System Performance

The new benchmark suite has left us in a bit of disarray, but thankfully we did have the CyberPowerPC Fangbook in recently. The Fangbook sports an Ivy Bridge CPU with exactly the same clocks as the i7-4700MQ powering the MSI GT70 Dragon Edition, and that gives us an excellent baseline for comparison.

I'd also like to personally thank NVIDIA for furnishing us with an Alienware M17x R4 equipped with a GTX 680M at the eleventh hour so we can get proper comparative results. Unfortunately, as you'll see the 780M in the GT70 Dragon Edition gets hamstrung by the CPU (more on this later) when tested playing games that hit the GPU and the CPU especially hard.

PCMark 7 (2013)

Cinebench R11.5 - Single-Threaded Benchmark

Cinebench R11.5 - Multi-Threaded Benchmark

x264 HD 5.x

x264 HD 5.x

Haswell is certainly faster clock for clock, but obviously owners of Ivy Bridge-based notebooks shouldn't be chucking their old kit. Under most circumstances, the i7-4700MQ is less than 10% faster than the i7-3630QM.

Futuremark 3DMark (2013)

Futuremark 3DMark (2013)

Futuremark 3DMark (2013)

Futuremark 3DMark 11

Pop on over to the 3DMarks and the 780M looks good, but unusually, not quite as good as it ought to. It has a commanding lead over the GTX 580M, and it should, but on Cloud Gate and Ice Storm it should also be blowing right past the 7970M. You'll get a better idea of what I'm talking about on the next page, when we look at actual gaming performance.

Introducing the MSI GT70 Dragon Edition Gaming Performance
Comments Locked

115 Comments

View All Comments

  • Shayler - Monday, June 24, 2013 - link

    I must say, after comparing this article with Notebookcheck's much more professional and thorough review, I'm disappointed in your outcome.

    With such comments as, "...a few runs of games and some stress testing..." I'm forced to question your methodology which throws your entire review into question.

    Please see Notebookcheck's test of this system (the non dragon edition) and reevaluate your review.

    http://www.notebookcheck.net/Review-MSI-GT70H-80M4...
  • JarredWalton - Monday, June 24, 2013 - link

    Given that Notebookcheck's review is of a slightly different build, it may not have the issues we noticed in this review. Our review is what we would expect an end user to get if they buy the GT70 Dragon with i7-4700MQ and GTX 780M right now; if you get a similar system with the MSI chassis but assembled by a boutique that uses different thermal paste, you may see better results. If you're willing to disassemble your notebook to replace the thermal grease, you may also see improved performance and lower temperatures, but we don't expect most end users to resort to such measures.
  • michael777 - Tuesday, June 25, 2013 - link

    Thats great but to come back to my question you cant have a laptop with these specs (considering MSI GT70H-80M4811B) for the price of 2000 euros. So what am I actually risking apart from the hot machine and few frames less in few games?
  • Shayler - Tuesday, June 25, 2013 - link

    Hi Jarred,

    Thanks for taking the time to respond.

    The internals on these systems are identical. The MSI GT60/70 (regardless of model name/number) and all bare bones use the same cooling system. In fact, you can remove the all of these parts and interchange them. Including the motherboards!

    My thought is, you received a bad sample. No one expects the end user to repaste their system (even though it's very easy on these models), but from personal experience with this model, I know for a fact that the temperatures you're reporting are far above normal even with silicon based thermal compounds.

    My concern with your review is the lack of documentation. Where's your methodology? What games did you play, at what settings, and for how long? Wheat benchmarks did you run? What were your ambient temperatures?

    Contact MSI, have them send you a new sample, document your methodology and results, and reevaluate your review.

    If you have the same results, post them and have your, "I told you so." moment.
  • JarredWalton - Tuesday, June 25, 2013 - link

    My point about the internals is that special editions (e.g. the Dragon Edition) may go through a modified assembly line at the factory. Dell and others do this as well -- it's an optimization for a SKU that they expect to sell more units, and it simplifies things as they don't have to custom build every single notebook. You just put the same parts in every Dragon and thus it's faster and easier -- but what if you happen to have a minor glitch on the thermal paste application?

    Oh, if only it were that easy. I'll see if Dustin wants to do a follow up...or I can ask as well. Sometimes it's difficult to get even one notebook sample in for review, and if it's not a favorable review they may just try to cut their losses. Anyway, no promises that it will happen, but I'll at least give it a try. Stay tuned....
  • michael777 - Tuesday, June 25, 2013 - link

    Looks like its a bit hard for the author(s) of that review to find an answer for my pretty simple question and, in fact, their own statement. So ill try again, what laptop with i7-4700MQ and gtx780 can you find for the similar money even if the cooling isnt the strongest point of this one?
  • JarredWalton - Tuesday, June 25, 2013 - link

    Here are a few vendors stocking Clevo builds (no mSATA RAID, but that's not a huge issue IMO)....

    AVADirect:
    http://www.avadirect.com/gaming-laptop-configurato...
    http://www.avadirect.com/gaming-laptop-configurato...
    http://www.avadirect.com/gaming-laptop-configurato...

    Eurocom (http://www.eurocom.com/) -- look at their gaming notebooks X3/X5/X7.

    OriginPC (tend to be a bit more expensive): http://www.originpc.com/gaming/laptops/

    Mythlogic: http://www.mythlogic.com/mobile.php (Pollux 1613)

    Alienware 17: http://www.dell.com/us/p/alienware-17/fs

    If you're looking at a $2700 notebook like this, there are certainly other options, including beastly SLI setups if you don't mind spending even more money and carrying around a gigantic notebook. They all have pros and cons for certain, but I'd be more inclined to spend a few extra hundred if it's the difference between throttling/cooler temperatures and not. Or maybe it really is just the thermal paste -- we'll see if we can get a second sample for testing.
  • Trkkr - Tuesday, August 20, 2013 - link

    I'm looking for the same answers as michael. Jarred, do you know if any of the manufacturers you have listed here have had their 17.3" GTX 780M models reviewed? If they haven't what's to keep their systems from suffering from cooling issues as well (they might have dual fans but that doesn't mean their systems will stay any cooler). Thanks.
  • Globemaster - Thursday, June 27, 2013 - link

    How do you reconcile the folks at gentechpc getting temps of 60 CPU and 65 GPU when running the similar benchmarks on the same Dragon Edition 2 at http://www.youtube.com/watch?v=jVjv2gcB7uc. I know they have a commercial interest, but doubt they faked the video. I'll be getting mine tomorrow and will try to replicate what you had here. Rest assured, if I get these results, it'll be going back. I just really think you got a lemon. That should be noted, but in my experience with reviews (even of SimCity V), most reviewers try to evaluate a working version of a product while making note of the possible quality control issues rather than just trash the product completely when results come in that seem unreasonable. Possibly at least worth a note in the article that you might have just gotten a lemon, rather than making people search through the comments to find alternate hypotheses.
  • Globemaster - Friday, July 5, 2013 - link

    Just upgraded to 320.49 drivers and got 5002 on Firestrike 3DMark 2013 (same config 4700MQ and 780M) and temps never exceeded 74 on CPU or GPU. Not sure why you got a better score on older drivers with such high temps here (was yours force overclocked???), but I had hoped for over 5000 and hit it. I was getting 4600 with the stock 311.xx drivers. The only scores I'm seeing higher are oc'd through bios or 3rd party programs. My 2 year old Sager with a 485M only got 1563, so this more than triples performance. I'm thriled with this laptop and will be keeping it.

Log in

Don't have an account? Sign up now