GTBTK

Sorry to tell you that your userbenchmark score, while not bad, is also not great either. I don't consider my 1070 "golden" but here is my userbenchmark score with an i7-2600 on a z68 Motherboard. http://www.userbenchmark.com/UserRun/1929485

no worries

I would only worry about that if you actually have a problem with the zotac. Have you tried readjusting the screws holding your existing card in place? The 2 screws in the top should hold the the end of the card steady. I have not seen that bracket before and I dont use one with my MSI card.

use a a custom fan curve. If you set the fans to 100% at 60 deg you will never have to worry about temperatures assuming that there is some airflow in your case

xtreme is a good card and highly overclocked on both the core clock and memory clock. Not much headroom for extra overclocking. If you are planning to overclock memory further, don't pay attention to the "I got +700 on my memory clocks" statements from other 1070 users because their memory is starting about 200mhz below yours to begin with. Think in terms of actual frequencies. The thing to watch is that because of the high overclock, it needs a high binned GPU to not crash. On some of the cards, the binning process gigabyte uses allows the odd lower binned chip to slip through as binning is not perfect so if you experience many crashes at stock speeds, RMA the card and get a new one.

OK fair enough, with the 4G vs 8g RX480, In some game titles there is no difference in performance and the bigger cooler will be beneficial particularly if you overclock the card but in others, the extra memory certainly gives you better frame rates at stock Ram speeds so that may offset the cooling and overclocking benefits that the better cooler has. If you go 8Gb, I would suggest that you just keep in mind the temps and case ventilation, particularly if you are in the southern hemisphere and coming into summer now. I wouldn't go out and spend on extra fans on day one but experiment first and find out if your installation actually has any high temperatures that causes the card to throttle significantly first. You may find extra case fans end up being beneficial by feeding the card more cool air.

rx480s run quite warm. The the hotter they get, the more they down clock to manage temps. The better cooler will give you better performance. I would get the card with the best custom cooler and avoid the blower style reference units.

Without the drive cages being in the way, you have loads of room. the Zotac is a beast. Just make sure you screw it in properly to avoid droop and putting undue load on the PCIE slot because is is quite a heavy card

memory frequency is more a product of motherboard than the CPU itself. I am not familiar with the gigibyte uefi on the motherboard you have. Most of teh kingston memory is plug and play so the motherboard will set the correct speed and timings for you. I am running the blue version of the 1600Mhz ram you linked to in my rig. As they are tye same price, I would suggest that you get the 1866, It should run fine in your motherboard and give you a littlle extra performance. As it is the same price, it makes the 1866 even more compelling. Just remember that If you buy a new/more updated motherboard in the future, Anything Skylake and newer will need DDR4 Ram and that is not compatible with the DDR3 ram that you are ordering here so I would not be thinking in terms of future proofing unless you plan on buying second hand Haswell components in the future. Make sure you measure the available space in your case. The Zotac is rather large both in length and width. The Asus is longer than say the MSI cards but the MSI cards are higher. The Strix is a good card and you wont go wrong if you end up choosing it. 650W gold power supply is plenty for even two 1070s in SLI. Strix needs one 8 pin PCIE power cable. Zotac needs 2 x 8pin and MSI needs 1x6pin + 1x8pin. The pascal cards typically draw less power then the older cards. Your 770 uses a 6+8 pin power solution I think so you should be fine using existing power connections.

The first two bioses that you mentioned are exactly the same with the exception of the Asus branding as opposed to Nvidia Branding Yes you could flash the 80.10.3E.00.01 OC bios, though you would want to have a case with fantastic ventilation or water cool the card as it disables GPU Boost and sets the power limit to 300W

GCN architecture of AMD cards has been designed to specifically support Async compute and do it in hardware. AMD developed Mantle to support it back in 2011 and it was exclusive to AMD hardware. Mantle is now dead but it did provide the foundations for Vulkan. Before Vulkan and DX12, there was no API that nvidia could use that made use of Async compute so the hardware was not designed for it. Even now, they are using software schedulers to provide that functionality on pascal cards that AMD is doing in silicon. Nvidia cards have been hardware optimized much more for DX11 non async applications. That is why we are seeing huge performance jumps with AMD 390 and Fury cards cards in Vulkan and well written DX12 titles and Nvidia is not really improving at all

don The maxwell cards are excellent cards. The only thing against them now in December 2016 is that they are 2014 technology that has been made obsolete by pascal. DX12 didn't exist when they were introduced to the market so what is going to become more important from this point of time forward was not included in the design of the card that will require more and more compromises over time. It happens. It it the very nature of technology. Time moves on and new tech emerges. If you have a 980, and are prepared to compromise game settings down from ultra and can run at dx11, there is probably no driving force to upgrade to a 1060. That being said, I cannot see the value in investing now in an obsolete technology that is already starting to ask for compromises when for teh same price, you can get something that doesn't ask for compromises at all.

