We'll get around to writing a whitepaper once we finish up coding our mobile benchmark application.
There is a long back-story to the battery life benchmark used in this review. Chris and I have decided to save that long explanation for another time. What you should know is that it is similar to BAPCo’s MobileMark, but it differs in a few ways. Like MobileMark, it is a workload-based benchmark, running processes through several programs. However, this is a benchmark that we've coded from scratch, so that is where the similarity ends.
We want to stress real-world usage, which is perhaps one of the biggest reasons we decided to tackle an in-house benchmark. This metric mimics what you should expect from everyday usage. Right now, we've programmed the battery life test to simulate a user typing at ~45 WPM and reading at ~200 WPM. So, this is a “Real Life Use” benchmark, hence the name--RLUMark (at least until we think of a better name).
Since we are testing netbooks, there is no need to include content creation programs like those from the Adobe CS5 suite. This limits benchmarking to the General Use Battery Workload.
This workload consists of the following programs
- Excel 2010
- Word 2010
- Outlook 2010
We will try and keep the benchmark as up-to-date as possible. Right now, everything has been updated to 8/28/2010.
In addition, we are always going to benchmark systems as they ship, including their default battery profiles. There is no clean wipe of the OS here. In real life, when you buy a notebook, system vendors rarely include a blank copy of Windows 7. Some of the included software is useless, such as trial software, but others programs are important for functionality, for example, ThinkVantage’s Power Manager.
Beyond turning down all of the “special offers” when starting the system for the first time and installing Office 2010 Professional Plus, we do not disable or uninstall any software. Bloatware naturally hogs more system resources during the benchmark process, so we want to encourage manufacturers to cut down on the inclusion of worthless software. We want to make a point that the time it takes to complete a benchmark workload is unaffected by included software.
Test Conditions for All Systems:
- Windows 7, all patches updated to 10/30/2010
- BIOS updates, current as of 10/30/2010
- Master Audio Volume: 50%
- System Drivers, current as of 10/30/2010
- Graphic Drivers, current as of 10/30/2010
- GMA 3150 - 22.214.171.1240AMD Mobility Radeon HD 4225 – Catalyst 10.11Nvidia Ion 2 - 206.99GMA 4500MHD - 126.96.36.1996
Some vendors tweak their Windows 7 battery settings a bit in order to maximize longevity. There isn’t anything wrong with this. Every company has a different battery strategy that it believes is the best for its system. For example, some notebooks are hard-wired to force hibernation at 5%. Other systems will let you go all the way to 0% and just die, no matter what you set in the battery profile. We are going to test at default shipping settings under the “Balanced” battery profile.
Some manufacturers have a different name for this profile, but this will always be the “Recommended” Windows 7 profile. To simulate the same visual experience, we only “untweak” the display settings to retail Window 7 settings. In addition, all displays have been set to ~100 nits brightness, which is usually the 50% brightness level for most systems.
|Other Benchmarks and Settings|
|WinRAR 3.93||Version 3.93 (32-bit), Benchmark: THG-Workload (334 MB)|
|Synthetic Benchmarks and Settings|
|PCMark Vantage||Version: 1.00, System, Memories, TV and Movies, and Productivity benchmarks, Windows Media Player 10.00.00.3646|
|World of Warcraft: Cataclysm 4.0.3||Low and Fair Quality Setting, V-sync off, 1366x768 and 1280x720, two-minute sequence, Fraps|
|Civilization 5||Modified Default Settings, vsync off, one-minute zoom and pan sequence, Steam Version, Fraps|
|Call of Duty: Black Ops||Modified Default Setting, vsync off, No AA, No AF, Texture Quality: Normal, Trilinear, two-minute game play sequence, Steam Version, Fraps|