Category: Benchmarking

BatteryXPRT’s future

on December 1, 2016

A few weeks ago, we discussed the future of HDXPRT. This week, we’re focusing on the current state of BatteryXPRT 2014 for Android, and how the benchmark may evolve in 2017.

BatteryXPRT continues to provide users with reliable evaluations of their Android device’s performance and battery life under real-world conditions. Originally designed to be compatible with Android 4.2 (Jelly Bean) and above, the benchmark continues to work well on subsequent versions of Android, up to and including Android 6.0 (Marshmallow).

Since Android 7 (Nougat) began to roll out on select devices in the past few months, our internal testing has shown that we’ll need to adjust the BatteryXPRT source code to maintain compatibility with devices running Android 7 and above. We developed the existing source when Eclipse was the officially supported SDK environment, and now we need to bring the code in line with the current Android Studio SDK.

Practically speaking, BatteryXPRT does run on Nougat, and to the best of our knowledge, battery life results are still accurate and reliable. However, the test will not produce a performance score. As more Nougat devices are released in the coming months, it’s possible that other aspects of the benchmark may encounter issues. If this happens during your testing, we encourage you to let us know.

Because MobileXPRT 2015 and BatteryXPRT 2014 performance workloads are so closely related, the next obvious question is whether MobileXPRT 2015 runs on Nougat devices. As of now, MobileXPRT 2015 does run successfully and reliably on Android 7, and this is because the most recent build of MobileXPRT 2015 was compiled using a newer SDK.

We think the best course of action for MobileXPRT 2015 and BatteryXPRT will be to eventually combine them into a single, easy-to-use Android benchmark for performance and battery life. We’ll talk more about that plan in the coming months, and we look forward to hearing your input. Until that transition is successful, we’ll continue to support both BatteryXPRT and MobileXPRT 2015.

As always, we welcome your feedback on these developments, as well as any ideas you may have for future XPRTs.

Justin

Posted in Android, Android L, Android M, Android N, Battery life, BatteryXPRT 2014 for Android, Collaborative benchmark development, MobileXPRT 2015 |

Machine learning

By Bill Catchings

on November 23, 2016

A couple months ago I wrote about doing an inventory of our XPRT tools. Part of that is taking a close look at the six existing XPRTs. The first result of that effort was what I recently wrote about HDXPRT. We’re also looking at emerging technology areas where the BenchmarkXPRT Community has expertise that can guide us.

One of the most exciting of these areas is machine learning. It has rapidly gone from interesting theoretical research (they called them “neural nets” back when I was getting my computer science degree) to something we all use whether we realize it or not. Machine learning (or deep learning) is in everything from intelligent home assistants to autonomous automobiles to industrial device monitoring to personalized shopping in retail environments.

The challenge with developing a benchmark for machine learning is that these are still the early days of the technology. In the past, XPRTs have targeted technologies later in the product cycle. We’re wondering how the XPRT model and the members of its community can play a role here.

One possible use of a machine-learning XPRT is with drones, a market that includes many vendors. Consumers, hobbyists, builders, and the companies creating off-the-shelf models could all benefit from tools and techniques that fairly compare drone performance.

The best approach we’ve come up with to define a machine-learning XPRT starts with identifying common areas such as computer vision, natural language processing, and data analytics, and then, within each of those areas, identifying common algorithms such as AlexNet, GoogLeNet, and VGG. We would also look at the commonly used frameworks such as Caffe, Theano, TensorFlow, and CNTK.

The result might differ from an existing XPRT where you simply run a tool and get a result. Instead, it might take the form of sample code and workloads. Or, maybe even one or two executables that could be used in the most common environments.

At this point, our biggest question is, What do you think? Is this an area you’re interested in? If so, what would you like to see a machine-learning XPRT do?

We’re actively engaging with people in these emerging markets to gauge their interest as well. Regardless of the feedback, we’re excited about the possibilities!

