NVIDIA GeForce RTX 3070 Launch Postponed to October 29th

[deu]

If you cant act > React. Everything Nvidia is doing right now says that they know that AMD is bringing the guns and that they are scared. Its been a long time since Nvidia have acted this way, so looking forward to AMDs launch, to see how it preasures nvidias prices even futher

 

[watzupken]

Do the reviewers warn us that their results are atypical since their cards boost 10 - 15% above "guaranteed boost clock" and that users might experience more than 10% lower performance? No?

If none of the reviewers mentioned this, how would you know then? This practice has been around for a long time, and it’s nothing surprising.

 

[cueman]

fiasco? some1 there took fiasco...amd fan...?

well...
i remembe radeon vii, furyX, rx480... continue.....?..ok vega64 and rx 500 series all.....


to point:

rtx 3080 is true 4K gpu, and that amd cant absolutly cant release with navi components...its pure 4K gpu.

...and,under that resolution,you cant get better gpu than 3060/3070 series ones...thouse also oc'd like hell,you can count of that.


rdna2 aka rx 6900 xt is pure rx 5700 xt Crossfire gpu...no doubt, and there is alot tested variant for net. google it......
example crossfire between gpus 5600 xt/5700 it loose clear rtx 3080 and its tdp is 464w!.. yes 464W!!
also rx 5700 xt Cf is there...no help. 500w easily!

all can make something clue for that....


fact:
rx 6900 xt use same core 7nm like rx 5700 xt and same memory gddr6...and have RT support..what btw,, eat more juice.. 50W min.
rx 6900 xt use gddr6 mem for sure, hbm variants raise price skyhigh..so nope,too pricey.

amd rx 6900 xt CANT be moore than that above, its pure true fact...and if rx 6900 xt tdp is really 300W......really if so, ...then rtx 3070 ti is enough.

if its about 390-420W...it might can battle near against rtx 3080,but still,not 4K level so fast like rtx 3080.


note:
rtx 3000 series selling like hell,all over world sell out....so, its clear that IF, i repeat if amd rx 6900 xt can performance good,
amd sure 'leaks' something of it out..
reason:
just stop that sold out maniac (what rtx 3080/90 really earn btw..)

so, if not next weekend end or max next 10 days nothing performance 'leaks' not coming amds side out...... i say, looks bad for amd fans...


my research and calculate for rx 6900 xt performance is about 70-75% rtx 3080 performance,meaning 10% faster than rtx 2080 ti...and this from 300W tdp.
....but there is coming rtx 3080/super and Ti models bfore xmas,much earlier.


but, also it is nice to get also intel Xe gpu out soon, then we have 3 competititon.
its always good!

prices not will be more issue anymore ,it will be same class,just performance and efficiency.

(and dont even dream that rx 6900 xt is cheap, i guess 699-799$,at least!)




- end -

 

[Valantar]

fiasco? some1 there took fiasco...amd fan...?

well...
i remembe radeon vii, furyX, rx480... continue.....?..ok vega64 and rx 500 series all.....


to point:

rtx 3080 is true 4K gpu, and that amd cant absolutly cant release with navi components...its pure 4K gpu.

...and,under that resolution,you cant get better gpu than 3060/3070 series ones...thouse also oc'd like hell,you can count of that.


rdna2 aka rx 6900 xt is pure rx 5700 xt Crossfire gpu...no doubt, and there is alot tested variant for net. google it......
example crossfire between gpus 5600 xt/5700 it loose clear rtx 3080 and its tdp is 464w!.. yes 464W!!
also rx 5700 xt Cf is there...no help. 500w easily!

all can make something clue for that....


fact:
rx 6900 xt use same core 7nm like rx 5700 xt and same memory gddr6...and have RT support..what btw,, eat more juice.. 50W min.
rx 6900 xt use gddr6 mem for sure, hbm variants raise price skyhigh..so nope,too pricey.

