Yesterday Apple announced two new Macs at WWDC: an all-new Mac Pro, and an updated MacBook Air. While the all-new Mac Pro isn't shipping until later this year, the new MacBook Air is available today and is already in peoples' hands. Results for the new MacBook Airs have started to appear in the Geekbench Browser.
Since Geekbench results are already available, we can find out how fast this year's MacBook Airs are compared to last year's MacBook Airs. I've gathered results from the Geekbench Browser for all five models and charted them below:
There's only a modest increase in performance between the 2013 MacBook Air and the 2012 MacBook Air. Depending on which models you compare, you can expect between a 3% and a 8% improvement in performance. However, last year's update brought between a 20% and a 25% improvement in performance. What's going on here?
The new MacBook Airs use Intel's new Haswell processors. With Haswell, Intel's focus was on reducing power usage. Rather than improve performance by 25%, Intel improved performance by 10% while consuming 25% less power.
Apple then went a step further. Apple reduced power usage even more by selecting processors with lower frequencies. In the low-end Air, the frequency has been reduced to 1.3 GHz from 1.7 GHz, while in the high-end Air, the frequency has been reduced to 1.7 GHz from 2.0 GHz.
Focusing on power usage instead of just focusing on processor performance leads to a massive increase in battery life for the new MacBook Airs. Apple claims that the new 11-inch MacBook Air has an 80% increase in battery life, while the new 13-inch MacBook Air has a 70% increase in battery life.
While it would've been nice to see a new MacBook Air that was 20% faster, when you consider the market the MacBook Air serves, I think almost doubling battery life is even nicer.
Last month I posted Geekbench 2 scores for the Samsung Galaxy S 4 with the Qualcomm 600 processor. Scores for the Samsung Galaxy S 4 with the Exynos 5 Octa processor were unavailable at the time:
We have no idea how the Samsung Galaxy S 4 with the new Exynos 5 Octa SoC will perform, but given how the different Samsung Galaxy S 3s performed I imagine the Geekbench score will be similar.
Now that Geekbench 2 scores for the Samsung Galaxy S 4 with the Exynos 5 Octa processor have started to appear in the Geekbench Browser, it turns out "similar" means "15% faster":
Some new points to consider:
While the Samsung Exynos 5 Octa is marketed as an eight-core processor, it's easier to think of it as two four-core processors, where one is a low-power processor and the other is a high-performance processor. Only one processor is active at a time, and the operating system can dynamically switch between the two depending on its power and performance needs.
Note the performance difference between the two Samsung Galaxy S 4 handsets: even though the Exynos 5 Octa runs at a lower frequency than the Qualcomm Snapdragon 600, the Exynos 5 Octa is the faster of the two processors.
The performance difference between the two BlackBerry handsets is also surprising, considering both use the same Qualcomm processor. However, the Q10 is running the latest BlackBerry 10 OS (10.1) while the Z10 is running the original BlackBerry 10 OS (10.0). This suggests BlackBerry Z10 owners will see a boost in performance once 10.1 is released.
Again, I'm amazed at how quickly smartphone performance has improved over the past few years. I'm looking forward to seeing what comes out over the next year and finding out if this pace can be maintained.
Come and influence the direction of one of the world's most popular benchmarks. Primate Labs is looking for a skilled software developer to come in and contribute to Geekbench. As a developer working on Geekbench, you will help design and develop future versions, as well as maintain the current version of the Geekbench application.
- Develop and characterize workloads for future versions of Geekbench.
- Develop and maintain the Geekbench application for desktop and mobile platforms.
- BA/BMath/BSc in a technical field (or in lieu of a degree, equivalent experience).
- Experience with C++ and the STL.
- Experience with at least one of Objective-C, Java, Python, or Ruby.
- Development experience on either Linux or OS X.
- Self-motivated, independent, and proactive.
- Low-level understanding of processor architectures.
- Experience writing cross-platform code.
- Experience writing desktop applications for OS X or Windows.
- Experience writing mobile applications for Android or iOS.
- Experience with GPGPU development using CUDA, OpenCL, or OpenGL.
- Competitive salary and vacation
- Flexible work hours
- Your choice of computer hardware
This is a full-time employment position, and telecommuting is required. Interested? Send me an email with your resume and any other relevant information about your background or work experience and we'll talk!
Android Authority ran Geekbench 2 on the new Samsung Galaxy S 4 last week and posted the results on YouTube. I thought it would be interesting to see how the Samsung Galaxy S 4 compares to other popular smartphones.
