, five times faster than a traditional hard drive. This is the sort of speed difference that’s dramatic, instantly noticeable, and likely to really improve your day to day Mac experience. Outstanding SSD performance is the key reason Apple has switched all of its MacBook laptops away from traditional hard drives to SSDs, and is beginning the same process with its desktop machines.


For Overall Computer Performance: Geekbench 3

Although there are a bunch of different “total computer benchmarking” apps out there, the one that’s easiest to recommend is Primate Labs’ Geekbench 3, since it’s partially free, works across multiple platforms (including Macs, iOS devices, PCs, and Android devices), and lets you compare one computer’s results to other computers and other users. Like Blackmagic Disk Speed Test, Geekbench 3 is designed to be simple to use — choose one setting, make sure your other apps are closed, and hit the “Run Benchmarks” button shown above. There’s only one hitch: the free version of Geekbench 3 only runs older (“32-bit”) benchmarks on your Mac; to see the superior performance you’d get from newer (“64-bit”) apps, you’ll need the full $10 version from the Mac App Store.


I could go into a lot of detail regarding Geekbench’s results, and there are a lot of them, sorted into three main categories, each with multiple tests. But the key numbers you need to know are the big two at the top: Single-Core Score and Multi-Core Score. Compared against results from other machines, Single-Core gives you a relative sense of how fast your Mac performs under most situations, when only one processing core is handling all of the Mac’s work. Multi-Core shows you how the Mac does when it’s being pushed to its limits and all of its processing cores are sharing a bigger workload at once.

The scores above show how my 2011 four-core iMac compares to my 2013 two-core MacBook Pro; under most circumstances, they’ll feel identical (3166 is only 3% faster than 3078), but when given big tasks to perform, the four-core iMac will deliver nearly twice the performance of the two-core MacBook Pro. Your numbers will vary a little bit from test to test; running the test multiple times will give you an average.


Geekbench offers a web-based Browser that lets you compare performance between your own machines, as well as the key hardware components found in each computer. You can see above that the differences between my iMac and MacBook Pro aren’t merely in their numbers of processor cores; look closely and you’ll note that each iMac core is faster, and the iMac has more memory. On the other hand, the MacBook Pro’s processor is newer, and its memory is faster.


The critical benefit Geekbench offers is the ability to compare your results against ones submitted by other users. Doing a search of the Geekbench 3 database for, say, “iMac 27″ would let you see how various 27-inch iMacs compare with your current computer. That way, if you’re going to shop for a new Mac, you can get a sense of the Single-Core and Multi-Core performance other users are getting from their machines. Do a little math (divide the new machine’s Single-Core number by your old machine’s Single-Core number) and you’ll get a sense of the performance boost. A 27″ Retina iMac with a score of 3980 would be around 26% faster than my current machine (3166) at most tasks. That’s a big jump by Mac standards, and unlike the 5%-8% benefits typically seen in annual Mac updates, one worth paying for.

Buying an all-new Mac is a big step, though, so you might prefer something simpler and cheaper, like adding extra RAM. Although that can deliver excellent performance improvements at a relatively low cost, Geekbench’s Memory test doesn’t show you the improvement you’d get from more RAM — its benchmark only shows the RAM’s raw speed, which typically can’t be improved over the top-specced RAM Apple ships in its Macs. Whatever RAM you buy as an upgrade will match the existing RAM’s speed, but the performance improvements you see will be real, including much-reduced hard disk accessing and better CPU utilization.


For Video Card Performance: Cinebench R15

Last but not least is Maxon’s Cinebench R15, a free tool that tests two things: graphics card performance using OpenGL, and CPU performance. The CPU test shown above checks how fast your computer’s main processor can render a photorealistic 3-D scene containing 2,000 objects with lights, reflections, shadows, and shaders. This test starts with a black window and fills out the image square by square over the course of several minutes; the higher the “point” total, the faster your CPU is. Like the other benchmarks, speeds in Cinebench R15 can be reduced if other apps are running.


The more distinctive benchmark is the OpenGL test, which uses three complex 3-D cars interacting on dimly-lit city streets to test your graphics card’s ability to handle nearly 1 million polygons at once with various special effects active. It’s a cool demo to watch, and the results will be displayed in frames per second (fps). My iMac hit around 68fps, versus around 23fps on my MacBook Pro. Doing a little research online, I found scores suggesting that the demo Mac Pros in Apple Stores were getting around 77fps last year.


That’s the major hitch with Cinebench: it’s hard to meaningfully compare your numbers against other Macs unless you search Google for “Cinebench R15 score” and the specific Mac you want to compare with. Maxon does include a small sampling of different OpenGL scores within a “Ranking” box, but all that tells you is that your machine (in orange) with X cores (C) and Y threads (T) running at a given GHz speed with a certain graphics card achieved Z frames per second. This isn’t “actionable information” in that you can’t do anything with it — most people won’t even be able to tell which Macs those specs pertain to. And unless you have a Mac Pro, the only current Mac with a replaceable video card, your only option to improve graphics card performance is to buy a new Mac.

My advice: if you’re interested in improving your current Mac’s performance, consider boosting the RAM and/or putting in an SSD. CPU and GPU improvements call for an all-new machine, and Geekbench is the best way to determine whether the performance differences will be meaningful enough to justify the added price.

More Great Ways To Improve Your Mac

To make the most of your Mac (or pretty much any other Apple device), I’ve written quite a few How-To and Best of guides, as well as reviews of worthwhile accessories. Read more of my guides and reviews for 9to5Mac here (and don’t forget to click on Older Posts at the bottom of the page to see everything)!

Filed under: How-To, Mac, Tips and Tricks Tagged: benchmark, Blackmagic Disk Speed Test, Cinebench R15, free, free download, Geekbench 3, Mac

Visit 9to5Mac to find more special coverage of Tips and Tricks, Mac, and Mac.

What do you think? Discuss "How-To: Benchmark your Mac with these three free downloads" with our community.

Comments Off on How-To: Benchmark your Mac with these three free downloads


Jeremy Horwitz

May 14th



Geekbench results from 12-inch MacBook show similar performance to 2011 MacBook Air


Even though the device is not yet available, an early unboxing gave us a hands-on look at Apple’s upcoming 12-inch Retina MacBook this morning, and now Geekbench results from the device have emerged giving us a look at what kind of performance we can expect from it. The Geekbench process tested the performance of the entry-level 12-inch Retina MacBook, which packs an Intel Core M-5Y31 processor clocked at 1.1GHz with Turbo Boost to 2.4GHz.

The 12-inch Retina MacBook was put through its paces twice with Geekbench. The laptop received single-core scores of 1924 and 2044 and multi-core scores of 4038 and 4475. If you compare those scores to previous Apple laptops, they’re in line with the 2011 MacBook Air that featured a 1.8GHz Intel Core i7 processor. Comparatively, the baseline 2015 MacBook Air received single-core and multi-core scores of 2881 and 5757 respectively. It features a 1.6GHz Core i5 processor.

Of course, it’s important to note that the new MacBook should offer major graphic performance improvements over previous MacBooks. The 12-inch Retina model features Intel HD 5300 graphics and significantly faster PCIe-based flash storage.

Apple unveiled the redesigned 12-inch Retina MacBook at an event last month, although 9to5Mac gave you an early look at the device’s features and design back in January. The 12-inch Retina MacBook will be available from Apple online and in-stores on April 10th. The device starts at $1299 for the baseline model with a 1.1GHz processor and 256GB of storage. For $1599 you can get the slightly faster 1.2GHz model with 512GB of storage. Both models feature 8GB of RAM.

Screen Shot 2015-04-01 at 9.27.46 PM

Filed under: Mac Tagged: 12" MacBook, 12-inch MacBook, Apple, benchmark, Geekbench, MacBook, Retina MacBook

Continue reading more about Apple, Mac, and MacBook at 9to5Mac.

What do you think? Discuss "Geekbench results from 12-inch MacBook show similar performance to 2011 MacBook Air" with our community.

Comments Off on Geekbench results from 12-inch MacBook show similar performance to 2011 MacBook Air


Chance Miller

April 2nd



New MacBook Pro’s Retina display reviewed and benchmarked

Retina resolution 1 Retina resolution 2 Retina resolution 3 Retina resolution 4 Retina resolution 5 Screen Shot 2012-06-12 at 2.14.45 PM Screen Shot 2012-06-12 at 2.14.57 PM Screen Shot 2012-06-12 at 2.15.07 PM Screen Shot 2012-06-11 at 8.36.33 PM Screen Shot 2012-06-11 at 4.29.36 PM Screen Shot 2012-06-11 at 8.21.29 PM Diablo Retina

