Anyone see any reviews where they mine on the new cards, be interesting to see if we are about to see another mining boom and cards being gobbled up....
Anyone see any reviews where they mine on the new cards, be interesting to see if we are about to see another mining boom and cards being gobbled up....
I really hope not, such an awful waste of electricity nevermind the dodgy ethics that crypto is inheritantly associated with.
Nothing official, but there is this:
https://videocardz.com/newz/nvidia-g...ng-performance
Cant see miners going for that, still, prices seem to be going up..
I was looking at just ordering a 3080 and a 3070, testing both, seeing which I needed and reselling, think I'll try for a 3080 and then wait for BigNavi and see what that does, got an Apple Watch to buy now as well as possibly a new phone, spare funds are running out lol
You'll probably be waiting a while.
Yeah looks like it, NVidia have really screwed us over here, created a load of hype and then delivered sod all by the looks of it, just left it all to the AIB guys, who have then, mostly, started at £700...
See what BigNavi brings and if that ends up being an actual launch rather than a fake one, if it does, NVidia have just shot themselves in the foot IMHO...
So is that not good then? As someone who knows nothing of mining i was just wondering how accurate the rumour of 3x mining performance was.
Given what the cards are going to cost I cant see it, I'll drop on Nicehash and see if they're already in the calculator and see what that says..
Ok, so looks like 3080's are on the calculator...
https://www.nicehash.com/profitabili...ia-rtx-2080-ti
The low price of the FE cards just makes anything else look expensive, you'd expect to pay more for a card from NVidia as its their stuff, in the same way as MS do with the Surface line, here guys, look when you can have if you pay a premium, then lets the AIB's make clones etc at a lower price point, Ampere has gone completely the other direction, it must be by design, maybe they only dropped small numbers because they wanted to get out of the gate before AMD's BigNavi..
If AMD can ace the launch, have them at a decent price and have good availability then it really could be AMD's time...
It's all a bit strange. Okay, the main reason for going with Samsung was the legendary Nvidia greed (after all JHH had a O'Leary-style rant about how the cost of transistors no longer decreases with new nodes).
However, the one advantage for consumers of going with Samsung should have been availability as it certainly wasn't efficiency as perf/watt hasn't really moved at all despite a node jump and so far all the reviewers have said there's next to no overclocking headroom (ergo, Nvidia are running these close the max out the gate).
So either Nvidia have rushed this release before they had built up stock, and/or the yields aren't that good. The dies aren't small at around 628mm² according to the TechpowerUp database and from the TDP they are running them way past their speed spot.
If they have lots of chips which can't hit the 3080 mark, they might release a 3070 Ti with them later.
I would imagine there’s a deal in place on Samsung ram chips if they get the gpu done with Samsung also, however Samsung’s track records not brilliant, so maybe this is less about nvidia and more Samsung..
Or is it really nvidia greed, they’re getting a huge discount for using Samsung so nvidia margins are good regardless of what the heat sinks cost...
Well, what I meant by greed was more what they supposedly tried to do: play off TSMC versus Samsung.
But since TSMC called their bluff they ended having to go with Samsung.
Whether or not Samsung over-promised is not really relevant: a 600mm² plus die was always going to be tricky whereas Nvidia's previous Samsung made GPUs were far smaller.
Of course, like Zhaoman on the Asus thread mentioned, Nvidia also have form with announcing FE prices below the AIB prices and then not really intending to deliver many at that price. Nvidia have a long history of getting release day reviews looking good and semi-fictional RRP help with that.
I'm curious how much of the lack of headroom is power limits - as the cards track current into the GPU then it complicates shunt modding. GPU-z screenshots still show power as the limit (not temps) even with the power slider cranked to the max, so I suspect that once someone figures out a way to cheap the GPU power reporting they'll see much better OC headroom. A wider design with similar performing cores will have a sweet spot at a much higher power level
There are currently 1 users browsing this thread. (0 members and 1 guests)