Category: Benchmarking

An anniversary update

on July 21, 2016

The Windows 10 Anniversary Update release is scheduled for August 2, and we’ve been running the XPRTs on the Windows Insider preview builds. While we can’t publish performance data from developer builds, we’re happy to say that WebXPRT and TouchXPRT run well on the Anniversary Update.

The story for HDXPRT 2014 is more complicated. Back in May, we reported that it would not run on more recent versions of Windows. However, we’ve identified steps that enable HDXPRT to run on the current stable Windows 10 build, as well as the latest Anniversary Update preview. It’s running well, but it’s possible that testers will encounter other issues as Microsoft releases new builds.

We have included the steps below. We’re considering an update to HDXPRT 2014 that will incorporate these changes. If you have any comments or suggestions related to HDXPRT, please let us know.

Justin

Summary
In addition to the normal system configuration requirements for HDXPRT, testers must also overwrite HDXPRT’s CPU-Z files with newer versions and change the default browser from Microsoft Edge to Internet Explorer. After configuring the system for HDXPRT testing, testers may encounter errors related to administrative privileges when attempting to launch Microsoft Edge. Returning User Account Control settings to their default pre-configuration state resolves the problem.

Process
1. Install the latest version of CPU-Z.
      a. Open any browser and download the latest version of CPU-Z for Windows
          (currently CPU-Z 1.76).
      b. Install CPU-Z on the system, using the default settings and installation path.
2. Install the HDXPRT 2014 benchmark using the default installation process. Reboot the system
    after installation.
3. Copy all the files from the C:\Program Files\CPUID\CPU-Z\ directory to the C:\Program Files
    (x86)\HDXPRT\bin, and overwrite the existing CPU-Z files.
4. Change the default browser from Microsoft Edge to Internet Explorer:
      a. Open the Windows Settings app and select System/Default apps.
      b. Under Web browser, click the Edge icon, and select Internet Explorer from the list.
      c. At the Before you switch window, click Switch anyway.
      d. Close the Settings app.
5. Adjust SmartScreen and security settings:
      a. Open Internet Explorer.
      b. Go to Settings/Internet options/Security, and make the following changes for the Internet
           and Trusted Sites zones:
            i. Select Custom Level.
            ii. Disable SmartScreen Filter.
            iii. Under Launching applications and unsafe files, click Enable (not Secure).
            iv. Click OK, and click Apply. If a warning message appears, click Yes.
6. Restart the system.
7. Open HDXPRT and run the benchmark normally.

If, after installing HDXPRT, you encounter an error related to administrative permissions when trying to open Microsoft Edge, return User Account Controls to the default setting, and restart the system. The default User Account Control setting is the third notch from the bottom: “Notify me only when apps try to makes changes to my computer.”

Posted in BenchmarkXPRT development community, Collaborative benchmark development, HDXPRT, HDXPRT 2014, Performance benchmarking, TouchXPRT, Uncategorized, WebXPRT 2015, Windows 10, Windows 10 Anniversary Update |

Open source?

By Bill Catchings

on July 14, 2016

We’re proud of the BenchmarkXPRT Development Community and its accomplishments over the last five years. We’re also thankful for the contributions the members of the community have made. One of the benefits of membership is access to the source code for all the XPRT performance tools. This has meant that the code is available to anyone willing to take the easy step of joining the community.

Behind our decision to use this model rather than a more traditional, open-source model was the need to control derivative works. The license agreement for the source allows members to modify the source, but not to claim that the results from that derivative code are XPRT results. For example, as a member, you may download the TouchXPRT source and modify the workloads for your specific purposes, but you can’t refer to the results as TouchXPRT results.

After much thought and discussion, we have come to believe that we can protect the benchmarks’ reputation within a traditional, open-source framework. While our original concerns are still valid, we think that the success and stature of the XPRTs is such that we can make it available via open source.

However, before we take this step, we want to hear the thoughts, concerns, and opinions of both our community members and the wider public.

Please note that if we do make the code open source, the other benefits of being a member—access to requests for comment, design documents, and community previews—will not change.