amd rx 6900 xt CANT be moore than that above, its pure true fact...and if rx 6900 xt tdp is really 300W......really if so, ...then rtx 3070 ti is enough.

if its about 390-420W...it might can battle near against rtx 3080,but still,not 4K level so fast like rtx 3080.


note:
rtx 3000 series selling like hell,all over world sell out....so, its clear that IF, i repeat if amd rx 6900 xt can performance good,
amd sure 'leaks' something of it out..
reason:
just stop that sold out maniac (what rtx 3080/90 really earn btw..)

so, if not next weekend end or max next 10 days nothing performance 'leaks' not coming amds side out...... i say, looks bad for amd fans...


my research and calculate for rx 6900 xt performance is about 70-75% rtx 3080 performance,meaning 10% faster than rtx 2080 ti...and this from 300W tdp.
....but there is coming rtx 3080/super and Ti models bfore xmas,much earlier.


but, also it is nice to get also intel Xe gpu out soon, then we have 3 competititon.
its always good!

prices not will be more issue anymore ,it will be same class,just performance and efficiency.

(and dont even dream that rx 6900 xt is cheap, i guess 699-799$,at least!)




- end -

a) You are for some reason entirely discounting the possibility of architectural efficiency and performance gains. That's a bit weird.
b) You can't compare CF/SLI results to single GPU results, as scaling is never linear, and is extremely dependent on drivers, game support, and various random factors. If one GPU has n number of CUs and another has 2n, two of the former in CF/SLI will (and has) never reliably performed the same as the latter.
c) "your research" - sources, please?
d) AMD leaking performance figures to stop Nvidia sales wouldn't make a dent. Nvidia's mindshare is far too strong for that. People would just get impatient and start complaining (even more than currently) that the new GPUs aren't out yet. Besides, leaking that would just entirely kill sales of their current products.

To be clear, I'm not saying the current rumors are true - I have no way of knowing, and some of them seem quite dubious. But your reasoning doesn't stand up to scrutiny.

 

[medi01]

...a price increase due to overwhelming demand.

Or underwhelming supply.

 

[EarthDog]

Do the reviewers warn us that their results are atypical since their cards boost 10 - 15% above "guaranteed boost clock" and that users might experience more than 10% lower performance? No?

lol wat?

Reviewers cards aren't boosting higher... its how boost works...it varies depending on factors.

 

[olstyle]

rtx 3080 is true 4K gpu

4k GPU with WQHD memory unfortunately.

Reviewers cards aren't boosting higher... its how boost works...it varies depending on factors.

One of the factors is temperature and since testers often use open benchtables they tend to get better results (but more like 1-5%, not 10%) there than in the average Joe's case.

 

[EarthDog]

4k GPU with WQHD memory unfortunately.

One of the factors is temperature and since testers often use open benchtables they tend to get better results (but more like 1-5%, not 10%) there than in the average Joe's case.

id be surprised if it was above 2/3% honestly. Nearly negligible.

 

[TheoneandonlyMrK]

While I am no Nvidia fan, but in their defence, Nvidia never guaranteed a boost above the boost clock they advertised. So say if your card can boost its clock speed in the high 1800 to 1900, these clocks are above and beyond the advertised boost clock speed. The fix if I am not wrong somewhat drops the max boost clock, but should still be comfortably ahead of the advertised numbers.

Imagine if Tesla juiced the first batch of model p100 so they could show faster actual track performance, then after receiving their car many owners started having thermal shut down issues on the way home from work , Tesla fix this issue with a OTA patch in two days.
All quite within the realms of possibility Cough Nikola, How do you think that would pan out, what are the customers saying?.

 

[Vayra86]

The fact half you guys here are still adamant AMD will just eclipse Nvidia with a single release is honestly a next level calibre of wishful thinking. I mean, how many times do you need reality played out in front of you to get the memo - you don't 'just' catch up 2 generations worth of performance. AMD's been at it since RX480, they didn't get an inch closer, in fact the gap has gotten wider.