Bill

Posted in Collaborative benchmark development, Drones, Future of performance evaluation, HDXPRT, Machine learning |

HDXPRT’s future

By Bill Catchings

on November 3, 2016

While industry pundits have written many words about the death of the PC, Windows PCs are going through a renaissance. No longer do you just choose between a desktop or a laptop in beige or black. There has been an explosion of choices.

Whether you want a super-thin notebook, a tablet, or a two-in-one device, the market has something to offer. Desktop systems can be small devices on your desk, all-in-ones with the PC built into the monitor, or old-style boxes that sit on the floor. You can go with something inexpensive that will be sufficient for many tasks or invest in a super-powerful PC capable of driving today’s latest VR devices. Or you can get a new Microsoft Surface Studio, an example of the new types of devices entering the PC scene.

The current proliferation of PC choices means that tools that help buyers understand the performance differences between systems are more important than they have been in years. Because HDXPRT is one such tool, we expect demand for it to increase.

We have many tasks ahead of us as we prepare for this increased demand. The first is to release a version of HDXPRT 2014 that doesn’t require a patch. We are working on that and should have something ready later this month.

For the other tasks, we need your input. We believe we need to update HDXPRT to reflect the world of high-definition content. It’s tempting to simply change the name to UHDXPRT, but this was our first XPRT and I’m partial to the original name. How about you?

As far as tests, what should a 2017 version of HDXPRT include? We think 4K-related workloads are a must, but aren’t sure whether 4K playback tests are the way to go. What do you think? We need to update other content, such as photo and video resolutions, and replace outdated applications with current versions. Would a VR test would be worthwhile?

Please share your thoughts with us over the coming weeks as we put together a plan for the next version of HDXPRT!

Bill

Posted in Application-based benchmarks, BenchmarkXPRT development community, Collaborative benchmark development, Future of performance evaluation, HDXPRT, HDXPRT development process, Performance benchmarking, What makes a good benchmark?, Windows 10 |

Tracking device evolution with WebXPRT ’15, part 2

By Justin Greene

on October 27, 2016

Last week, we used the Apple iPhone as a test case to show how hardware advances are often reflected in benchmark scores over time. When we compared WebXPRT 2015 scores for various iPhone models, we saw a clear trend of progressively higher scores as we moved from phones with an A7 chip to phones with A8, A9, and A10 Fusion chips. Performance increases over time are not surprising, but WebXPRT ’15 scores also showed us that upgrading from an iPhone 6 to an iPhone 6s is likely to have a much greater impact on web-browsing performance than upgrading from an iPhone 6s to an iPhone 7.

This week, we’re revisiting our iPhone test case to see how software updates can boost device performance without any changes in hardware. The original WebXPRT ’15 tests for the iPhone 5s ran on iOS 8.3, and the original tests for the iPhone 6s, 6s Plus, and SE ran on variants of iOS 9. We updated each phone to iOS 10.0.2 and ran several iterations of WebXPRT ’15.

Upgrading from iOS 8.3 to iOS 10 on the iPhone 5s caused a 17% increase in web-browsing performance, as measured by WebXPRT. Upgrading from iOS 9 to iOS 10 on the iPhone 6s, 6s Plus, and SE produced web-browsing performance gains of 2.6%, 3.6%, and 3.1%, respectively.

The chart below shows the WebXPRT ’15 scores for a range of iPhones, with each iPhone’s iOS version upgrade noted in parentheses. The dark blue columns on the left represent the original scores, and the light blue columns on the right represent the upgrade scores.

As with our hardware comparison last week, these scores are the median of a range of scores for each device in our database. These scores come both from our own testing and from device reviews from popular tech media outlets.

These results reinforce a message that we repeat often, that many factors other than hardware influence performance. Designing benchmarks that deliver relevant and reliable scores requires taking all factors into account.

What insights have you gained recently from WebXPRT ’15 testing? Let us know!

