680 was supposed to be the 660Ti?

Status
Not open for further replies.
Permabanned
Joined
20 Oct 2012
Posts
1,800
Ive been reading this more and more latley.

The 680 was really supposed to be the 660Ti?

And Nvidia decided to rob everyone and make it the 680 because it beat the 7970 up until them wonder drivers?

If this is the case it still shows Nvidia are a MUCH better company for making graphics cards! If a card that was supposed to be around £200 could beat a card that was £400+ on launch (is still £300+ now) and costs way less to produce just shows they make better cards does it not?

Even now the 770 is the same as the 7970 and is still cheaper to make, has less Vram but yeah it should do the same job at 1080p as the 7970 until games that use 2GB as standard start rolling out.

If the 780 was supposed to the the 680 why is it a cut down Titan? Im sure that was never planned or maybe it was i dunno. But if it wasnt then do Nvidia have a card similar specs as the 7970 hidden away and are just using the scraps of the Titan to make more dollar? Maybe they just give up on it as it would be another card produce and cost more then to use Titan scraps?

Either way it shows probably by the 800 series if they dont rip peoples eyes out they should be really good cards and indeed kill AMD?

So at the end of the day Nvidia have made a killing and nobody really knows whats what, but if the 680 was indeed the 660Ti and was then changed they sure did produce good cards but cashed in big time!
 
No it wasn't.

Bait thread.

How is this a bait thread? Its a Question and my views on it?

I own a 7970 btw.

I can see the point though, why go 2GB when they have already seen the 7950 and 7970 released with 3GB, it does sound right in a way being a mid range card and then it was soo good it beat the 7970 so they though what the heck its our flagship now!

Can you see the connection though?

But yeah it could have been the flagship all along, either way it was cheaper to make and the 770 re badged is just as good as the 7970 now, still cheaper to make.
 
How is this a bait thread? Its a Question and my views on it?

I own a 7970 btw.

I can see the point though, why go 2GB when they have already seen the 7950 and 7970 released with 3GB, it does sound right in a way being a mid range card and then it was soo good it beat the 7970 so they though what the heck its our flagship now!

Can you see the connection though?

But yeah it could have been the flagship all along, either way it was cheaper to make and the 770 re badged is just as good as the 7970 now, still cheaper to make.

No I can't, because they went 2GB and 256bit bus to cut costs.

The GTX680 is the way it is because nVidia wanted to cut costs.
 
How was the 680 supposed to be the 660Ti?

Google it im sure it will tell you.

But its said because of the spec of the card it doesnt match up to the spec of the AMD cards and overall seems a more budget card, specially to produce. But because of the performance compared to a 7970 it was then classed as a top of the range card and sold as one.

Either way it shows that the 7970 should be even faster then it already is, say Nvidia made the 7970 and the 680 im sure the 7970 would be a lot faster then the 680, if you get me?

I dont know which is true either but it could be if you look at everything it does make sense.
 
No I can't, because they went 2GB and 256bit bus to cut costs.

The GTX680 is the way it is because nVidia wanted to cut costs.

That is also a good point but doesnt change the fact that a cheaper card matches a more expensive card and therefore proves Nvidia make better cards, does it not?

Like i said above if Nvidia released a 680 with the specs of a 7970 then Nvidias 7970 would be better then AMDs 7970? I know its kinda confusing but im sure you understand me?

And no the 780 is not Nvidias 7970 as it is a more expensive card then the 7970 to produce.
 
That is also a good point but doesnt change the fact that a cheaper card matches a more expensive card and therefore proves Nvidia make better cards, does it not?
What? How've you worked that one out?

Like i said above if Nvidia released a 680 with the specs of a 7970 THAN Nvidias 7970 would be better THAN AMDs 7970? I know its kinda confusing but im sure you understand me?

It's not confusing, you're just mistaken in your assumptions.

You seem to think that die size and production costs and games performance are intrinsically linked between AMD and nVidia, they're not really.

