No announcement yet.

3400MHz 16GB (2x8GB) vs 3200MHz 32GB (4x8GB) optimum performance and requirements

  • Filter
  • Time
  • Show
Clear All
new posts

  • 3400MHz 16GB (2x8GB) vs 3200MHz 32GB (4x8GB) optimum performance and requirements

    After much research into DDR4 ram, including considering all the top manufacturers and at times being close, have come to the conclusion that G.Skill Ripjaws and Trident Z are amongst the very best. I have ended up choosing Trident Z for its optimum quality, also like the looks of best and the top performance, quality,and its the premium range. I have just about decided, am awaiting the Ram on order, worth waiting for and am still awaiting an air cooler that I decided to go for instead of budget to mid range water cooling and am not planning on record over-clocking or heavy gaming, but wanted the best most reliable and fast OC Ram of at least 3000MHz and up to 3400MHz.

    I posted the following to a top enthusiast hardware computing site,and it was suggested by a well experienced quality contributor that I go for 32GB, probably because of the software I have, which is a good choice, but then again they both have their advantages, so thought I may pose the same question here as its totally whether to go for the Trident Zs or the Trident Zs:

    I have all the kit assembled for a new Skylake 6700K build and have left the ram to last as has been rapidly changing and at last become more affordable for very high specs. My question could relate to any make of DDR4 RAMM, but have decide din favour of G.Skill TridentZ as one of the best, although had also come close to going for Corsair Vengeance Pro, but eventually gone for the Trident as in many ways it's on a par to Corsair Dominator platinum, but seems better value.

    My PC setup is primarily as workstation for Microsoft Office suite, including mostly Word, OneNote and Visio, multiple open browsers and open tabs, also working with Dragon Naturally Speaking professional, qualitative data analytic software. I also have music production software, with Cubase, Halion and Steinberg VST instruments, with the largest set Halion Symphonic orchestra that specifies minimum requirements as 4GB Ram Photoshop photo editing Premiere Pro and After Effects, also Sony Vegas Pro for video production hobbies, using the Adobe suite.

    Although this software when used by professionals may require 32 or even 64Gb GB for optimum working, if adding some professional VST libraries can be much more Ram intensive than the standard Steinberg ones I use, or for advanced layers or 4K video rendering, for my usage though, I don’t usually require anywhere near as much capacity, as I am not running a studio. I only use 1080p video and am a beginner with Photoshop with 12 megapixel photos and video compositing it Isn't my main area that I use my PC for, as am not a professional musician or video editor, so it is more for my own rather than productive use.

    After much consideration have come to deciding between 2 DDR4 ram kits, based on optimum quality and affordability:


    Alternatively for £28 more but twice the amount of Ram, but also one step lower in speed and GTZB instead of GTZ so possibly not as tight a Cas latency at the extreme:

    I also noticed the following: F4-3200C16Q-32GTZ
    This is also 32Gb and in a 4 dimm set, GTZ with slightly tighter Cas latency sequence. It also does not seem to be as readily available, so the decision appears to be between the 2 sets above this one.
    I would've also considered the 3400 in 32GB sets, but the much higher price at about double makes it far over budget.

    All are Trident Z series, specifically designed and tested for Skylake. Both specified as premium, so if considered the GTZB wondered if quality as good, but would seem to be except possibly tested speed is not a tight, but build quality as good. Also GTZ in same capacity and speed tend to be slightly more expensive, probably due to the tighter tested speeds. This may limit overclocking above the tested figures possibly. But if over clocking to less than tested maximum speed may mean both GTZ and GTZB may perform close to or even equal , with the main difference being in pushing it to the tested limit.

    My mainboard is an Asus Z170-A, so is Skylake dual channel, with 4 dimms. It has all the features and specifications required, but has a limit, as do most of the boards up to midrange of 4 x DIMM, Max. 64GB, DDR4 3466(O.C.) The mainboard having limit of 3466 (OC) means that the 3400MHz set presumably could be clocked at 3400 or possibly 3466 with a few tweaks to maximise performance on the board. Then again the 3200MHz set would also run well at that speed and possibly could also get closer to the board limit.

    I would like to know if GTZ vs GTZB is also a factor consider or is only related to pushing speeds to limit of tested frequencies and otherwise would perform just as well up to that point, which I think it possibly will do, as both types are premium DDR4 Trident Z series and so are both high quality.

    But an important factor is that as the 16Gb 3400MHz set is still not quite as expensive as the 32GB 3200MHz set, if I only actually require 16Gb I may as well have the absolute optimum in performance at a lower cost as the spare ram capacity would not be utilised. But of course for a relatively small amount more could double Ram capacity, with slightly lower but still very high performance.

    There is another factor, in one set using all 4 Ram DIMM slots on the mainboard, whilst the other uses just two. As long as used in matched pairs and sets that are compatible, and the Trident Z are specifically designed to be used on Skylake boards, so they should perform very well as a full set using all 4 DIMM slots. But from many previous comments on this board and from the internet that I've read on this, it is meant to relate to stress on IMC memory controller. I have also read in one review though, that the Skylake i7 5700K processor is meant to have one of the best and strongest memory controllers yet. This may mean the extra stress may not effect durability of the processor to the same extent as may have done in earlier models, and may not even have any detrimental effect any more at all.

    Very often people also make suggestion of using 2 slots means can add more ram later, but also often correctly emphasised since DDR3 and 4 and especially with higher frequencies and overclocking mixing ram sets is not recommended, as need to be manufactured at same time and selected as a set. This means if want to upgrade capacity or type with later innovations, would need to replace the entire set, so the main factor is on performance.

    I have also read 16Gb modules may not necessarily perform as well as 8Gb ones, but not sure on this, but could mean 4x8 may in some respects be better than 2x16. There seems to be very little information available on-line that compares performance of 4GB, 8Gb and 16Gb individual modules of same frequencies and Cas latencies with one another.

    But also using 4 dimms does mean doubling voltage and possibly adding heat, I seem to have read somewhere possibly by as much as 10 c and if so needs to be considered. The Asus mainboard instructions does mention if using all 4 dimm slots to ensure adequate cooling though, so perhaps it does mean increased heat, but then again that may only be important if there isn't sufficient overall cooling, although the using of all dimms may add to heat that needs to be dissipated. As long using all 4 DIMM slots isn't a major factor in affecting CPU durability though, and the board and processor are designed to work with 4 DIMMS and at double capacity of 32GB for that matter, with 4x16GB up to 64GB, then it may no longer be as important as it used to be, or be important at all in this case. It may mean using all 4 DIMMs in a matched set may even perform just as well as using one matched pair.

    All the software would work with either RAM though. So it’s a series of trade-offs, if I opt for 16GB when 32GB may have been better, although most of the time I expect 16 GB would be sufficient and for my primary uses of the PC that may be optimum using al the current software. However if 16 GB is sufficient for the vast majority of the time on the PC then there isn't really any trade-offs and the highest performing 16GB set would then seem to be the best choice.

    So in summary I'm wanting to decide which set to go for, either the very highest performance 3400MHz GTZ C16 16GB (2x8GB) or the also very high performance 3200MHz GTZB also C16 32GB (4x8GB). If 16GB is sufficient though, for the way I use my PC, then that would be the optimum. But if 32GB were needed, then a slightly lower performance, together with using all 4 DIMM slots, but still with high performance, would then be the best option.

    Any good advice in deciding would be appreciated. Also points mentioned may be of interest to many, enabling various factors to be considered or debated, that could help others decide on how best to set up their PCs too.

    For several days, weeks I stayed with the 3400MHz 16GB order, then as had to wait patiently as both sets have been in short supply and very new, then changed it to 3200MHz 32GB. Now can change back or stay the same, once it arrives I would have to pay extra postage to swap if don't open, once opened shall of course install it and am sure shall be pleased with either choice. If only the 3400MHz 32Gb set was much closer cost to 3200MHz, but its a big difference anyway a few areas of debate and any recommendations appreciated. I shall follow up a post when I receive the order and install and use it, I'm sure I shall be well pleased with either set.
    Last edited by Creative Tiger; 11-20-2015, 01:39 PM.

