makes sense, multi-sourcing it will be cheaper, and the modules are inherently faster than 6x even at the same clock. Micron starts production in Q1

Cheaper to make, but "new" tech is always a reason to charge more.

cheaper for them double the price for us?

16gb/128bit is possibility with gddr7. It has enough bandwidth for nvidia to consider such configuration.

tsunami231:

cheaper for them double the price for us?

Sorry but do you think R&D is free, it cost AT LEAST twice as much hence the price increases, should be grateful this new process allow us to use less resources, time, energy, space somehow it is twice expensive cause R&D /s Bollocks.

hijodeosiris:

Sorry but do you think R&D is free, it cost AT LEAST twice as much hence the price increases, should be grateful this new process allow us to use less resources, time, energy, space somehow it is twice expensive cause R&D /s Bollocks.

your probably one those people that think the prices NVidia is charging for cards are also reasonable. which at this point are anything but

hijodeosiris:

Sorry but do you think R&D is free, it cost AT LEAST twice as much hence the price increases, should be grateful this new process allow us to use less resources, time, energy, space somehow it is twice expensive cause R&D /s Bollocks.

tsunami231:

your probably one those people that think the prices NVidia is charging for cards are also reasonable. which at this point are anything but

You guys deserve each other.

tsunami231:

your probably one those people that think the prices NVidia is charging for cards are also reasonable. which at this point are anything but

Graphics card are a luxury rather than a commodity, so I see their prices as such that they can charge whatever they like and its fine to do so, it just means "I" don't buy at those prices. Others don't seem to agree with me on that and do happily buy one for the price. In saying that, i'm also going to lean towards the cost is the full suite of capabilities, which these days is more than just a card and drivers, I'm sure once AMD has an equivalent stack they'll do no differently.

Astyanax:

Graphics card are a luxury rather than a commodity, so I see their prices as such that they can charge whatever they like and its fine to do so, it just means "I" don't buy at those prices. Others don't seem to agree with me on that and do happily buy one for the price. In saying that, i'm also going to lean towards the cost is the full suite of capabilities, which these days is more than just a card and drivers, I'm sure once AMD has an equivalent stack they'll do no differently.

usually how it works but my issue with price is price have more doubled for some series of cards. pay that and get almost not gains from gen to gen in case of majority 4xxx series except the stupidly highly priced card is just wrong. People are gona buy what they gona buy reguardless if price right or stupidly and profits will be made off that stupidly priced cards. I am sure AMD would do the same thing if they had the hardware. but way thing are right now most people that where in hobby are not gona have keep the hobby if they card they can afford are crap. Never mind the obsession with DLSS/FSR main sell or the fact game seem to need them now just to run ok. it is what is. I want a 4070 or 4070ti but zero chance i paying those prices and 4060 cards are not worth there cost to performance. I do agree it is luxury purchase. GDDR7 chips would atlest make 128bit bus usable again. the question is at what cost and is nvidia gona butcher bandwidth more. thew whole setup on 4060 card reason enough for me to never touch them And I know i complain alot do most venting here, people are welcome to ignore block me what ever. there tons people on here i have on ignore

tsunami231:

I want a 4070 or 4070ti but zero chance i paying those prices and 4060 cards are not worth there cost to performance. I do agree it is luxury purchase.

I had to go look up the price of the 4070 now to see what you were "those prices" you say you have "zero chance" of paying.... about £600 to £700... really?? For basically 80-90% of the performance of the 4080 which people paid £1500 to £1700 for?? And I did go from 1070 to 3070 to 4080 so I know the kind of performance jump between each generation. And here is my previous purchase prices: In Jan 2021 Gigabyte GeForce RTX 3070 Aorus Master 8GB OC GPU £679.99 In Oct 2016 Gigabyte GeForce GTX 1070 XTREME GAMING 8GB GDDR5 £469.99 About the same prices at the time for much slower hardware. You should be happy you are getting that level of performance (in 4070) so relatively cheaply in comparison.