Justin

Posted in Apple, Application-based benchmarks, Benchmark metrics, Benchmarking, Browser-based benchmarks, iOS, Performance benchmarking, Uncategorized, WebXPRT |

Tracking device evolution with WebXPRT ‘15

By Justin Greene

on October 20, 2016

The XPRT Spotlight on the Apple iPhone 7 Plus gives us a great opportunity to look at the progression of WebXPRT 2015 scores for the iPhone line and see how hardware and software advances are often reflected in benchmark scores over time. This week, we’ll see how the evolution of Apple’s mobile CPU architecture has boosted web-browsing performance. In a future post, we’ll see the impact of iOS development.

As we’ve discussed in the past, multiple factors can influence benchmark results. While we’re currently using the iPhone as a test case, the same principles apply to all types of devices. We should also note that WebXPRT is an excellent gauge of expected web-browsing performance during real-world tasks, which is different than pure CPU performance in isolation.

When looking at WebXPRT ’15 scores in our database, we see that iPhone web-browsing performance has more than doubled in the last three years. In 2013, an iPhone 5s with an Apple A7 chip earned an overall WebXPRT ’15 score of 100. Today, a new iPhone 7 Plus with an A10 Fusion chip reports a score somewhere close to 210. The chart below shows the WebXPRT ’15 scores for a range of iPhones, with each iPhone’s CPU noted in parentheses.

Moving forward from the A7 chip in the iPhone 5s to the A8 chip in the iPhone 6 and the A9 chip in the iPhone 6s and SE, we see consistent score increases. The biggest jump, at over 48%, appears in the transition from the A8 to the A9 chip, implying that folks upgrading from an iPhone 6 or 6 Plus to anything newer would notice a huge difference in web performance.

In general, folks upgrading from an A9-based phone (6S, 6S Plus, or SE) to an A10-based phone (7 and 7 Plus) could expect an increase in web performance of over 6.5%.

The scores we list represent the median of a range of scores for each device in our database. These scores come from our own testing, as well as from device reviews from media outlets such as AnandTech, Notebookcheck, and Tom’s Hardware. It’s worth noting that the highest A9 score in our database (AnandTech’s iPhone SE score of 205) overlaps with the lowest A10 Fusion score (Tom’s Hardware of Germany’s iPhone 7 score of 203), so while the improvement in median scores is clear, performance will vary according to individual phones and other factors.

Soon, we’ll revisit our iPhone test case to see how software updates can boost device performance without any changes in hardware. For more details on the newest iPhones, visit the Spotlight comparison page to see how iPhone 7 and 7 Plus specs and WebXPRT scores stack up.

Justin

Posted in Apple, Benchmark metrics, Benchmarking, Performance benchmarking, WebXPRT, WebXPRT 2015 |

Taking a detour

By Eric Hale

on October 13, 2016

Back in April, Bill announced that we would be starting development of a cross-platform benchmark. This announcement generated a lot of interest and we got lots of good feedback and ideas.

We knew from the start that getting a cross-platform benchmark right would be hard. However, it proved to be even trickier than we thought. As I explained before, benchmarks not only have to run well, but the results must be fair to all platforms involved. Achieving both of these requirements has been a challenge.

At the same time we’ve been devoting a great deal of effort and resources to the cross-platform benchmark, some increasingly popular new use cases have been receiving less attention than they deserve. We’ve decided that the cross-platform benchmark is not the best use of the Community’s resources, and are going to put it on the shelf for a while. This will free up the resources to let us really dig into some newer technologies.

Thanks again to everyone who responded.

Eric

Posted in Benchmarking, BenchmarkXPRT development community, Collaborative benchmark development, Cross-platform benchmarks |

Category: Benchmarking

BatteryXPRT’s future

Machine learning

HDXPRT’s future

Tracking device evolution with WebXPRT ’15, part 2

Tracking device evolution with WebXPRT ‘15

Taking a detour

Check out the other XPRTs: