Asus G751 coming with Maxwell GTX 980M and GTX 970M

Discussion in 'ASUS Gaming Notebook Forum' started by Cloudfire, Sep 19, 2014.

Thread Status:
Not open for further replies.
  1. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    7,081
    Messages:
    20,411
    Likes Received:
    25,194
    Trophy Points:
    931
    eleron911, these questions, and more, have been answered in this very long thread already :)

    But, to reiterate, the benefits of Optimus, a few more minutes on battery mode, pale in comparison to the limitations of running all the video out the Intel iGPU instead of directly from the very expensive Nvidia GPU. Optimus isn't worth it.

    Briefly, keeping the iGPU in the video chain keeps it on all the time, adding heat to the CPU die/carrier, even when using the Nvidia GPU for rendering.

    This raised the G750JM/J S/JZ CPU temperatures to 10c higher than the same 100% usage tests on the G750JW/JX/JH - essentially the same CPU's running the same load, 10c hotter - which caused many to go into Thermal Throttling:

    Things to try ~CPU overheating JM/JS/JZ models~
    http://rog.asus.com/forum/showthread.php?48594-Things-to-try-CPU-overheating-JM-JS-JZ-models

    And, putting the iGPU in front of the Nvidia GPU disables all the video controls and features that the Nvidia GPU would have if directly connected to the displays (internal/external), including higher than 60hz operation, 3D, G-Sync, DSR(?), and color management.

    With the new 980m/970m their power uaage is way down from the 700m/800m, so it is less of an advantage to hobble yourself with the very slow Intel Iris GPU.

    Remember, the Intel GPU isn't magic, the rendering efficiency of that part isn't any better, and likely worse, than the Nvidia GPU, so it is going to draw the same power for the same performance, so by using less power it is giving less performance. You may be saving some battery time, but the performance will be less too.

    The only good comparison I have is a battery test done on the G750JH compared to the same test on a G750JZ, the JZ ran on battery 15-30 minutes longer, an improvement to be sure, but not enough to put up with the limitations of Optimus.

    The G751JT/JY don't have Optimus.

    The G750JY, just coming out, has Optimus

    So you have a choice :)

    It would be fun if someone ran a battery test on the G751JY with Battery Boost vs G750JY with Optimus/Intel GPU, and also compared in game performance of both on battery, to see in this next generation what has changed.

    As Nvidia/ATI makes better and better mobile GPU's, with variable de-tuning for battery time vs performance, the need for Intel GPU / Optimus for laptops with discrete GPU's will disappear.
     
    Last edited: Dec 27, 2014
    eleron911 likes this.
  2. eleron911

    eleron911 HighSpeedFreak

    Reputations:
    3,886
    Messages:
    11,104
    Likes Received:
    7
    Trophy Points:
    456
    Thank you sir, I've seen your posts on the asus forums also.

    The thing is, being out of the loop for so long makes me highly susceptible to lack of info or potential. I've done my research for the laptop, I've already owned and changed a 750JS, which ran much warmer than this one (on average, as you said, +10C) but the 970m is a big improvement to the 870M.

    However, I have always undervolted my systems (I've run my X9000 from stock 3.4 Ghz and 1.47V at 3.8 and 1.3V, with temps under 75C :) ) and I'm going to do the same with this laptop, since power usage=temps, which I want to keep as low as possible with this setup.

    I will try what settings I've seen here for XTU, for now, -30mV is very stable with all stress tests and benchmarks I've run, but I get a score of under 500, which is disturbing to say the least. I'll do my research and post updates later on.

    Thanks for the info so far!
     
  3. eleron911

    eleron911 HighSpeedFreak

    Reputations:
    3,886
    Messages:
    11,104
    Likes Received:
    7
    Trophy Points:
    456
    I must be doing something wrong to get these results, I mean it's almost 50% less than what I see for the 4710hq... xtu results.jpg
     
  4. d0r1n

    d0r1n Notebook Guru

    Reputations:
    0
    Messages:
    63
    Likes Received:
    6
    Trophy Points:
    16
    i had the same on 3 of 5 laptops g751jy...i changed 5 laptops because of the bleed,but the 5th one is good,the temps are good.
    i had 2 laptops returned because of the temps too high,about 93proc and 82 video,the 5th one only gets 80proc and 79video
     
  5. eleron911

    eleron911 HighSpeedFreak

    Reputations:
    3,886
    Messages:
    11,104
    Likes Received:
    7
    Trophy Points:
    456
    The temps are ok, highest I've seen for the GPU was 71-72C at an ambient temperature of 25C.

    What worries me is this score, I've checked online and others with the same setup hit 700+. I don't know if that affects performance or not... but I want to make sure whether I should keep it or not.

    How did you get those temps? I'm assuming your setup is a bit different, I have the JT with the 4710hq and 970m GPU, so maybe I should expect lower temps overall.
     
  6. Meaker@Sager

    Meaker@Sager Company Representative

    Reputations:
    8,731
    Messages:
    54,423
    Likes Received:
    15,864
    Trophy Points:
    931
    What performance state are you running in? High performance?
     
    hmscott likes this.
  7. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    7,081
    Messages:
    20,411
    Likes Received:
    25,194
    Trophy Points:
    931
    eleron911, there have been some with low scores, I have been seeing a few more with 50% normal score as well, I haven't debugged XTU recently. If you are running 5.1.1.25 you have the same version as I do, and I am getting the higher score.

    If you are getting normal scores in other benchmarks, I wouldn't worry about the XTU score, XTU is great for tuning, but less so for benchmarking, monitoring, etc. It is meant to provide a single base to do the tuning and monitor the results, but there have been issues with XTU scores for a long time.

    Again, if you are getting expected scores in other benchmarks, don't worry about XTU benchmark results - it has turned into a rat hole for many.

    About a year ago there were still some software incompatibilities, AIDA64 (uninstalled it), spybot (uninstalled for install, then reinstalled), and early on XTU needed a pagefile for full features - Intel fixed that so I am back to running without a pagefile. There were some library requirements, but those should have been resolved as well. For a while there was a bug that some people were getting 50 points less, but that went away as well.

    For some, not all, they need to Run As Administrator to get the XTU installer to work.

    Another rathole bug was when an XTU version was "causing" low temperature Thermal Throttling in some models, I think it was predominantly the JX. That wasted a lot of time for some, maybe even caused them to return their laptops, but uninstalling XTU and using hwinfo to monitor Thermal Throttling showed there was no problem with XTU uninstalled. A new version of XTU fixed it... detecting a pattern? [​IMG]

    I don't think any of those considerations apply any longer, this is new. So I am posting the Intel support, docs, community links, and HWBOT is a partner for the XTU project and has equal or greater support for optimizations and problem support I would go there first - but also check out the Intel support forum.

    I don't have the problem, so I can't work on a fix directly, but I encourage you to carry on and do that and post the results here. As long as you feel you are learning keep going, but if it becomes a frustration, drop it for a while.

    Intel Support hub for XTU
    Support for the Intel® Extreme Tuning Utility (Intel® XTU)

    Intel Docs for XTU
    Documents & Guides for the Intel® Extreme Tuning Utility (Intel® XTU)

    Intel Support forum for Intel Processor Tools
    https://communities.intel.com/community/tech/processors

    HWBOT support forum for XTU
    Intel XTU HWBOT Integration - hwbot.org

    On the other hand, I wouldn't get caught up in XTU benchmark scoring issues.

    There are lots of other benchmarks, and I encourage you to find, install, and test for comparison to assure yourself that your hardware is running fine.

    The normal considerations for performance apply to those benchmarks as well as XTU, so as you optimize your Windows configuration across all the benchmarks, improvements for XTU may occur, and a new release of XTU may fix the current incompatibility that is holding your score low.

    Over time I have seen people waste way too much time on XTU score / problems, so please don't get caught up in that.

    If XTU is working for setting CPU option tuning, that is enough for me.

    Tip: The XTU interface app doesn't need to be running to keep CPU options set. I exit the app/systray after I am done with it. Leaving it running affects performance, just like any other app monitoring the system. I also exit Asus Tweak Tool systray apps after boot.
     
    Last edited: Dec 28, 2014
    eleron911 likes this.
  8. Meaker@Sager

    Meaker@Sager Company Representative

    Reputations:
    8,731
    Messages:
    54,423
    Likes Received:
    15,864
    Trophy Points:
    931
    Exactly with benchmarks and information software it's good to have a couple at least to cross reference.
     
    hmscott likes this.
  9. eleron911

    eleron911 HighSpeedFreak

    Reputations:
    3,886
    Messages:
    11,104
    Likes Received:
    7
    Trophy Points:
    456
    If you mean the performance type for the power state, it's on balanced. That shouldn't have an impact though, since while benchmarking the frequencies stay on par and go up to 3.44, according to the test bench. Which is odd, I did set it on x37 multiplier, so I was expecting a 3.7 Ghz confirmation...
    Yes, 5.1.1.25 also.

    I had no Asus apps installed, it was a fresh boot, since my purchase was with Free-Dos.
    I've repeated the test several times, right after boot, 10 mins after boot, no difference.

    I'll stick to XTU for undervolting purposes and try to do the other benchmarks to see if there is a difference.

    Thanks for all the input so far.
     
    hmscott likes this.
  10. Meaker@Sager

    Meaker@Sager Company Representative

    Reputations:
    8,731
    Messages:
    54,423
    Likes Received:
    15,864
    Trophy Points:
    931
    Try it on high performance. A lot can happen from micro second to microsecond that impacts performance ;)
     
    hmscott likes this.
Loading...
Thread Status:
Not open for further replies.

Share This Page