After posting initial benchmark data yesterday for the new Retina MacBook Pro’s SSD and USB 3.0, AnandTech published a longer analysis today about the notebook’s display. The report first took a closer look at the new resolution preferences for Retina MBP users and described the advantages of the different scaling options displayed in the gallery above:

Retina Display MBP owners now get a slider under OS X’s Display Preferences that allow you to specify desktop resolutions other than 1440 x 900. At 1440 x 900 you don’t get any increase in usable desktop resolution compared to a standard 15-inch MacBook Pro, but everything is ridiculously crisp… Even at the non-integer scaled 1680 x 1050 setting, the Retina Display looks a lot better than last year’s high-res panel. It looks like Apple actually renders the screen at twice the selected resolution before scaling it to fit the 2880 x 1800 panel (in other words, at 1920 x 1200 Apple is rendering everything at 3840 x 2400 (!) before scaling… Everything just looks better.

As illustrated in the images above showing benchmark data, the review found greatly improved viewing angles, black levels, and contrast when compared to the previous generation high-res MacBook Pro model. AnandTech then looked at Apple’s claims that the new MacBook Pro display reduces glare by 75 percent from previous generations:

By removing the cover glass Apple reduces the number of reflections and thus glare, however it’s important to point out that this still isn’t a matte display… Compared to my matte MacBook Pro, the Retina Display is obviously more glossy but at the same time remarkably close. I’ll reserve my final judgement until I’ve used the display in more varied conditions however.

The last part of AnandTech’s analysis takes a look at Diablo III running in Retina resolution. Diablo III is game that Apple noted would be updated to support the new MBP during its unveiling. As noted in the report, apps will clearly have add Retina support, but certain games are already recognizing the 2,880-by-1,800 resolution. The screenshot of the game in the gallery above shows Diablo III running at full resolution. AnandTech explained:

Diablo III is actually quite playable at 2880 x 1800, at least in the earlier levels (I haven’t had time to make it far enough in the game to tell how bad it can get). I managed to average 20 fps at 2880 x 1800 in the most stressful scene I have presently unlocked. Obviously things are smoother at lower resolutions. Diablo III exhibited some graphical anomalies at 1920 x 1200, but was fine at other 16:10 resolutions…Not all games will let you do this however. The Unigine Heaven benchmark for example wouldn’t expose any resolutions higher than 1920 x 1200.

Comments Off on New MacBook Pro’s Retina display reviewed and benchmarked


Jordan Kahn

June 12th


Google Tablet Spotted in Benchmarks, Running Android 4.1 ‘Jelly Bean’ [Rumors]

The long-rumoured Android tablet supposedly being made for Google by Asus as we speak has been discovered in some tech benchmark results, revealing the 7″ Nexus tablet to be running on Nvidia's Tegra 3 chipset. More »

Comments Off on Google Tablet Spotted in Benchmarks, Running Android 4.1 ‘Jelly Bean’ [Rumors]


Gary Cutlack - gizmodo uk

May 30th


Ivy Bridge MacBook Pro and iMac Benchmarked? [Apple]

Intel's hot new next-gen chipset looks like it'll be making an appearance in Apple's computer ranges rather soon, with Ivy Bridge processors popping up in some Apple tech benchmarks. More »

Comments Off on Ivy Bridge MacBook Pro and iMac Benchmarked? [Apple]


Gary Cutlack - gizmodo uk

May 14th


Benchmarked: New iPad’s A5X vs iPad 2′s A5 vs Tegra 3

At the launch of Apple’s third-gen iPad, the company’s Marketing Chief Phil Schiller claimed the device’s new A5X processor with quad-core graphics provided up to 4x the graphics performance of NVIDIA’s Tegra 3 chip. Schiller also claimed the new chip provided 2x the graphics performance of the iPad 2′s A5 chip. NVIDIA was skeptical of the benchmark data behind the claims, but early benchmarks seemed to show A5X outperforming a Transformer Prime running Tegra 3 in the majority of tests.

New benchmark data provided by IGN shows the iPad 2′s A5 chip outperforming both the A5X and Tegra 3 with the A5X’s improved graphics going largely towards powering the new iPad’s high-resolution Retina display of 3.1 million pixels. The A5X shows a significant increase in performance over iPad 2 and Tegra 3 devices only when the chip is not forced to power the Retina display in “off-screen” benchmarks.

First IGN ran three tests: GeekBench for raw CPU power, and GLBenchmark 2.1 Egypt, and GLBenchmark 2.1 Pro for graphics. It also ran “off-screen” versions of the GLBenchmark tests to show the performance of A5X without having to power the Retina display.