  • #2
    Sounds like you should consider capacity as priority.


    This kit seems ideal for your situation.


    • #3
      Thanks for such a quick response, and a top quality recommendation.

      I appreciate you taking the trouble. I have noticed G.Skills staff respond quickly on the forum and noticed when researching about my compute build and components on New Egg to people who have any issues you're always quick to offer help to ensure people get the very best. I had a top contributor of a top hardware website also suggest I should go for the larger capacity one and of course after all, they are only one step apart in tested frequency sets. They also are a big fan of G.Skill Ripjaws and Tridents. They also have the same kit right now.

      I suppose my dilemma was built up from researching the topic so well that I went to the finest points. In the real world the difference in performance of both kits will be negligible only detected on test bench under extreme testing. I somehow think both kits well exceed anything that can be thrown at them right now and yes in usage terms having double the Ram has many advantages. It shall mean for all my usages I wont max it out and will give a performance boost to all applications.

      I do also tend to multi task, with multiple browsers, I've about 10, don't always use all at once, but have several open with multi tabs for different Twitter feeds, Hootsuite, Tweetdeck, academe professional, social issues, technology, music video and use several Microsoft desktop Office applications at once, whilst using Dragon Naturally Speaking Professional and whilst using others such as Photoshop, Premiere Pro, After Effects or Vegas Pro and if video editing would want to run any rendering in the background whilst use computer for writing or social media or music video etc.

      The 6700K i7 processor is ideal as hyperthreaded and 4 cores and for my work with these programs is enough, besides the biggest gain is on using 4 cores, also the best settings on the Adobe software is to give 2 to 3 GB per core, but if use say Premiere and After Effects at the same time, that's double, plus use other programs and need to keep a good amount for OS and programs working on whilst these can run at same time.

      Yes can envisage could hit limit using 16GB and 32 GB shall give me all that is required and still at ultra high specifications and probably so close to each-other, as they are after all are just one frequency set apart,and even then variables of having more ram performance advantages in the real world usage compared to less, as well as speed having an effect and both are ultra high quality, excellent timings, hand checked for compatibility as the best are the best.

      It would probably require test bench to detect any difference and because of being 2 different sizes, may make an interesting test bench comparison for that matter as I haven't come across any test bench where 3400MHz 16GB has been compared with 32ooMHz 32GB. I dare say the advantages of having more ram may compensate in real-world scenarios when using programs or multitasking where ram capacity affected performance.

      Also G Skill Ripjaws and Tridents out-perform any programs requirements currently and shall probably be ahead of program applications for quite a while, so in real world use probably just as good as each-other, but with the advantage of having double the quantity of the world's best Ram.

      Its one of those questions where there can be more than one answer, yet all can be good and for different reasons , and depending on how one uses a computer, which for me is probably appropriate in view of the programs I have but still wanting to make the best decision and even then as mentioned above if both these identical kits were compared they may turn out to be even more similar in real world usage, where of course more ram can help.

      So yes thanks for providing a quick answer, its helped to encourage me, to stay with what I've ordered, I am sure I shall be really pleased and impressed by the quality and performance of the Tridents


      • #4
        I didn't read the entire thread, however, currently most motherboards experience XMP issues, the issues are more common with 4 sticks, so I suggest 2 sticks, for the short term

        Long term 4 stick issues might be solved for good, I wouldn't risk it myself


        • #5
          Originally posted by GSKILL TECH View Post
          Sounds like you should consider capacity as priority.


          This kit seems ideal for your situation.
          Thanks for the recommendation of getting 32GB as the combination of programs could mean I can exceed 16GB and whilst probably often use under this and not likely to push the 32GB to its limit, the available ram is useful . There are also speed enhancements when using certain programs as well as especially multitasking.

          To the person above who is also unsure on using 4 dimm slots, as I was originally, on testing this ram kit have found it works perfectly and it pays to get a tested set when using 4 sticks together.

          I am pleased with the ram, excellent quality. I have written a full update to my post on Toms Hardware site that received some comments from a top poster on the site also a G.Skill fan.

          I have completed a moderate overclock on 6700k processor o 4.6GHz overclocked 3700k at 1.328v for cpu manual and installed the ram effortlessly and flawlessly using xmp profile achieving full G.Skill tested advertised frequency and latencies. I then later carried out some tweaking and have managed to get the same profile at 3333 MHz. Can’t get the same profile at 3400 MHz or next step at 3466MHz, but can get it to work if change settings. However, anyone wanting the advertised low latencies and high performance at 3400 MHz would need to get a 3400MHz set. I would’ve done but cost at the moment is a lot higher than 3200MHz. Also my mainboard an Asus Z170-A is close to its maximum, but 3200 is already in category of super Ram and 3333 MHz for a complete 4x8GB ram kit, using all dim slots is absolutely excellent.

          I have dropped Ram frequency back to what appeared stable 3333MHz at same good latency settings 16-18-18-38 at 1.36v I tried setting adaptive mode on cpu voltage from manual at 1.33v that in bios sets to 1.328v and ram at 1.35v but crashed. On changing this to manual 1.33v, shows in bios as 1.328v and Ram set as 1.36v shows varying between 1.360 and 1.376v it works. I also seem to recall that xmp profile for ram set voltage was very slightly above 1.35v, so same as xmp for ram. I reset to cpu manual at 1.33 and bios shows 1.328v and ram set in bios to 1.360, although appears to vary on bios display between 1.360 and 1.376 but working. I then tested resetting to xmp, actual value is 1.3530v and did not realise had to change from auto on ram voltage setting, was trying to change it in the above option, discovered could change in ram voltage in bios from auto to 1.35 and now shows as 1.3530 that matches xmp.

          It worked with same settings as 3200MHz and passed stress tests using 1.36v, however at same settings and 1.35v it crashed. So seems requires 1.36v needed to get same 16-18-18-38. But when tried 17-19-19-41 at 1.35v it worked at 3333 MHz at correct xmp voltage and with 4 dimms all populated opted to keep it to 1.35v at slightly higher cas latency. However, it is hit and miss at 3333 MHz. They do seem to have extremely good xmp profiles at 3200 MHz for anyone simply wanting to know the ram is capable of tested speed.

          Shall have to do more testing to ensure working at 1.35v with slightly higher latency settings but still at 3333MHz.There are no issues in running 3200MHz at xmp settings at 1.35v though. So it’s a trade-off between 3200MHz and 3333MHz at very similar settings and bearing in mind at higher frequency, whatever it is good to match the tested frequencies and to have scope to exceed by timing changes and at respectable settings.

          What this proves is G.Skill Tridents achieve full tested rates and with no errors, flawless and for anyone wanting to go beyond using xmp profiles it is possible to also achieve just beyond tested levels also with the same high quality flawless results.

          So I am pleased as a customer and from these results can recommend getting Trident Z 3200 MHz (OC) GTZB 32GB (4x8GB) and makes a good combination with 6700K overclocked using mid-range Skylake mainboard. I should also consider any of the Trident Z range ram sets in any combination or frequency to fully perform at tested ratings and with enough room to go further for tweakers, over-clockers and extreme over-clockers.
          Last edited by Creative Tiger; 12-17-2015, 11:44 AM.


          • #6
            Those are pretty good results. Don't think there are many faster personal computers.

            Did you run any benchmark tests?

            You posted a lot of good information I'm sure many people will find useful.

            Good stuff- Keep us posted!


            • #7
              Thanks, I am learning and am new to overclocking, in fact my first pc build. After a lot of booting and rebooting and finding myself around trying to optimise my set-up without excessive voltages, yet ultlising above stock ones for performance. I found running ram at rated 1.35v best and to my surprise all settings at xmp and yet at 3333 MHz. I didn't expect to get that stable, thought may have to chose between using xmp that is already excellent or loosening to get 3333, with 3400 being a bridge to far, but I think may be more related to limits of the mainboard plus my desire to limit cpu voltage to 1.35v. I managed to get an incredible 3467MHz stable at 1.35v on the ram kit uisng all 4 dimms at 32GB at 4.6Ghz, but noticed too many voltage spikes up to 1.4v on the cpu. I noticed performance improvements and much lower voltages and very cool running using 4.5GHz and by not trying to push any closer towards 3400 MHz on the ram, but excellent on 3200 and 3333 (OC).

              It gets even better, I think G.Skill may like the benchmarck chart for the ram, its at the very peak on 'PassMark' memory mark. I have done some tests at 4.5 GHz on 6700K with what seem typical voltages for that speed, set at 1.328v with 0.020v negative offset, that allows low voltages when idle or not under stress and seems to maintain core voltages in high 1.2's to low and up to about 1.35v when under stress testing, with maximum temperatures on Noctua NH D15 air cooler with 2 fans not exceeding 750rpm at around 62 degrees C and with the 3200 (OC) ram profile same latency settings of 16-18-18-38 also at 3333MHz.

              After letting it run for a few hours using regular programs, checked in bios and offset cpu core had changed to 1.264 v and still minus offset. Not sure how that happened unless it had taken it down from about 1.340 when I first applied the negative offset. One of the experts on this site may have an explanation, but whatever its running fine at low temps and at 4.5 GHz with ram at 3333 MHz.

              On Prime 95 stress testing after running for 2 hrs at 4.5GHz and Ram at 3333 MHz and passing all tests, whilst using browser and OneNote, plus CPU-Z, Asus Dual Intelligent Processors 5 with 2 input one output case fans set to max cpu usage at 100% cpu temp varies, with peaks on most demanding tests temporarily up to 75 to 81 degree max, with case fans on max, but appears to range between mostly 62 and 64 degrees. As not a game, although if get chance to play in future with this set-up may consider more gaming when get a graphics card (I am going to wait for the Pascal series of graphics cards to come out in 2016 as many new features, geared for 4k and designed to work well with Skylake series and onwards mainboards and processors, to add to rig). When working on media files not usually rendering for over an hour or for hours on end, usually do not use heavily intensive tasks for over an hour, but can use pc for hours with Office, browsers and programs that may use ram but not too intensive on CPU or GPU, So 2hrs of Prime 95 whilst using Office, email web browser and Youtube seems a good enough test, together with incredibly high performing scores on other benchmark tests, including 11 1/2hrs of MemTest 86 seems a thorough test, the set-up has passed. Also as using integrated graphics without a graphics card this also means am using cpu and Ram for all graphics functions, including driving 2 1080p monitors, so possibly when add a graphics card this may help performance even more.

              Results on Cinebench R15 and Passmark (v8.0.1053) are astonishing, with Ram at the very top and incredibly even some graphic tests notably in 2D matching 980 graphics cards on the comparison benchmarks and even exceeding in several tests that are cpu and ram intensive, using the on-board Intel integrated graphics. On the overall rankings it is only beaten by 8 core and 12 core xeons and on programs utilising above 4 cores, but in practise most people would be utilising at most 4 cores, but even then it rates using 4 cores not that far behind 8 and even 12 cores and on 4 or less cores it exceeds them. There could be a combination of a good moderate over-clock on cpu, plus a very good over-clock on the 3200 (OC) run at 3333MHz Trident Z s and utilising 32GB Ram.
              Last edited by Creative Tiger; 12-19-2015, 07:23 PM.


              • #8
                Niiice good to hear. It will be the fastest PC for some time. Like you said even most work station computers with more cores may not perform as well.


                • #9
                  Thanks for the enthusiastic reply. Since my initial query on selecting ram I've been running it smoothly and tweaking a bit more, Trident Z is running amazing. I am also pleased ive gone for the 32Gb ram, as I may add using virtual machines to tasks needed as well as using many programs that are helped by having plenty of ram. I thought as it was running well as the better than expected settings, I would try a bit more tweaking in case it may be possible to reach 3400MHz. from trial and error, always maintaining 1.35v for the ram and aiming for a good moderate over-clock on cpu at 4.5Ghz. I achieved best stability with cpu set to 1.35v offset to allow for boost up to 1.37 and at 3466 MHz manual over clock aiming to limit to max peaks within 1.42v. I find mostly it runs at between 0.8 and 1.35v for the cpu. The ram at a steady 1.35v.

                  Note all tests and usage is using only the Intel HD530 CPU integrated graphics without any graphics card and running 2 monitors at 1080p, as I want to wait for a Pascal video card that will be better optimised for the features of K series Skylake processors along with G.Skill Trident Z super performance Ram. With discrete graphics card meaning much less demand on the processor for graphics it may be possible to exceed all benchmarks, although the fact that this processor has integrated graphics equivalent to what was a top end graphics card from only a few years ago, but as programs can use integrated features as well as discreet graphics or for integrating processor and graphics card, the high quality of graphics even without graphics card is incredible. But this means with say an Nvidia Pascal graphics card when it is released the benchmarks are likely to all be surpassed, including tweaking to improve latencies and lower voltages as well as for maximum workable clock rate.

                  The latency settings of 16-18-18-38 have worked steady at 3333MHz. I then tried various slightly loser settings to try for 3400MHz and found it has reached a steady usage setting at 17-18-18-38 at 3466 MHz. Funnily i originally tried loser settings that seemed to go down. These settings are close to the 3600MHz kit latencies but using the 3200 MHz kit and achieving at similar latencies 3466MHz. I shall have to see how this runs under stress. The 3333MHz passes all the stress tests and for the last week it has been running smoothly in normal use at 3466MHz at 17-18-18-38 CL. That is also at the limit of the main board and a significant over-clock far exceeding any program specifications.

                  I also wonder if losing but on same ratios may achieve same high speed but can set cpu voltage a bit lower, possibly at similar levels to 3333MHz at 16-18-18-38, maybe at 18-18-18-38 or 18-19-19-38. I'm not an expert on this but am learning as going along. I may also take it back down to 3333 MHz if need to maintain cpu at slightly lower temperatures as at those settings was finding that at 3333MHz, that the cpu could run at 1.29 to 1.35v, mostly 0.8 to 1.32v and in reality the difference for real world usage of 3333 and 3466 MHz must be very small, both very fast.

                  Update: Update: After running successfully for couple of weeks at 3333MHz set at CL 16-18-18-38 on 32GB set of 4 ram dimms at 4.5GHz, using only the integrated graphics as waiting for Pascal before adding graphics card, thought I would lower to 4.4Ghz until add graphics card then may take it back to 4.5GHz. All tests have been with BLCK at 100. I also noticed on offset voltages running higher than liked, and setting at 4.4Ghz didn’t lower voltages with settings as they were. Also using Asus on-board 5-way set-up it seemed to also keep it running higher. Manual lowered the peak rates to the rate set and found at 4.4GHz could run at 2.9 v., but also never went down. Then discovered adaptive seems best for my purposes as its mostly much lower voltages, very low temperatures, extremely quiet and still allows peaks when demand increases. Found it crashed a few times at 2.8v, but was stable at 3.0v but upped it to 3.2v to allow for higher demands and found it mostly ran near 0.784v and still did so when set at 3.2v, just that allowed more room for heavy demand programs.

                  With it running smoothly thought would have a go upping ram towards mainboard tested maximum and found it has been stable set at 3466MHz, 1.35v for ram, with CL at 17-18-18-38. This is a fantastic result from 3200MHz kit and I expect similarity high performance surpassing advertised tested benchmarks for all in the range. On my main board this maxes out the board and for anyone with boards that go higher similar maxing out may be achieved on even faster ram kits, but for me this means I have achieved full maximum with a 4 dimm 32Gb kit. Although again as speeds are so fast they exceed any software currently available, but possibly in the future means its future proofed to keep any system running to work with any software for an extended period before any future upgrades. its also a very good reason to use any of these ram kits.

                  G.Skill Trident Z is excellent ram, very impressive!
                  Last edited by Creative Tiger; 01-21-2016, 05:00 AM.


                  • #10
                    Can some one from GSK can post the STRAP and BLCK of each of this ram kits above 3200Mhz I know 3200 works /w 100S & 100Bclk, but what about 3333, 3400, and so on, are any other ram frequency above 3200 that works w/ 100s and 100blck and listen to my ?. Above 3200 able to work with 100strap and 100 clck XMP. And I said XMP cause then its certify from factory that works at those setting. Yes I know that anything above 3200 you could tweak it to work with 100 and 100 but you have to do a lot of tweaking but if any other frequency above 3200 can work with 100 and 100 XMP certify I will like to know, so I know that is unto my IMC's cpu to get it to work. So far my CPU is able to handle a 128GB at 2800Mz kit that I own and a 3200Mz 16GB kit as well, so I want to find out what any other kit work w 100 and 100 XMP so I can try, Like I said above 3200, TIA.
                    Case:Corsair, Obsidian 650D
                    MB: Asus X99-Deluxe, Bios-1801
                    CPU: i7 5960X @ 4.5Ghz-1.265V, Cache:4.0Ghz @1.255V
                    CpuCooler: EK Supremacy WB with Airplex Modularity Radiator.
                    GPU:EVGA 980HC @1565Mhz
                    PSU:EVGA 1300W SuperNova G2
                    Memory:F4-3200C16Q-16GRK @ XMP settings, Testing:F4-2800C15Q2-128GRKD
                    Timing: 16-16-16-36-1T
                    BD: Intel 400GB NVMe Pci-e SSD
                    OS: Windows 8.1 Pro, UEFI Boot mode
                    " I'm not an OCer, just looking 4 a good mentor "


                    • #11
                      Try kits listed on our QVL, those models are guaranteed to work.


                      • #12
                        Highly recommended Ram kit

                        I thought I'd post screen shots, as in your post you request people to. They relate to previous posts I wrote in this thread.

                        I have overclocked the Intel i7 6700K processor at 4.4GHz with 100 BCLK, adaptive mode. I found setting it at 1.30v works but 1.32v allowed stable operating for more demanding tasks and mostly at very low voltages such as 0.795v. I had a close call using offset, as there was an error when using Asus software tweak, immediately corrected and was not my fault, as I had only set it to 0.01 and ensured I had, but has put me off using offset. I also found offset tended to use higher than required voltages, as did the Asus software set-up. Manual allows full control but meant to ensure stable , always running more than required. Adaptive is proving very good, mostly running very low voltages and very low temperatures, with Noctua NH D15 air cooler maximum temperate using an Intel CPU and ram stress test, 63 degrees C. with Ram set at 3466MHz (Intel XTU Benchmark on Asus Dual Intelligent Processors 5 software). Using Intel XTU features in following video

                        As I am using my desktop pc just with integrated graphics until Pascal graphics cards are released, didn’t want to overstretch the processor as all graphics are currently being run from the CPU only, I have therefore changed the overclock to 4.4 GHz and when add a graphics card may up it to 4.5Ghz. I have also added Intel Tuning Replacement Plan via Intel website as discovered overclocking Is not included in standard 3-year warranty terms, but that allows for any major error that may occur including not user error, so no matter how careful one is, software, hardware UEFI bios etc, that may damage the processor. Of course when one gets it running well don’t want to change processors, just as well to be covered, so can just enjoy the performance of the processor with ram.

                        The Asus version of CPUZ shows actual ram frequency, the CPUZ latest version shows the rating for each channel so has to be multiplied by 2x to get ram frequency.

                        I found the 3200MHz 32GB (4x8GB) kit F4-3200C16Q-32GTZB performed perfectly in all cases set at 1.35v, at 3200 MHZ CL 16-18-18-38 and 3333 MHz CL 16-18-18-38 and these screen shots all show it running at 3466 MHz CL 17-18-18-38 and this appears to be running stable for all usual computing.

                        This seems stable for normal use and I have used lighter tests in this case, but it has passed the Asus on-board benchmark and the CPUZ Intel benchmarks including stress tests at this setting.

                        I must give this ram kit a 10/10 meeting and surpassing all advertised benchmarks and can highly recommend G Skill Trident Z.
                        Attached Files
                        Last edited by Creative Tiger; 01-21-2016, 09:50 AM. Reason: typing error, link html to Intel replacement plan and the G Skill Trident Z exact kit


                        • #13
                          Great information, this is a very popular combination of hardware so I'm sure many people will find this info useful.

                          Thanks for your post and experience

                          Keep us posted with any overclocking results!