-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
** NVIDIA GTX 1080 FOUNDERS EDITION: WANNA PRE-ORDER?
- Thread starter Gibbo
- Start date
Nvidia release GPU's based on certain chips. There is the high end chip, which always ends in 0, like the GF100, GF110, GK110 etc. The mid range chip always ends in 4, the GF104, GF114. You get the idea.
Just name a few of the cards associated with those chips.
GF100 - GTX 480, GTX 470.
GF110 - GTX 580, GTX 570.
GK110 - Titan, 780, 780Ti.
GF104 - GTX 460,
GF114 - GTX 560,
GK104 - GTX 680
GF114 - GTX 770.
These mid range chips are easier and cheaper to make than the full fat chips.
The 780 for example was a high end card, it was a GK110 chip. The 680, 980 and now 1080 are not high end, they are built on mid range parts.
If we are using the reckoning that it's the highest performing card out now and that makes it a high end card, then every mid range card ever released was also high end and that's absurd.
Nvidia have very cleverly adjusted the market with releasing the high end chips much later than the mid range and renaming the mid range as X80, X70 (980, 970 for example) And raised the price accordingly too.
And now people are not only paying extra for mid range cards, but also convinced that they are high end cards!! A stroke of genius by Nvidia, getting people to pay high end prices for cards that are cheaper for them to make.
Something people seem to be missing about the pricing of the GTX 1070 and 1080.
There probably is very little difference in manufacturing cost of the two cards.
Yes I know the 1080 uses GDDR5X but it won't be a massive amount more expensive than GDDR5.
The 1070 also has some cores disabled but it is still the same chip.
My point is there will be a massive difference in pricing between the 2 cards.
Some would argue that the 1070 uses chips that can not be used in the 1080 but even this is lame as the 70 will probably outsell the 80 by a factor of 5 to 1.
There probably is very little difference in manufacturing cost of the two cards.
Yes I know the 1080 uses GDDR5X but it won't be a massive amount more expensive than GDDR5.
The 1070 also has some cores disabled but it is still the same chip.
My point is there will be a massive difference in pricing between the 2 cards.
Some would argue that the 1070 uses chips that can not be used in the 1080 but even this is lame as the 70 will probably outsell the 80 by a factor of 5 to 1.
£620 - 650 for a 699$ Card.
This price should be for a Ti not an average joe 1080
THIS IS NEARLY AS FUNNY AS THE SIMILAR PRODUCT TO THE 1080.
At these prices it is XMAS for OcUK.
Soldato
- Joined
- 6 Feb 2010
- Posts
- 14,582
This is what I have been saying for a while...Nvidia are not pricing the card based on what they are anymore...they are calling 60/60Ti 70 and 80 cards, with a price to match...no, now even higher than that, just because the 1080 is a tinny bit faster than 980Ti (when both are overclocked) which is a last gen card.
I will buy three to see what gains I can get out of them; i can always resell if there is literally no point in it! It is possible and I want to test it!
I think you, and the other peeps on here who have been kind enough to educate me, are right about the FE; the 8+6 pin power debacle seems to have been confirmed(?) and therefore the gains over the FE from the partner cards will be much larger.
I will see what happens on the 27th!
Thanks mate
Believe me, there is no challenge here. It's only a lot of effort for nothing. Nvidia has essentially made it so complicated to get 3 and 4 way systems going that it has virtually killed it off. And if Nvidia aren't wasting time with 3-4 way setups can you really see developers going to the bother? I mean SLI and crossfire support is pretty bad already.
Me too. It's just annoying me a little that people are comparing it to the 8800 GTX Ultra and previous high end cards like that and saying the price hasn't changed.
Changing the name of the mid range cards to X80 and X70 was brilliant. You gotta hand it to them. Here take a look at this, it shows the top and mid range Fermi cards, based on chip size with the Maxwell cards. Where they are in line and what they would have been named if Nvidia had kept with same numbering scheme that they used up until Kepler.
580 GTX(GF110) -- Titan X (GM100) -- 980 GTX
570 GTX(GF110) --- 980Ti (GM100) ---- 970 GTX
560 Ti (GF114) ---- 980 GTX(GM104) -- 960 Ti
560 (GF114) ---- 970GTX(GM104) -- 960
I wonder would people say that the 980 was high end if it was called the 960Ti?
That's very true. It's a real shame if they are trying to kill of 3-4 setups. It doesn't make much sense from a profit perspective though does it!
Thanks for the advice Melmac, much appreciated!
Soldato
- Joined
- 19 Oct 2008
- Posts
- 6,042
That's very true. It's a real shame if they are trying to kill of 3-4 setups. It doesn't make much sense from a profit perspective though does it!
Thanks for the advice Melmac, much appreciated!
Should be automatically scalable now across graphic cards. I'm not sure why the devs have to care about multi-gpu setups, it should just work from either hardware/OS or DirectX - some kind of controller that just pushes out the work across however many GPU's are in a system. Like a RAID controller .
Last edited:
Agreed. Seems crazy - all the work loaded onto the developers when they could just use virtual scaling architecture within the bios (supported by the OS) to deliver the use of multiple GPU's to the application/game like a pool.
But it seems Nvidia have other ideas!
Cheers
That's very true. It's a real shame if they are trying to kill of 3-4 setups. It doesn't make much sense from a profit perspective though does it!
Thanks for the advice Melmac, much appreciated!
You are welcome!! It is amazing that they are making it so hard to do. There are people out there who want to trick around with things like this. Wouldn't it be great if there was some kind of way for 3 way and 4 setups that you could make your own profiles? Something that would take a lot of trial and effort but would make a difference if you got the right balance. It would be heaven for people like you and Kaap
I see what your both saying, and thanks for the info.
But what I was referring to before was the performance and price in comparison to other cards. It doesn't matter what the chip is called or the name of the card, but simply where they stack in the performance tree.
I'm not saying these, as you put it, 'mid range chips' cost nvidia what 'high end' parts should cost, or if they are overcharging or not.
Just simply the performance in relation to the previous generations and current generation puts them in the high end bracket.
Yes, I know what you are saying, it's what others are saying, but that has only started happening in the last few years.
New mid range cards after a die shrink have always been faster than the older high end cards, but never before has anyone said that those cards were high end. Look at the 8800GT, it was released before the 8800GTX but nobody suggested that the 8800 GT be called a high end card, even though it performed nearly as good as the 8800GTX and way faster than anything previously.
It's a mid range card pure and simple. Only clever marketing has changed people's perceptions and allowed for price increases.
Ultimately it is the people who buy the cards that set the MSRP.
If they don't buy the cards the MSRP will go down.
If they do buy the cards and the demand is high the MSRP will go up.
Supply and demand, a force more powerful than nuclear fusion.
Unfortunately Kaap you couldn't be more wrong.
Customers buying the cards do indeed alter the 'retail price', but not the MSRP as by definition it is the Manufacturer Suggested Retail Price. as suggested by the manufacturer.
![[Damien]](/styles/overclockers/avatars/4.png){kind=avatar}