I've charted the Geekbench 2 scores for the Samsung Galaxy S4 along with several Android phones, the iPhone 5, and the BlackBerry Z10. If you're not familiar with Geekbench 2, higher scores are better.
Some things to consider when looking at the results:
It's clear that the Samsung Galaxy S 4 will be the fastest smartphone available when it is released in April. What's not clear is how the Galaxy S 4 will handle the increased power and cooling requirements that generally come with faster processors.
The Samsung Galaxy S 4 is twice as fast as the Samsung Galaxy S 3. Given that the Samsung Galaxy S 3 is less than a year old, that's a remarkable achievement. I am amazed at how quickly smartphone technology is improving.
The Samsung Galaxy S 4 is also twice as fast as the Apple iPhone 5. Apple has improved performance dramatically in the past (there was 2.5x increase in performance from the iPhone 4S to the iPhone 5). Will they be able to make a similar improvement for the next iPhone?
The Samsung Galaxy S 4 is also twice as fast as the BlackBerry Z10. However, unlike the Galaxy S 3 and the iPhone 5, the Z10 launched this year, not last year. Will BlackBerry feel compelled to release an updated handset with a faster processor sooner rather than later?
Android Authority only tested the Samsung Galaxy S 4 with the Qualcomm Snapdragon 600 SoC. We have no idea how the Samsung Galaxy S 4 with the new Exynos 5 Octa SoC will perform, but given how the different Samsung Galaxy S 3s performed I imagine the Geekbench score will be similar.
Both the HTC One and the Samsung Galaxy S 4 use the Qualcomm Snapdragon 600 SoC, although the HTC One runs at a slower speed (1.7 GHz vs. 1.9 GHz). The HTC One is also running an older version of Android (4.1.2 vs. 4.2.2), so the HTC One's performance should improve slightly once it is on the latest version of Android.
Again, I'm amazed at how quickly smartphone performance has improved. We've gone from the original iPhone with a Geekbench score of 130 to the Samsung Galaxy S 4 with a Geekbench score of 3100 in only six years.
I have added several devices (both new and old) to the Geekbench Browser Android Benchmark Chart including:
- ASUS Transformer Pad Infinity
- ASUS Transformer Pad TF300T
- ASUS PadFone 2
- HTC Desire S
- HTC Droid DNA
- HTC One
- HTC One X+
- LG Optimus G
- Motorola RAZR i
- Sony Xperia T
- Sony Xperia Z
- Xiaomi MI-2
Right now the HTC One is the fastest Android device on the chart with an average score of 2675. I doubt the HTC One will keep the lead position for long though as the upcoming Samsung Galaxy S 4 uses the same processor clocked at a higher speed (1.9 GHz vs 1.7 GHz).
Last week Apple quietly updated their Retina MacBook Pro lineup with faster processors. The 15-inch Retina MacBook Pro now comes with a 2.4 GHz, 2.7 GHz, or a 2.8 GHz quad-core processor. The 13-inch Retina MacBook Pro now comes with a 2.5 GHz, 2.6 GHz, 2.9 GHz, or 3.0 GHz dual-core processor.
How do these new Retina MacBook Pro models perform compared to the original models? To find out I've collected Geekbench 2 results from the Geekbench Browser for all of the Retina MacBook Pro models and charted the scores below.
If you're not familiar with Geekbench, it's our cross-platform benchmark utility. Higher Geekbench scores are better, where twice the score means twice the performance. If you're curious how your computer compares, you can download Geekbench to find out.
15-inch Retina MacBook Pro
Below is a summary of the processor upgrades for the different 15-inch Retina MacBook Pro models:
Each model received a 100 MHz increase in processor speed. Besides the differences in speed, the processors for each model have the same specifications.
So, how do the new 15-inch MacBook Pros perform?
The new processors bring between a 3% to 5% increase in performance to each MacBook Pro model according to the Geekbench 2 scores. This increase is entirely due to, and in line with, the increase in processor speed.
One thing to note is that the new mid-range Retina MacBook Pro has the same speed processor as the old high-end Retina MacBook Pro. However, the new mid-range model is slightly slower than the old high-end model. While this seems surprising at first, the difference is easily explained by comparing the two processors: the old high-end processor has more cache than the new mid-range processor.
13-inch Retina MacBook Pro
Here is a summary of the processor upgrades for the different 13-inch Retina MacBook Pros:
Like the 15-inch Retina MacBook Pro, each 13-inch Retina MacBook Pro model receives a 100 MHz increase in processor speed.