As for raw CPU power, a Tegra 3-powered ASUS Transformer Prime and Galaxy Tab 10.1 scored higher in GeekBench than both the iPads:

Tegra 3′s quad-core configuration blazes past the dual-core A5X, garnering GeekBench scores of 1540 and 750, respectively. Interestingly, the A5X’s average score fell a few points short of the iPad 2′s standard A5 chip, 753. Both the A5X and the A5 also fell shy of the Samsung Galaxy Tab 10.1′s Tegra 2, which received an overall score 905. The gains made by the Tegra 3 are easily chalked up to its two extra cores, but it is also boasts the highest clocking speed of the group at 1.6GHz, compared to the 1GHz clock of the A5X, A5 and Tegra 2.

However, Apple’s stats showing 4x performance were specifically related to the new chip’s quad-core graphics. When it comes to the graphics tests with GLBenchmark, iPad 2 scored higher than the third-gen iPad in both tests. Meanwhile, both iPads beat out the Android tablets. IGN noted the tests “ran at the native resolution of whatever device it was running on.”

For the Egypt test, the iPad 2 (1024×768) produced 6,709 frames at a framerate of 59 frames-per-second, while the new iPad (2048×1536) ran 5,974 frames at 53 FPS and the Transformer Prime (1280×800) generated 5,955 at a rate of 52 FPS. The Galaxy Tab 10.1 (1280×800), on the other hand, produced only 2,465 frames at a surprisingly low 21 FPS.

The results of the off-screen tests (pictured above) show the A5X producing 15,412 frames at a rate of 138 FPS compared to the iPad 2 that has 10,143 frames at 90 FPS. The A5X significantly outperformed Tegra 3 when not powering the Retina display, but it did not quite provide the 4x performance that Apple’s stats claimed.

Related articles

Comments Off on Benchmarked: New iPad’s A5X vs iPad 2′s A5 vs Tegra 3


Jordan Kahn

March 30th


New iPad’s A5x chip beats out Tegra 3 in most benchmark tests

When Apple launched the new iPad on Friday, it did so with a new dual-core A5x processor and quad-core graphics inside. During the product’s unveiling, Vice President of Worldwide Marketing Phil Schiller talked about the new chip noting that it provides four times the performance of Tegra 3. Nvidia was quick to question the slide displayed by Apple onstage (pictured right), which did not provide any specific benchmark data. We now finally have some solid benchmark tests courtesy of Laptop Mag that provide us new insight.

For the benchmark tests, Laptop Mag used an ASUS Eee Pad Transformer Prime, which is powered by Tegra 3, and put it up against the new iPad in GLBenchmark 2.1, Geekbench, and browsers’ benchmarks with Sunspider and Peacekeeper. In its last test (video above), the publication did a side-by-side subjective gaming performance test to try to spot any noticeable differences between the same title running on both devices. Here is what the publication found:

Starting out with GLBenchmark 2.1, the new iPad wins out by processing 6718 frames at 60 fps in the Egypt Standard 3D animation test. In comparison, the Transformer came in at 5,939 frames at 53 fps. Similar results appeared in the program’s other standard tests. The new iPad came in at 7,530,524 frames at 57 fps during the Geometric test compared to the Transformer’s 3,523,926 at 27 fps. However, the Geekbench test, which Laptop Mag pointed out “measures raw processing power rather than graphics,” is where the Tegra 3 was able to outshine the new A5x:

“The quad-core Tegra 3 blew its competitor way as it achieved an overall score of 1,571 to the A5X’s 692. On the integer (1391 to 614), floating point (2408 to 825) and memory subtests (1076 to 784), the Tegra 3 dominated, but the A5X bested it by a small margin of 324 to 266 on the stream subtest.”

Moving onto browser benchmarks, Android’s stock browser against Safari provided almost the same results in Peacekeeper, but the new iPad was able to surpass the Transformer during the SunSpider JavaScript test by completing the test in 1810 milliseconds compared to the transformer’s 2216 milliseconds.

Lastly, Laptop Mag did the side-by-side graphics test with Shadowgun and Riptide running on both devices. However, as you will see in the video above, many of the advantages attributed to the iPad are thanks to the device’s new 2048-by-1536-resolution Retina display and not necessarily the new quad-core graphics.

