MSI GT70 (GTX 970M) Notebook Review

[Redaktion]

Gigantic with drawbacks As expected from MSI's GT lineup, we meet uncompromising readiness to perform. The update of the GT70 is faster and in some cases even cheaper. However, the weaknesses found in the old tests unfortunately have not been eliminated.

http://www.notebookcheck.net/MSI-GT70-GTX-970M-Notebook-Review.142664.0.html

[Unigine Heaven Benchmark?]

Where's your Unigine Heaven Benchmark stress test that you have normally been doing recently?  This is far more useful to consumers than your Prime & Furmark stress testing which is an unrealistically & artificially high workload which tells nothing about the performance capabilities of the laptop.  Your extended Unigine Heaven Benchmark stress test was more indicative of a long gaming session load, and you used to show the stable CPU & GPU clocks & temperatures that were achieved in this test - this is what really dictates performance in real scenarios.  If I was to buy a laptop, then this is the test I would seek out, combined with your various 3DMark & gaming benchmark results.  Please start reinstating the Unigine Heaven Benchmark test, it's one of the most useful pieces of your reviews - if you need to cut down testing workload just go ahead & take the bold step of removing the Furmark & Prime 95 tests, they don't tell us anything, they're outdated tests with modern hardware.

[Klaus Hinum]

I think Furmark & Prime are very important to test the cooling solution if it can handle the stress. Especially as the cooling solutions get worse over time (dirt, heat paste dries) a bit of a "headroom" always helps. And e.g. Battlefield 3 on some laptops even produced higher power consumption readings in my tests than the stress test, therefore its not that far fetched in my opinion.

[Unigine Heaven Benchmark?]

The problem with Furmark is that it is a power virus, and the NVidia drivers detect the presence of the Furmark.exe running & throttle the GPU artificially as soon as that particular program is detected, if they didn't do this it would override the Wattage limits of what they deem safe for their GPUs - they've been doing this since the introduction of Fermi (early 2010).  This makes Furmark irrelevant as a test.

In fairness, Prime95 is probably more relevant than Furmark, as Intel doesn't artificially throttle their CPUs when Prime95 is run - instead they just throttle for thermal and/or when & if they hit their maximum power limits - just as they would if any other power intensive program was run on their CPUs.

Unigine Heaven Benchmark along with the stable GPU & CPU clocks that notebookcheck used to show is far more relevant than Prime95 and Furmark tests.  If all notebooks throttle when running Prime95 & Furmark then the test tells us nothing, because it's just too strenuous, therefore there are no distinctions between laptops to be made when running those tests when no laptop can clear the bar.