However, unlike the 15-inch Retina MacBook Pro, which only ships with the new processors, the 13-inch Retina MacBook Pro ships with both the old and the new processors. The old processors appear in 13-inch models that start with 128GB of storage, while the new processors appear in 13-inch models that start with 256GB of storage.
How do the new 13-inch MacBook Pros perform?
It turns out the new 13-inch MacBook Pros receive a boost similar to the new 15-inch MacBook Pros. The new processors bring a 3% to 5% increase in performance according to their Geekbench 2 scores. Again, this increase is in line with the increase in processor speed.
With Haswell, Intel's new processor architecture, at least four months away, I'm not surprised Apple updated the Retina MacBook Pro lineup with the latest Ivy Bridge processors. I am surprised Apple didn't update the MacBook Pro lineup at the same time. Both the 13-inch and 15-inch MacBook Pro come with the older Ivy Bridge processors (although you can upgrade the 15-inch MacBook Pro to the newer Ivy Bridge processors). I wonder if this is an indication that the Retina MacBook Pro is more popular than the MacBook Pro, or perhaps that this is the last generation of the MacBook Pro.
I imagine we'll find out what happens to the MacBook Pro later this year when Apple releases new Haswell-based laptops.
I'm excited to announce that Geekbench 2, Primate Labs' popular cross-platform benchmark, is now available for BlackBerry 10. Geekbench 2 for BlackBerry is available for download from BlackBerry World for $0.99.
While Geekbench 2 for BlackBerry will look familiar for those who have used Geekbench 2 for iOS or for Android, Geekbench 2 for BlackBerry features a new interface written for BlackBerry 10 using the Cascades API.
Writing a new interface on a new platform can be challenging. That wasn't the case with BlackBerry 10, though. Cascades, which is based on Qt, is a very productive API. From start to finish, the new BlackBerry interface took half the time to write and has a third the code of the equivalent Android interface.
Even though the interface is different, Geekbench 2 for BlackBerry uses the same benchmarks as Geekbench 2 on other platforms. Not only can you measure the performance of your BlackBerry, you can also compare the performance of your BlackBerry with the performance of other handsets (such as the iPhone 5 and the Samsung Galaxy S III) using Geekbench.
Also, since Geekbench 2 was designed with multi-core processors in mind, Geekbench 2 for BlackBerry can measure the full potential of the dual-core processors found in the new BlackBerry 10 handsets.
Finally, Geekbench 2 for BlackBerry works with the Geekbench Browser. You can upload your Geekbench results from your BlackBerry and share them with other Geekbench users. You can also compare your results with any one of the 1.6 million results in the Geekbench Browser.
If you have a BlackBerry 10 device (such as the BlackBerry Z10), you can grab Geekbench 2 for BlackBerry from BlackBerry World today for only $0.99.
Now that 2012 is almost over I thought it would be interesting to look at all of the Macs released this year and see how they perform compared to one another. I've collected Geekbench results from the Mac Benchmark chart for all of the Macs that are currently available from Apple.
Not familiar with Geekbench? It's a cross-platform benchmark that measures processor and memory performance. Higher scores are better. Curious how your Mac compares (especially if you're wondering if it's time to upgrade)? Download Geekbench and find out how your Mac measures up to the Macs on this chart.
Mac Performance (2012 Models)
A few things from the chart stood out at me:
- Even though the Mac Pro hasn't had a significant update in over two years, the 6- and 12-core Mac Pros are still the fastest Macs available. However, the Mac Pro processor architecture is starting to show its age. Both the 4-core and 8-core Mac Pros have been eclipsed by Apple's latest 4-core laptops and desktops.
- There's only a minor performance difference between the quad-core Core i7 desktops and the quad-core Core i7 laptops. You no longer need a large Mac (or, for that matter, a desktop Mac) to have a fast Mac.
- When the MacBook Air was introduced in 2008, many were concerned whether it was fast enough to get "real work" done. Now, while the low-end 11-inch MacBook Air is still the slowest Mac available, the high-end 11-inch and 13-inch MacBook Airs are competitive with the rest of Apple's portable lineup. Only the 15-inch MacBook Pros, with their quad-core processors, are significantly faster than the MacBook Air.
- Both the mid- and high-end Mac minis are surprisingly fast for such small computers. It's now possible to buy a Mac that scores over 10,000 in Geekbench, uses less than 11W of power when idle.
