toastybread

usually rule of thumb with next generation GPUs is look at the high end and mid range. the higher end will likely be what will be of the mid range.. so typically and ideally with competition driving things forward.. GTX 1080 (or even 1080 Ti if competition was fierce) would be mid range next generation.. that's typical rule of thumb, however Nvidia will likely stagnate a bit, they are a company, and they mostly in the business of making money, without competition pushing them forward... what else is there to expect.. so 1080Ti will likely be a solid pick for next 2 generations assuming playing field is identical to what it is now.. technically speaking, the GTX 1080 is a mid range die cut for pascal already.. the 1080Ti is slightly above mid range cut. just in model numbers/part numbers of their dies it basically indicates that...so we are already in stagnation with Nvidia.. and from looking back we been in stagnation for years/generations.. we gamers need fierce competition in GPU market, and badly.. AMD really needs figure their stuff out and come out swinging harder...

i noticed it in a lot of newer games at higher native resolutions when they used mixed resolution assets. i can't speak on what causes it, and what fixes it.. but GTA 5 an example people bring up a lot does suffer heavily from this because of mixed resolution asets im guessing.. game that is just straight up horrible about it however for me is Ghost Recon Wildlands, and yes GRW is brutal on GPUs, even n Nvidia's and it is a gameworks title, but it doesn't seem to be a GPU load issue.. does weird flickering, shimmering, jaggy edges and messy scaling on both AMD and Nvidia cards when sampled outside of native resolutions specifically and no AA that does a good job and temporal AA method used in the game is just shameful..

if there is positive ROI it is effectively free.. i cannot believe that is a difficult concept to wrap one's head around. i mean yea welcome to the real world, have to pay to play, but if you are earning more back on a daily basis than you are putting in, and have effective zero net loss...where is the argument that it is not "effectively free" . if there was a hard setup process and it needed constant maintaining, then ya maybe i would be swayed the other way with this argument.. but for most serious minors especially, the process takes of getting multiple GPUs to mine sync with each other on a rig takes an hour at most and usually one time process.. sure avg joe from windows only experience may struggle a bit at first.. but once its up and running, that's all has to be done for life time of the rig unless wish to add more hardware.. most things can be checked remotely and even edited remotely.. also as an update, the guy the originally purchased one of my RX480's for $450 during that ETH spike, called him back and he going to help me get my current RX480 running as a mining rig he offered to help, make sure i optimize it correctly, i mean i know what i am doing but he does it for a living so im going let him tinker with it. and i have 2KW solar panel array in my barn and some deep cycle batteries, and several selections of boards and CPUs laying around doing nothing.. ..so it should be good for off the grid mining rig.. he said with avg cost of power here and RX480 optimized right, should be making a couple dollars per day counting losses, with no energy bill to run it and winter fast approaching, it probably be a dollar or so more per day maybe..

you dont want to put another gpu on a failing psu either way...

not hot enough to be thermal shut off limit.. i would grab a new Power supply try it out, doesn't work you can take it back and get money back..

i had a few EVGA supplies fail specifically and usually when i contact support they give me a free upgrade once i submit my claim...i never had my PC get damaged from any reputable brands failing.. and they are always good about getting it fixed..

i suspect its the power supply... is the CPU overclocked? and how much? how hot does the GPU run under idle and full load?

hard to really say.. could be the power supply as well.. may have triggered over current protection. typically a dying GPU exhibits artifacts at stock settings.. usually discoloration/lines and weird shapes.. a PC just shutting off under load tends to lead me to think maybe power supply tripped over current protection or something got too hot and kicked the PC off.. what power supply do you have, what are the specs of your system? and age of it..

it depends on the graphics card generation and the brand since AMD and nvidia do handle it a bit differently and different generations handle it differently as well.. there is no one answer, and both do alot of the same thing. however that being said, most stable and easiest way to reduce power usage and cut thermals is to reduce power target. the most effective way is to reduce boost clocks.. messing with voltages may result in stability loss if you are unfamiliar with it.

but i rather put one of those 3 brands in my system.. way i look at it is that the PSU is the front line to the system stability, and protection from not only crappy dirty power but also from itself if something shorts out inside.. i can also add delta power supplies to that list as well.. that's just me..i would rather janky ass PSU from a reputable brand, over a 50/50 shot in the dark PSU from a company that i'm not all that familiar with, but heard of.

ya, thinking about it, i got my eyes on a 980Ti i seen a few in upper $200 range and a few locally that sell fast for $250.. still little steep, but $250 for what is basically a 1070..that's not too bad actually..

400-450 watt quality power supply is totally fine for your rig... but has to be a quality power supply from a reputable brand.. like i would go with corsair or EVGA or cooler master.

i personally am 50/50 on this subject... i do want cryptocurrency to be a thing, it has really good strengths and works on its own terms that are not influenced by economic shifts.. however, it is very annoying landscape for pc hardware right now because of these annoying miners.. Z97/X79 sandy bridge/ivy bridge boards are like $200-$500 ??? some of these high speed 8 core xeons like E5-2657W (i think is the number, but prob wrong) used to cost like $80 on ebay used are now $180+ used.. GTX 980 Ti used to cost spring of 2017 around $175-$180 used, now averaging out to $300+ used. RX480 i sold one of my 2 i owned and was using in crossfire.. paid $170 for it, 8GB XFX GTR model.. sold it for $450 locally to someone..he admitted to being an ETH miner and told me hit him up if i get my hands on anymore.. i was tempted to sell him my other RX480, but i backed out since this xeon doesn't have an iGPU. derp move on my part since ETH since has settled out a bit..

maybe you need to define what "free" means to you. but for most people Free is basically getting more back than you put in..or not having to pay for it even if it costs money. i call this "effectively free" when your ROI is positive.... the major difference between this, and lets say a normal business, is have effectively zero man-hours. you just turn it up and start making money.. most places on earth, you can make between a few cents to a few dollars even on a single GPU after factor in power usage.. if you have positive ROI day after day, then its "effectively free". because your money that is being made can afford the daily power usage as well as afford you a little bit of money on the side. you dont calculate GPU cost day to day, that's a long term cost.... so that's a dumb argument if you start that one up, and proves that you dont understand the business of mining

of course is speculation and educated guessing until then. i think main reason Nvidia would launch a 1070Ti is the V56 market.. and right now, doesn't even exist really. if V56 took off, and was available openly and affordable enough i dont even think would be much time for rumors to have been stirred up, it would be out already. then have other side of speculation... sometimes when NV cuts down something from higher end product, they accidentally make it faster than product its cut from in gaming... not sure how it happens exactly i assume just less overhead and leakage and etc.. so it maybe possible a 1070 Ti being cut down from a 1080 could accidentally ended up faster in gaming than the 1080. if that's the case NV maybe delaying it and working out some way to limit it from making the 1080 pointless..