Please let us know that you think. Email us or contact us on Twitter.

Bill

Posted in BenchmarkXPRT, BenchmarkXPRT development community, Collaborative benchmark development, Open Source, TouchXPRT, What makes a good benchmark? |

Seeing the future

By Eric Hale

on July 7, 2016

Back in April we wrote about how Bill’s trip to IDF16 in Shenzhen got us thinking about future benchmarks. Technologies like virtual reality, the Internet of things, and computer vision are going to open up lots of new applications.

Yesterday I saw an amazing article that talked about an automatic computer vision system that is able to detect early-stage esophageal cancer from endoscopy images. These lesions can be difficult for physicians to detect, and the system did very well when compared to four experts who participated in the test. The article contains a link to the original study, for those of you who want more detail.

To me, this is the stuff of science fiction. It’s a very impressive accomplishment. Clearly, new technologies are going to lead to many new and exciting applications.

While this type of application is more specialized than the typical XPRT, things like this get us really excited about the possibilities for the future. Have you seen an application that impressed you recently? Let us know!

Eric

Posted in Benchmarking, BenchmarkXPRT development community, Collaborative benchmark development, computer vision, Future of performance evaluation, IDF, Internet of things, IoT, Virtual reality, What makes a good benchmark? |

I couldn’t wait!

By Eric Hale

on June 24, 2016

We’ve created a series of videos about the XPRT Women Code-a-Thon. We’ll be talking more about the series, and an exciting new channel, “Women Coding for Change,” in next Thursday’s blog post. However, we posted the first video today, and I wanted to let you know so you can check it out right away!

Eric

Posted in Collaborative benchmark development, Women Coding for Change, XPRT Women Code-a-Thon |

Watching students become masters

By Eric Hale

on June 23, 2016

As you know, last year, PT sponsored a senior project at the Senior Design Center of North Carolina State University (NCSU). The students created Nebula Wolf, a mini game that might evolve into a future benchmark test. It was a valuable collaboration for us and a very educational experience for the students involved.

I’ve talked before about the emerging technologies we’re considering for new benchmarks. Today, I met with the folks at the NCSU Senior Design Center to discuss a possible future project. We’re hoping to harness the immense energy of these students by having them explore one of these new technologies, and then build on what they discover. Nothing is set yet, but we will, as always, keep you informed as things develop.

We’ll be sharing some exciting news about the XPRT Women Code-a-Thon tomorrow. Check back to find out more! Meanwhile, we hope you enjoy as much as we did the University of Washington Tacoma article on student Viveret, the first place winner of the XPRT Women Code-a-Thon.

Eric

Posted in Benchmarking, BenchmarkXPRT development community, Collaborative benchmark development, Future of performance evaluation, Nebula Wolf, North Carolina State University, What makes a good benchmark?, XPRT Women Code-a-Thon |

Getting it right

By Eric Hale

on June 2, 2016

Back in April Bill announced that we are working on a cross-platform benchmark. We asked for your thoughts and comments, and you’ve been great! We really appreciate all the great ideas.

We’ve been using code from MobileXPRT and TouchXPRT as the basis for some experiments. In his post, Bill talked about the difficulty of porting applications. However, even though we have expertise in porting applications, it’s proving more difficult than we originally thought. Benchmarks are held to a higher standard than most applications. It’s not enough for the code to run reliably and efficiently, it must compare the different platforms fairly.

One thing we know for sure: getting it right is going to take a while. However, we owe it to you to make sure that the benchmark is reliable and fair on all platforms it supports. We will, of course, keep you informed as things progress.

In the meantime, keep sending your ideas!
Eric

Posted in Android, Application-based benchmarks, Benchmark metrics, Benchmarking, BenchmarkXPRT development community, Collaborative benchmark development, Future of performance evaluation, iOS, MobileXPRT, Performance benchmarking, TouchXPRT, WebXPRT, Windows 10 |

Category: Benchmarking

An anniversary update

Open source?

Seeing the future

I couldn’t wait!

Watching students become masters

Getting it right

Check out the other XPRTs: