Jump to content


  • Content count

  • Joined

  • Last visited


This user doesn't have any awards

About MageTank

  • Title
    Fully Stable
  • Birthday October 27

Contact Methods

  • Battle.net

Profile Information

  • Gender
  • Location
    United States, Ohio
  • Interests
    Gaming, Computer Hardware
  • Occupation
    Slim Jim Enthusiast


  • CPU
    Core i7 8700k 5.4ghz Cinebench Stable (best kind of stable)
  • Motherboard
    ASRock Z370 Fatality K6
  • RAM
    32GB (2x16GB) G Skill Ripjaws V 3200mhz C14 (Overclocked to 3600mhz C14-14-14-28-CR2)
  • GPU
    EVGA GTX 1080 Ti Hybrid FTW3
  • Case
    Cooler Master MasterCase Pro 5
  • Storage
    Samsung 960 Evo M.2 500GB
  • PSU
    EVGA 650W Supernova G2
  • Display(s)
    Dell S2417DG 165hz G-Sync TN
  • Cooling
    EVGA CLC 280 AIO
  • Keyboard
    Logitech G810 Orion Spectrum
  • Mouse
    Logitech G403
  • Sound
    Sennheiser Game One
  • Operating System
    Windows 10 Pro

Recent Profile Visitors

7,267 profile views
  1. This is going to be a tech support nightmare if true. It's bad enough walking people through their system when they accidentally enable tablet mode on their PC, so if they were to enable this, it's not going to be a good time...
  2. DDR3 in DDR4 Slot

    That would be the VCCIO/SA voltages I referred to. I thought you were referring to VDIMM when you told him "different voltages", because VDIMM is the only advertised voltage on memory.
  3. DDR3 in DDR4 Slot

    Let's completely ignore the IMC limitations and massive difference inner timings between the two, and focus strictly on electrical signals and physical pinouts of the two types of DIMM's. DDR3: DDR4: Ignoring the fact that you would somehow need to account for the lack of those missing 48 pins, you would also need to completely reassign the function of the pins themselves, and alter the trace topology of the DIMM's to account for that change in signal. You would need to first get by this physical limitation before you could even address the aforementioned issues with the IMC and vastly different timing configurations. The CPU itself would likely be fine (assuming you are referring to VDIMM and not VCCIO/SA), but the board might not have a good time, lol.
  4. I ordered a graphics card but...

    My brain actually saw "under" in the first place. Funny how that works, lol.
  5. I ordered a graphics card but...

    That is correct, there are a few GTX 1050 Ti models that use 6 pin auxiliary power: https://www.gigabyte.com/Graphics-Card/GV-N105TG1-GAMING-4GD#kf https://www.msi.com/Graphics-card/GeForce-GTX-1050-Ti-GAMING-X-4G/Specification https://www.asus.com/us/Graphics-Cards/ROG-STRIX-GTX1050TI-O4G-GAMING/specifications/ https://www.evga.com/products/product.aspx?pn=04G-P4-6258-KR Almost every manufacturer has an "overclock" edition 1050 Ti that has the 6 pin connectors. They honestly don't make much of a difference for overclocks from my personal testing, but I suppose they might help for some when it comes to clock stability.
  6. It will work fine, but I would highly recommend liquid electrical tape. Easy to peel off, and doesn't require acetone to remove, just fingernails. https://www.lowes.com/pd/Gardner-Bender-1-in-x-1-ft-General-Duty-Electrical-Tape/3385480 You can get it from basically any hardware store, and one jar will last you for years if stored properly.
  7. That would make a lot more sense, but even that seems unlikely, given the 950m was about 10% faster than a desktop GTX 750, and the Ryzen 2400G needs overclocking to outpace the GTX 750, not to mention it likes fast memory when doing so. Unless that board supports high speed SO-DIMM's, I just don't see the numbers adding up.
  8. Yeah... No. The Ryzen 2400G with 11 CU's was barely on par with a GTX 750. This having 8 CU's isn't going to be anywhere close to a GTX 950. While it is certainly a very interesting unit, I find that specific piece of information to be extremely deceptive. Now, if this thing used a cutdown version of the Vega that's inside Hades Canyon (which out-performs a GTX 1050 Ti) I'd be more inclined to believe it, but that's extremely doubtful, given AMD has already published the information on these chips: https://www.amd.com/en/products/specifications/embedded/8216+8206+8201+8196 Hopefully this doesn't turn into one of those cleverly marketed kickstarter scams that we've been seeing time and time again.
  9. Alienware 15R4 overheatting

    He already tried liquid metal and heavy undervolting. Liquid metal won't solve the lack of thermal mass on the heatsink, and it certainly won't help with the poor design of sharing the heatsink with the GPU. He is already throttling under a CPU only test, a combined test would only make it worse. From what I was able to read on the NBR forums, there also appears to be an issue with the heatsink surface being uneven, along with poor mounting pressure in general for the CPU. Polishing it won't help fix an uneven surface, he would need to thoroughly lap it, and trying to make a laptop heatsink flat by hand is no easy task, even with experience. The sad part is, this isn't even a new issue. They've been having issues with this heatsink design for years now. Here is a thread full of people with similar issues on the Kaby Lake version: http://forum.notebookreview.com/threads/how-to-fix-aw17r4-15r3-13r3-cpu-core-temperature-differential-issue.805062/ As tempting as it is to use liquid metal in these situations, liquid metal won't fix poor mounting pressure and uneven surfaces. You'd be far better off with a very viscous paste like Phobya Nanogrease Extreme, which performs better than Kryonaut on laptops in my personal tests, where mounting pressure is a concern. In cases with good mounting pressure, they are practically dead even in performance. @Dalian I still wouldn't bother trying to make the 15R4 perform better, as it's a losing battle. Many others have tried, but from what I've seen in the past, none have succeeded. This is before you factor in yours having 2 additional cores on top of that.
  10. Alienware 15R4 overheatting

    For the price you paid, you can find Clevo laptops with socketable desktop CPU's and MXM modules that would not only out-perform those "i9" CPU's, you'd also be able to upgrade them/replace them in the future if need be. I'd look at the P870TM-G. Even if you don't get the vapor chamber, the single GPU solution is still superior to that Alienware from a design standpoint: Do be warned, you will not find desktop level performance in a thin and light system. You make a sacrifice in portability, in return for desktop-like performance.
  11. Alienware 15R4 overheatting

    The CPU might be capable of overclocking to 5ghz, that does not mean that every chassis will allow for such an overclock. The same is said in desktop environments. Your thermal solution is a big part of that equation, and ultimately dictates how high you will be able to push without throttling. Sadly, no matter what you do, the design of this particular laptop will not yield promising results from a thermal perspective. Take a look at the design of the heatsink: You have the CPU heatsink that has a faction of the mass that the GPU has, and both of them use a joined unibody heatsink. If that GPU gets hot, it's going to dump additional heat towards the CPU no matter what you do. Liquid metal won't solve this either, because you are dealing with a thermal mass issue on top of a dissipation issue, as those fans are not going to be enough to properly dissipate that heat. Liquid metal is fantastic for solving the bottleneck of thermal transfer from the die to the heatsink, but if the heatsink and fans are the bottleneck, it won't perform a miracle. I honestly cannot think of a single laptop with a thermal solution that can handle these i9 mobile CPU's on top of a GTX 1080. Maybe the Clevo P870's unibody vapor chamber, but that's still pushing it. If it makes you feel any better, you are certainly not alone: http://forum.notebookreview.com/threads/i-have-returned-my-alienware-17-r5-i9-8950hk-gtx-1080-but-still-like-it.816635/
  12. nvidia gtc keynote

    Next you are going to tell us he is wearing the same leather jacket in every presentation...
  13. overclocking memory issue

    tFAW should be 4x tRRD, and shouldn't have that big of an impact on stability as long as you adhere to that rule. Any speed is technically possible, but you need to figure out what your IMC likes, along with what the board can handle. Again, the board has that 2666mhz rating listed, but it's difficult to know what is required to achieve that, as there are thousands of timing/clock combinations one could try. If you want to try 2400, I would not go any looser than 11-13-13-26-2. If you have to loosen primary timings any more than that to make 2400 stable, it will be worse than your 2133 C10 setup.
  14. overclocking memory issue

    2133 is perfectly fine. Also, a lower tRFC is actually better from a performance standpoint. I am glad that you were able to overcome the 1600mhz barrier. If you have the time to tinker, you might be able to see if 2133 C9-11-11-22-1 boots. That would be far superior to most 2400 and 2600mhz XMP's, and should result in a nice performance uplift.
  15. overclocking memory issue

    I am not questioning the safety of the utility, I am questioning overclocking ram from within the OS, as you risk corrupting your OS install with failed memory overclocks. There is very little you can do to force anything beyond 1600 if it's not taking it. Multiple factors are likely limiting you. If you are thinking about using AfuDOS to flash a custom bios with the VTT flags revealed, I would advise against this unless you have a cold flasher or EEPROM programmer to get yourself out of a sticky situation if it goes south. If it were me, I would keep the 1600, and tighten the timings as much as physically possible. Since you have a highly binned kit, you should be able to tighten it quite a bit. You can also try to tinker with your tertiary timings, or try using a single DIMM at a time to see if it lets you push any higher. If a single DIMM allows you to bypass the 1600mhz limitation, you know that it could be an issue with specific channel timings. If your sticks are multi-rank (which most 8GB DIMM's are for DDR3), you can adjust the multi-rank tertiary timings to be looser as well, to relieve some of the stress off the IMC.