But, dream on. Its easy to shit over things that don't exist yet. Do come back when you've been proven wrong I'll do the same.

Get ready for the big surprise: Ampere isn't entirely what it was expected to be/marketed as, and neither is Navi, so its likely a choice of evils. History repeats. Again... and again...

 

[Vya Domus]

you don't 'just' catch up 2 generations worth of performance.

Of course you can, you just make the largest chips possible right at the reticle limits of the node like Nvidia did for those past two generations. AMD simply chose not to do it because of the intense efforts Nvidia have put into making the high end segment nonviable for AMD in the past decade.

 

[Vayra86]

Of course you can, you just make the largest chips possible right at the reticle limits of the node like Nvidia did for those past two generations. AMD simply chose not to do it because of the intense efforts Nvidia have put into making the high end segment nonviable for AMD in the past decade.

Ah, yes, AMD was doing what they did out of pure comfort, I get it now

But yeah, as I posted earlier, dream on. Even the fabled Zen arch didn't manage to catch up to a sleeping giant with the first or even the second gen. They're just now getting performance parity on every front while not losing on any of them, we're moving onto 5000 series already. So yeah, there's reality for ya.

Did it occur to you that AMD forced itself into the midrange with Polaris because the age old 'make the biggest chip you can' nearly got them into bankruptcy? Navi is the first mid-high end range product in many years now that actually has some semblance of margins. You need to wake up, and do it fast. These world upside-down comments you make are pretty strange.

 

[Vya Domus]

Ah, yes, AMD was doing what they did out of pure comfort

You can call it that, if you prefer to dumb everything down to the simplest terms.

Even the fabled Zen arch didn't manage to catch up to a sleeping giant with the first or even the second gen.

Why is it fabled ? Zen is a figment of our imagination or what ? What does that even mean ?

Anyway, like many, you have a strange definition of "catch up" which is implied to mean something among the lines of "it must crush everything out there".

because the age old 'make the biggest chip you can' nearly got them into bankruptcy?

Funny how that never quite works out the same way when Nvidia does it. Almost as if there is something else ...

These world upside-down comments you make are pretty strange.

Only up-side down thing is believing that AMD somehow can't make large high performance chips. They can, the landscape is such that it's not worth it for them, you can continue to cry about that or just deal with it.

 

[M2B]

Ah, yes, AMD was doing what they did out of pure comfort, I get it now

But yeah, as I posted earlier, dream on. Even the fabled Zen arch didn't manage to catch up to a sleeping giant with the first or even the second gen. They're just now getting performance parity on every front while not losing on any of them, we're moving onto 5000 series already. So yeah, there's reality for ya.

Did it occur to you that AMD forced itself into the midrange with Polaris because the age old 'make the biggest chip you can' nearly got them into bankruptcy? Navi is the first mid-high end range product in many years now that actually has some semblance of margins. You need to wake up, and do it fast. These world upside-down comments you make are pretty strange.

And we should also keep in mind that AMD is going to be on a much better node with RDNA2, Samsung's 8nm process has decent density but it's absolutely no match to tsmc's N7P/N7+ in performance and efficiency.
Nvidia could've easily used a much better node but I think aside from the samsung node being less expensive, they just didn't want to make ampere too good, as weird as it sounds.
They kinda made that mistake with pascal and pascal made everything else look underwhelming, including Nvidia's own Turing generation.
Now this is my own theory and I could be totally wrong, but I think they just intended to make ampere good enough, and not great in the first place.
The point I'm trying to make is that Nvidia is kinda milking the market in their own way and they can probably do much better if they feel threatend in a serious way.

 

[wolf]

It appears that AMD have got Nvidia covered this generation. They are not impressed by the 3090. Nvidia cannot beat AMD in a price war as they operate with a higher margin. I feel that 'big navi' will beat the 3080 and therefore can beat any 3070 and can do it at a lower selling price. They will take some market share from Nvidia this time around.

Sounds like a MLID regurgitation to me, I really enjoy how much people claim to know before the cards are even announced. And as for the 3090, what a vague statement that is, not impressed does not equal they can handily outperform it or even match it, and really who was genuinely impressed? 10-15% better perf over a 3080 for more than double the cost, yeah, not impressive stuff.

nGreedia are very worried about AMD, aren't they! So much so that they blame the pushed back launch of the 3070 on their customers... Nothing to do with that they have no idea how to price or market this card yet...

nGreedia's marketing are truly customer hating scum.

When people say or type this word is the moment I know they're not worth listening to. NVIDIA and AMD alike are both companies, not charities, they will charge the most they can get people to pay, get over it, it's unhealthy and you'll feel better for rising above.

The 3070 is gimped with a paltry 8GB memory and is made on the awful Samsung 8nm process as we all know, so power draw is going to be awful compared to RDNA2.

I've heard AMD are going to match 2080 Ti performance with a 40CU card clocked around 2.4-2.5Ghz. This will pull Nvidia's pants down as it'll draw less power, have 12GB of memory (GDDR6) and cost $50 less.

You can count the amount of people that got Ampere cards at those prices on two hands probably. This is Nvidia's marketing 'trick'.

You've heard hey?

And plenty of people already have 3080's and 3090's, plenty got them day 1 and many more since then at or acceptably close to MSRP too. Yes the launch had issues but the cards are out in the wild in ever growing numbers.

Reduced boost clocks is not a fix bud.

Right, that must be why owners report either same, slightly lower or even sometimes slightly higher boost clocks, with better stability, overclockability and better performance, what a garbage driver update.

NVIDIA, the company people love to hate.

 

[Deleted member 185088]

While I am no Nvidia fan, but in their defence, Nvidia never guaranteed a boost above the boost clock they advertised. So say if your card can boost its clock speed in the high 1800 to 1900, these clocks are above and beyond the advertised boost clock speed. The fix if I am not wrong somewhat drops the max boost clock, but should still be comfortably ahead of the advertised numbers.

Why did they let the GPUs clock higher in the first place. Those things happen.

 

[P4-630]

Photo surfaced of alleged GA104-300 chips in the validation process.

The GA104-300 die size is 392,5 mm2.
The full GA104 chip contains 48 SMs, but the RTX 3070 will have 46 active.
This means that a RTX 3070 Ti / Super may be based on a fully enabled GA104 chip.

 

[r9]

Sounds like a MLID regurgitation to me, I really enjoy how much people claim to know before the cards are even announced. And as for the 3090, what a vague statement that is, not impressed does not equal they can handily outperform it or even match it, and really who was genuinely impressed? 10-15% better perf over a 3080 for more than double the cost, yeah, not impressive stuff.

When people say or type this word is the moment I know they're not worth listening to. NVIDIA and AMD alike are both companies, not charities, they will charge the most they can get people to pay, get over it, it's unhealthy and you'll feel better for rising above.

You've heard hey?

And plenty of people already have 3080's and 3090's, plenty got them day 1 and many more since then at or acceptably close to MSRP too. Yes the launch had issues but the cards are out in the wild in ever growing numbers.


Right, that must be why owners report either same, slightly lower or even sometimes slightly higher boost clocks, with better stability, overclockability and better performance, what a garbage driver update.

NVIDIA, the company people love to hate.

I have nothing against nvidia, it's just I don't buy their excuse that they gonna make people wait longer so they don't wait longer. That's like fucking for virginity.

 

[BoboOOZ]

I have nothing against nvidia, it's just I don't buy their excuse that they gonna make people wait longer so they don't wait longer. That's like fucking for virginity.