Cross-posted on 9to5Google.com

Comments Off on New iPad’s A5x chip beats out Tegra 3 in most benchmark tests


Jordan Kahn

March 19th


The New iPad’s 1GB of RAM and 1GHz Processor Confirmed [Ipad]

Though it was almost certainly going to be the case, a leaked benchmark has confirmed what we expected: the new iPad comes with 1GB of RAM, but the clock speed of its CPU stays at 1GHz. More »

Comments Off on The New iPad’s 1GB of RAM and 1GHz Processor Confirmed [Ipad]


Jamie Condliffe

March 13th


iPhones beat out Nokia Lumia 800 Windows Phone in browser benchmark, despite earlier reports

359gsm (via My Nokia Blog) recently pitted the iOS 4.3-based iPhone 4 and the iOS 5-based iPhone 4S against a Windows Phone 7.5 Mango-based Nokia Lumia 800 smartphone in various benchmark tests. In every test, both the iPhone 4 and iPhone 4S significantly outperform one of the best Windows Phone options on the market. This is in contrast to reports in April, prior to Mango’s release, that claimed Window Phone’s IE 9 on Mango devices beat mobile Safari in similar browser benchmark tests.

You can view the results in the video above or get the full results below (via 359gsm). As you can see, the iPhone 4S significantly passes the Lumia 800 in all tests, including: Browsermark tests, Speed Reading test, Sunspider, Acid3, and HTML5. The iPhone 4 also outperforms the Lumia in most tests, although by a narrower margin.

Browsermark Test: Higher is better

iPhone 4 (iOS 4.3) – 37 503

Nokia Lumia 800 (WP7.5 aka Mango) – 30 452

iPhone 4S (iOS 5) – 86 702

Speed Reading Test:

iPhone 4 (iOS 4.3) – 2 fps (iPhone 4 with iOS 5.0 – around 37 fps)

Nokia Lumia 800 (WP7.5 aka Mango) – 40 fps

iPhone 4S (iOS 5) – 60 fps

Sunspider Test: Lower is better

iPhone 4 (iOS 4.3) – 4018.2 ms

Nokia Lumia 800 (WP7.5 aka Mango) – 7188.7 ms

iPhone 4S (iOS 5) – 2266 ms

Acid3 Test:

iPhone 4 (iOS 4.3) – 100/100

Nokia Lumia 800 (WP7.5 aka Mango) – 100/100

iPhone 4S (iOS 5) – 100/100

HTML5 Test:

iPhone 4 (iOS 4.3) – 210

Nokia Lumia 800 (WP7.5 aka Mango) – 141

iPhone 4S (iOS 5) – 296

Comments Off on iPhones beat out Nokia Lumia 800 Windows Phone in browser benchmark, despite earlier reports


Jordan Kahn

January 2nd


Report: Intel releasing Ivy Bridge CPUs April 8, will likely land in future Macs

According to Taiwanese PC manufacturers (via DigiTimes), Intel is preparing to release the first round of 22nm Ivy Bridge desktop, notebook, and ultrabook CPUs around April, 8 2012.  Seventeen models are slated to be launched, but some desktop and notebook CPUs will likely replace current CPU options in next-generation iMacs and Macbooks.

The 3820QM and 3720QM i7 CPUs, detailed by DigiTimes, are candidates for a next-gen MacBook Pro. Meanwhile, quad-core Core i7 and Core i5 desktop models could land in a future iMac. Ultrabook CPUs, possibly headed to future MacBook-air-like devices are scheduled for later in the year (most likely May).

A few weeks back, a leaked roadmap for Intel’s upcoming Ivy Bridge mobile CPU lineup gave us a look into the new standard voltage M-series and ultra low voltage U-series lineups. The document shows the CPUs will probably make their way into MacBook Pros and MacBook Airs, and will most likely be available to OEMs sometime in April or May 2012. The document leaked days after a report claimed Ivy Bridge desktop processors would roll out to partners sometime in the second quarter of 2012.

The new Ivy Bridge lineup marks a significant performance increase over previous Sandy Bridge models found in current Macs. According to 3DMark Vantage GPU benchmark from Intel, there is an average 199 percent improvement in graphic intensive applications, and 25 percent overall better performance over Sandy Bridge CPUs. Improved power consumption could lead to increased battery life, while the lineup will also include OpenCl 1.1 support and DirectX 11 for more than 30 percent faster graphics performance. They could also support 4K video.

Comments Off on Report: Intel releasing Ivy Bridge CPUs April 8, will likely land in future Macs


Jordan Kahn

December 28th

October 2015
« Sep