It does but Maxwell async compute performance is no different than it is without enabling the setting. Async compute improves pascal and and to an even greater extent AMD Polaris performance

Kingston HyperX Fury RAM is good quality, overclocks well if you end up upgrading to a z97 motherboard at some stage and is typically better priced than than Corsair ram for similar or better specs. The i5 you have will bottleneck the card slightly but it will probably still be in the good enough category. Asus strix cards are good, MSI Gaming cards are slightly quieter and have higher power limits than the ASUS card. The Zotac Amp extreme is a beast with the highest power limit of any 1070 card. Having said that all 1070s can overclock to about the same levels but the ability of the cooler and power limits are the features that will allow the higher clock speeds to remain more stable If you are on a tight budget you may also look at the MSI Armor series. Cheaper and slightly lower power limit than the gaming cards but the cooling is similar

less headaches with a single card

the 980 does not support DX12 and async compute as well as either of the other two. That will impact the longevity of your purchase

I know that Fermi and Kepler cards were, in general voltage locked unless you did a bios tweak. not sure about MSI cards in afterburner. You are probably better to ask that question at guru3d.com in the afterburner support forum. The user Unwinder over there is the developer of AB

Sorry, you quoted someone out of context. At the time that statement was made, The general assumption, particularly being reinforced by the "Cover your A**" throw thermal pad reaction at the problem by EVGA, was the same assumption you are making now, that the VRM were running way out of spec and too hot because of the design of the card. Subsequent to that statement being made they did actually do some analytical testing that showed that while the VRM cards were running towards the higher end of their rated temperature range, they were, in fact still within the specifications for those components. If a Supplementary component that is connected to the VRM parts blows up after 3 months because the temps are in the higher range of the specified range of operating temperatures, there is a problem with the supplementary component, not the VRM design itself or the cooling design because teh VRM components are still running within their specified range. I also note that FTW cards have not all universally exploded which would seem to indicate that the capacitors in some cards can actually cope in that operating environment. The higher range of temps may certainly have accelerated failure of the faulty capacitors but that faulty component was doomed the moment that it was installed in the factory. Manufacturers cannot test every card at max load for 3 months before they ship them, They would never be able to sell anything. They test samples to see that they function properly when they initially power up and rely on the Quality control of their component suppliers and their warranty to cover any other faults that slip past quality control which is exactly what EVGA are doing.

That was in reference to the overheating "issue" and thermal pads that was being blamed at the time and since been show to be a non issue anyway. Not in relationship to the quality of a batch of capacitors that , when working within spec, are not causing a problem.

Have you tried opening nvidia control panel and going into the setup multiple displays section?

The card will only clock up to the maximum if there is load on it that requires the performance, all the while the 1070 will be managing things to save power. LOL is very light when it comes to GPU requirements so I suspect that it is not putting enough load on the card for it to ramp up.

for future reference, the MSI also has a memory ok type button that will try and find lowest common denominator settings so it will boot with some types of high speed memory

I'm not there to look obviously and i dont have an Asus Z170 to test it on but It still seems that something in the bios is conflicting to let those voltages hit 1.4v. Those high voltages will impact the life of the CPU, particularly if the cooling is not designed to cope with that sort of heat load. I would suggest that you think about resetting the bios to optimized defaults. Just make sure you check the sata interface settings and make sure they are the same as what you are running now. On my Z68 board, the sata setting is reset to raid mode and disables the hard drives if you reset to defaults so I have to change it back to AHCI and re-enable the drives before the PC will boot. I just did a quick search and found this. it is just a run through to show you what is in the bios. I am sure there are other vids around that will give you a step by step guide to set up the bios for overclocking etc. Asus does have step by step automatic OC utilities you can use if you want but that may have been the cause of the high voltages in the first place. Worth taking a look and seeing what happens, you can always reset everything if id doesnt work out well.. To set the cpu voltages settings manually, you need to run the UEFI settings in advance mode (press F7 at the uefi front page) a and go into the AI tweaker section.

I know that feeling

1.4V is very high. I guess the vcore bios setting is currently "auto"? At that voltage, the CPU should be under an AIO rather than a 5 yo air cooler I Suggest that you manually set vcore to 1.25V in the bios and your temps should end up being much more under control.