It's obviously a marketing move. They want to steal some of AMD s thunder, and if Navi is looking too good, they will announce some bundles, a 3080 or 3070 KO, or they might even adjust some prices.
Nvidia is a company doing what companies do when they are almost a monopoly, and this isn't a good situation. The market needs a strong showing from the Radeon group really badly.

 

[John Naylor]

NV already did try to steal the thunder with their 3080 and 3090 release. You can tell me how did that go. Since we will have them in the market AMD will be out with their products anyway.

Not commenting on the quality of the products but I expect it would expect same as it did with the 780n Ti .... millions of AMD marketing down the tube with a single press release.

"Just wait till you see our new big thing ....the 290 / 290x will crush both nVidia's top offerings. Then 5 days before, vVidia responded saying we've had this design sitting on the shelf waiting for something to challenge the 780, it will be in stores 5 days after the 290x. And then the reviews came out, the 2xx series was so aggressively clocked that it ran at 95C, had no OC headroom while the 7xx series 25 - 30% OC meant nvidia still had the 2 fastest cards

I mean it's not as if we have ever seen this level of scarcity with new release .... well except with thew 2xxx series .... and the 1xxx series .... and the 9xx series..... marketing is expensive.... easier and cheaper to let the press do all your marketing for free by writing over and over about scarcity and unprecedented demand. The early market is tied to the market segment who aligns their self worth with "being the first to have the next big thing .... "look at me look at me ... look what I got. This will get me likes on instagram !

So they will get their thumbs up, they will pay more, they will suffer the bleeding edge cuts ... and nvidia will get tons of fee pres,. And those who wait, will experience no stress, no hours spent in late night sniping, no immature BIOSs, no bugged designs and lower prices for improved products.

All the whining and complaining is just free press for the product. Don't really care when it comes out because won't buy any new product in 1st 3 months ... will wait for the new and improved later steppings and I will pay less for a better product/

 

[olstyle]

The big design was already launched and even the smaller card for the big chip. That's the opposite of the 780ti situation.
The 3070 can only attack on price.

 

[ratirt]

Not commenting on the quality of the products but I expect it would expect same as it did with the 780n Ti .... millions of AMD marketing down the tube with a single press release.

"Just wait till you see our new big thing ....the 290 / 290x will crush both nVidia's top offerings. Then 5 days before, vVidia responded saying we've had this design sitting on the shelf waiting for something to challenge the 780, it will be in stores 5 days after the 290x. And then the reviews came out, the 2xx series was so aggressively clocked that it ran at 95C, had no OC headroom while the 7xx series 25 - 30% OC meant nvidia still had the 2 fastest cards

I mean it's not as if we have ever seen this level of scarcity with new release .... well except with thew 2xxx series .... and the 1xxx series .... and the 9xx series..... marketing is expensive.... easier and cheaper to let the press do all your marketing for free by writing over and over about scarcity and unprecedented demand. The early market is tied to the market segment who aligns their self worth with "being the first to have the next big thing .... "look at me look at me ... look what I got. This will get me likes on instagram !

So they will get their thumbs up, they will pay more, they will suffer the bleeding edge cuts ... and nvidia will get tons of fee pres,. And those who wait, will experience no stress, no hours spent in late night sniping, no immature BIOSs, no bugged designs and lower prices for improved products.

All the whining and complaining is just free press for the product. Don't really care when it comes out because won't buy any new product in 1st 3 months ... will wait for the new and improved later steppings and I will pay less for a better product/

Sure but how I see it, AMD doesn't have to beat the 3090 or even 3080. With what the prices for these two are, there are 2 things that matter here. Availability and price. If AMD get products out and make sure of the availability and proper pricing its a win.
With the demand I wouldn't be so sure. As I've seen it is simply supply scarcity. The demand is there for sure but most of that is due to supply and a lot of Tech youtube channels confirmed it or has same conclusions. So, availability and pricing will play a huge role in AMD RDNA2 success on top of other things but still.