While 2012 has been a good year for Mac performance, I'm excited to see what Apple has in store for 2013. I'm hoping that the Mac Pro rumors are true, and that we'll see new Mac Pros sometime next year. I'm also excited to see what Intel's new Haswell micro-architecture will bring to the Mac, both in terms of higher performance and lower power use.
Now that the new 27-inch iMac is available for purchase I've been receiving requests from people who are wondering if they should purchase a 27-inch iMac, or whether a 21.5-inch iMac is fast enough for their needs.
To help answer that question I've collected Geekbench results from the Geekbench Browser for both the 27-inch iMac and the 21.5-inch iMac. If you're not familiar with Geekbench, it is a cross-platform benchmark that measures processor and memory performance. Higher Geekbench scores are better.
If you're curious how your Mac (or PC!) compares to the new iMacs you can download Geekbench and find out (it only takes a couple of minutes to run).
iMac Performance (27-inch 2012 and 27-inch 2011)
Before comparing the new 27-inch iMac to the 21.5-inch iMac, I was curious to see how the new 27-inch iMac compared to the old 27-inch iMac. As you can see from the chart, the new 27-inch iMac models are between 13% and 19% faster than last years models (when comparing equivalent models). This increase is thanks to both the new Ivy Bridge processors and, for the mid- and low-end iMac, a slight bump in processor speed.
iMac Performance (27-inch 2012 and 21.5-inch 2012)
Somewhat surprisingly, the new 27-inch iMac models aren't that much faster than the new 21.5-inch iMac models. Only the low-end 27-inch iMac is significantly faster (13%) than the low-end 21.5-inch iMac. The other 27-inch models are only 3-4% faster than the equivalent 21.5-inch models.
Should you get a 21.5-inch iMac or a 27-inch iMac? Unless you're looking at the low-end models, processor performance isn't different enough to recommend one over the other. Which Mac you purchase will come down to personal preferences. Do you want a bigger screen with a more powerful GPU? Or do you want a smaller screen that takes up less space? Either way, regardless of which iMac you choose, you'll be happy with its performance.
While Apple introduced the new iMac last month, Apple only started shipping the new 21.5-inch iMac today (and won't start shipping the new 27-inch iMac until next month). The new iMacs, along with a redesigned body, feature Intel's new Ivy Bridge processors. How do these new processors perform compared to the Sandy Bridge processors found in the previous-generation iMac? How does the new iMac perform compared to Apple's other desktop computers?
To answer these questions I've charted Geekbench results gathered from the Geekbench Browser for the following Macs: iMac (Late 2012), iMac (Mid 2011), Mac mini (Late 2012), Mac Pro (Mid 2012), and Mac Pro (Mid 2010).
If you're not familiar with Geekbench, it's a cross-platform benchmark that measures processor and memory performance. Higher scores are better, where double the score means double the performance. You can download Geekbench and use it to find out how your computer compares against the Macs in this article.
The benefits of the new Ivy Bridge processors are clear; the new high-end 21.5-inch iMac is almost 10% faster than the old high-end 27-inch iMac. The new mid-range and low-end iMacs show similar improvements as well.
iMac and Mac mini Performance
Both the iMac and the Mac mini use Ivy Bridge processors, so neither Mac has an architectural advantage over the other. What's interesting here, though, is how the quad-core Core i5 iMacs perform compared to the quad-core Core i7 Mac minis. Since Core i7 has hyper-threading technology (and the Core i5 does not), it can execute more instructions at once, leading to higher performance.
Here this means that the mid-range Mac mini is faster than the mid-range iMac that's almost twice the price. True, you do get a display and a discrete GPU with the iMac, but these Geekbench results show how powerful the new Mac mini is despite its size.
iMac and Mac Pro Performance
Since Apple hasn't meaningfully updated the Mac Pro since 2010, the Mac Pro is stuck with the older and outdated Westmere and Nehalem processors. This handicaps its performance to a point where the new iMac is faster than some of the current Mac Pros. In fact, only the 12-core Mac Pros really distinguish themselves from the new iMacs (both in performance and in price).
The new 21.5-inch iMac isn't a bad upgrade (at least from a performance perspective). Intel's Ivy Bridge processors give the new iMac a nice boost over the previous-generation iMac, and help make it competitive with all but the high-end Mac Pro.
I'm still not sure, though, whether I'd choose a new iMac over a quad-core Mac mini (especially since I already own a nice display).