yes happy series that which i payed 450$ for is now 800$ for same series that was general same price gen to gen till 4xxx and 3xxx series. every has the idea of what normal price and what screw up and what "inflation" is acceptable. btw nvidia prices go well beyond those inflation. this nvidia know people will pay it and people will even pay 3x there price to scalapers. BTW paying more for 3070 performance with 4070 is not good thing in alot people eyes same for 4060 etc etc it why alot reviews bomb nvidia for or and why nivida now change how reviews can be done it is what it is So long nvidia goes bastardized there bandwidth more and charge more for less and they trying justify price with more AI software to make up HW deficient

tsunami231:

your probably one those people that think the prices NVidia is charging for cards are also reasonable. which at this point are anything but

offtopic note the " /s" at the end of the text... jezz I guess is just way to subtle or is just a thing "boomers" don't catch? It stands for "sarcasm". ontopic. Yeah lots of people including lots of guruian have that posture, cause "capitalism gonna capita-lly(??)" so to them any price, move, tactic is valid, and the rest should either waltz at their rythm or STFU and go back to eat beans or w/e poor peasants do, let the "luxury" (LOL) items to the royalty. Go figure the mental gymnastics. tune up for later news "Water is a luxury, and just people who can afford should have access to" /SSS

hijodeosiris:

offtopic note the " /s" at the end of the text... jezz I guess is just way to subtle or is just a thing "boomers" don't catch? It stands for "sarcasm". ontopic. Yeah lots of people including lots of guruian have that posture, cause "capitalism gonna capita-lly(??)" so to them any price, move, tactic is valid, and the rest should either waltz at their rythm or STFU and go back to eat beans or w/e poor peasants do, let the "luxury" (LOL) items to the royalty. Go figure the mental gymnastics. tune up for later news "Water is a luxury, and just people who can afford should have access to" /SSS

Your posts are so well thought out and smart, that I feel dumber every time I read one.

AuerX:

Your posts are so well thought out and smart, that I feel dumber every time I read one.

Can you get your IQ lower than room temp? Is that a challenge,a bluff or a simple apreciation of pride about your condition, either way good for you, keep on the good dude/dudette w/e.

Hopefully nvidia wont shrink the bus size again so 5070 can have 256bit/16gb configuration same as 4080. That would be great if they actually keep the pricing same as 4070.

no need for 7, not yet - we have barely scratched the surface with mesh shaders and new pixel and vertex shaders. Games are just not that big anymore, you can have one texture and use scaling in the engine for loading that one texture through one sample without constantly trying to shunt vram in lots of directions. This is not making any sense, unless they got a better deal than the one they got - as the thing that really pushes pixel, vertex and mesh shaders is X threads and X core counts, as these shaders are part of HLSL which in itself is able to run this concurrently...therefore more threads and cores is more important than the speed of the memory. 5 and 6 are fine and dandy, 7 is for sure awesome, but it is a waste when you consider the awesomeness of these new shaders and old shaders with new versions and instructions.

Loobyluggs:

no need for 7, not yet - we have barely scratched the surface with mesh shaders and new pixel and vertex shaders. Games are just not that big anymore, you can have one texture and use scaling in the engine for loading that one texture through one sample without constantly trying to shunt vram in lots of directions. This is not making any sense, unless they got a better deal than the one they got - as the thing that really pushes pixel, vertex and mesh shaders is X threads and X core counts, as these shaders are part of HLSL which in itself is able to run this concurrently...therefore more threads and cores is more important than the speed of the memory. 5 and 6 are fine and dandy, 7 is for sure awesome, but it is a waste when you consider the awesomeness of these new shaders and old shaders with new versions and instructions.

You can rationalize it any which way, but extremely unlikely Nvidia would not use GDDR 7 when its out. They are going to carve out every ounce of performance they can out of the 5xxx series to be able to justify whatever price gouging they have in store for us. The timing of the 5xxx series and GDDR 7 makes it a virtual certainty they are going to use it.

alanm:

You can rationalize it any which way, but extremely unlikely Nvidia would not use GDDR 7 when its out. They are going to carve out every ounce of performance they can out of the 5xxx series to be able to justify whatever price gouging they have in store for us. The timing of the 5xxx series and GDDR 7 makes it a virtual certainty they are going to use it.

Oh, naturally - but if you give developers 98gb to play with, they are just going to use it, and demand for another 48gb! HLSL of API's from everyone (even AMD!) are extremely powerful now, but devs are not tapping the power - Starefield is a good example of this, with next to ZERO development of new shader languages, relying on old shaders from DX9 days, maybe even expanded upon from DX8 days as I mentioned in a prior thread. Just because you compile a shader from 2010 in DX12, does not mean it is using DX12 HLSL for vertex, pixel and geometry...more power to developers, but seriously, faster vram is just not needed right now, not with development where it is.