And no the 780 is not Nvidias 7970 as it is a more expensive card then the 7970 to produce.

Huh?
 
Theres quite a bit of evidence that tends to indicate that the GTX680 was never originally planned to be on the GK104 process and nothing to indicate that it was a cost driven decision other than speculation.
 
How i worked it out -

680 beats 7970 on release due to 7970 bad drivers.

7970 catches up and passes due to new drivers, not by much mind.

770 released still 680 really but matches and beats 7970.

770 is cheaper to produce = Nvidia makes better cards?

I know they are totally different cards all round but spec wise the 7970 should be much faster then the 680/770 should it not?

Take prices away and say the 680 is cheaper it doesnt make a difference its still a better card for what spec it has?

Last bit was in case people said the 780 ect.
 
Theres quite a bit of evidence that tends to indicate that the GTX680 was never originally planned to be on the GK104 process and nothing to indicate that it was a cost driven decision other than speculation.

Nothing? Really? Everything about what nVidia have done with Kepler points to cost cutting.

Had they not rinsed the crap out of GK104 killing off the FP64 performance, it would have been a bigger chip.

256bit bus, cost cutting, 2GB, cost cutting, voltage lock, cost cutting, 300mm² GPU, cost cutting.

500mm² GPUs have been unsustainable, nVidia needed to reduce production costs, with Kepler, they reduced production costs and raised price points, which has meant big profit at the expense of the customer.
 
How i worked it out -

680 beats 7970 on release due to 7970 bad drivers.

Unoptimised drivers, this is quite a typical situation with major architectural changes.

The gap wasn't massive in the first place though.

7970 catches up and passes due to new drivers, not by much mind.

The 7970 is quite a bit faster, especially when it comes to high res situations.

770 released still 680 really but matches and beats 7970.

It matches a 7970GE, but it's just an overclocked 680.

770 is cheaper to produce = Nvidia makes better cards?

How does this make sense? AMD's Tahiti GPUs have an order of magnitude greater compute performance than GK104/GTX660/670/680/770.

I know they are totally different cards all round but spec wise the 7970 should be much faster then the 680/770 should it not?

The 7970's GPU is around the same size as GK104, so will cost roughly similar amounts to produce.

AMD's cards are more expensive to produce due to the greater amount of RAM, 384 bit memory bus and beefier voltage circuitry.

So I'm really not sure how you're coming to the conclusions you are coming to.

Take prices away and say the 680 is cheaper it doesnt make a difference its still a better card for what spec it has?

No :confused:
 
Nothing? Really? Everything about what nVidia have done with Kepler points to cost cutting.

Had they not rinsed the crap out of GK104 killing off the FP64 performance, it would have been a bigger chip.

256bit bus, cost cutting, 2GB, cost cutting, voltage lock, cost cutting, 300mm² GPU, cost cutting.

500mm² GPUs have been unsustainable, nVidia needed to reduce production costs, with Kepler, they reduced production costs and raised price points, which has meant big profit at the expense of the customer.

Thats all speculation tho - nVidia has "needed" to reduce production costs in the past for several generations and never did so, 256bit bus is as much if not more so an indication of a process being developed with a mid-range product in mind than it is a cost cutting measure, coupled with the configuration of the SMs with cache and scheduler implementations not typical of those found in a GPU intended for the high end product and it far from conclusively points to being a cost cutting measure and when combined with other aspects like earlier roadmaps showing the 680 having a different configuration to the other 600 cards, the 680 device ID being identified as the 670ti in drivers around release and 660ti in much older development drivers, along with the process numbering being a departure from the normal convention for nVidia's high end GeForce cores and several other factors - some of it mentioned here ( http://www.techpowerup.com/162901/did-nvidia-originally-intend-to-call-gtx-680-as-gtx-670-ti.html ) it tends to tell a different story than just being cost cutting.
 
Status
Not open for further replies.
Back
Top Bottom