Skip to main content

MSI GT60 2PC Dominator Review: A Fast Notebook With Battery Boost

Power, Battery Life, And Efficiency

Our power tests require the notebook be plugged in, though that shouldn’t dissuade anyone who might like to calculate unplugged energy draw based on the battery’s watt-hour rating. Certainly the higher-clocked parts in the newer machine need a bit of extra juice to operate at full performance.

Battery Boost’s approach to power savings is to give you all the performance you need to maintain a playable frame rate, and nothing more. That means if your settings are conservative enough to allow ultra-high frame rates while plugged into the wall, the software can drastically reduce GPU frequency (and power consumption) while unplugged to further extend battery life compared to previous, percent-based power reduction methods.

Reading and Working refer to PCMark 8’s benchmarks, which automatically exit at 20% battery life and thus require recalculation to reach 5%. The Viewing figure comes from Windows Media Player using a 720p, H.264-encoded video.

The Gaming test requires some additional explanation. In order to create a looping, repeatable workload, I started with 3DMark 11's first test and ran until the systems shut off. Here's the thing, though: because 3DMark is a synthetic metric designed to tax graphics hardware, Battery Boost is likely given no opportunity to scale back clock rates, and thus appears to have no effect. Instead, we see the more powerful host and graphics processors exact a hit to battery life. Once we put all of the data together and realized that this wouldn't reflect real-world gaming, we went back to create a more applicable workload.

Metro: Last Light makes for an ideal follow-up because it can be looped and its settings are highly tunable. When I crank up the detail settings to bring average framerate down (to around 34 FPS with the power adapter plugged in), the results I recorded in 3DMark are largely repeated. I can drop the graphics quality, however, generating a much lighter load, and increase battery life by an astounding 96%.

That still doesn't reach the previous machine's battery life, though, since that older system sports less powerful hardware.

Perhaps the biggest issue with this test is that Metro: Last Light's most mainstream DirectX 11 settings are still quite taxing to a mobile GPU. Plugged in, I was only seeing 90 FPS at my entry-level combination of options. But as stressful as Metro appears to be, we've seen Battlefield 4 and Far Cry 3 run at similar performance levels using their medium quality settings.

We're not going to strike off in search of a modern game that generates hundreds of frames per second at ultra-low detail levels simply to validate the potential gains of Battery Boost. Instead, we'll note that the technology's best application is as a performance normalizer when you're away from a wall socket, where battery life can be extended without pushing frame rates you don't need. Fixed-frequency power-saving mode would have caused minimum framerates to drop below playable performance levels, as shown by the GeForce GTX 770M in the Battlefield 4 line graphs a few pages ago.

Although Nvidia's software-based technology was designed to limit performance in exchange for greater battery life, letting the hardware run at full speed shows that MSI's newer notebook can often be less efficient than its predecessor. There's simply no free lunch. None of the components benefit from more advanced manufacturing technology, and the ones that affect performance are operated at higher clock rates to move the needle.

If you do choose not to stress the hardware when running off the battery, you’ll be rewarded with over six hours of reading and over five hours of movie viewing. Those are the two applications where the GT60 2PC shines, and they’re also the two activities you'll most likely engage in as you sit on an airplane.

Thomas Soderstrom
Thomas Soderstrom is a Senior Staff Editor at Tom's Hardware US. He tests and reviews cases, cooling, memory and motherboards.