NVIDIA ‘Ampere’ 8nm Graphics Cards

[Undesirable]

Are they really going to wait until the new consoles have stolen all the limelight before they release any juicy specs / benchmarks of the new card? I don't care about the frigging release schedule / trade show promotion / marketing / competition / NDAs, just give us the bloody card for christ sake!

Not only had they best hurry up, but it also be fairly priced and live up to performance expectation or I'm buying one of those new consoles and a nice new OLED TV to tide me over, and I don't care what they do if I'm forced to go down that road.

Fed up with the GPU / monitor market.

 

[Panos]

Are they really going to wait until the new consoles have stolen all the limelight before they release any juicy specs / benchmarks of the new card? I don't care about the frigging release schedule / trade show promotion / marketing / competition / NDAs, just give us the bloody card for christ sake!

Not only had they best hurry up, but it also be fairly priced and live up to performance expectation or I'm buying one of those new consoles to tide me over, and I don't care what they do if I'm forced to go down that road.

Thats something raised on the other discussion. Everyone expected at GDC that the keynote from Jensen to make some big announcements where the company is heading this year with the new gen. Nobody expected specs etc but architecture and the vision. However while everyone else published their speeches & presentations on video through YT & Twitch for the GDC, AMD had to transform the 24fps RDNA2 RT video to 60fps for YT, others made the videos from their bedrooms, Nvidia hasn't uploaded anything in relation to the GDC. On the NV website states that they will publish something but nothing 12 days later.

Someone said in here that I read between the lines, however if NV had something impressive surely there would be a hint other than they support DXR1.1 (a half arsd statement) with RTX2000 series. Seriously where is Jensen's keynote? They could have published something, even in text format. Or he was going up there on stage without anything planned to say to the developers? But even standup comedians have some generic plan & script of what to say.....

 

[Spook187]

Why has this thread turned into console thread, away with you peasants

 

[Grim5]

Did Panos really say PC gamers are peasants

hahaha this guy is ridiculous

 

[Richdog]

Are they really going to wait until the new consoles have stolen all the limelight before they release any juicy specs / benchmarks of the new card? I don't care about the frigging release schedule / trade show promotion / marketing / competition / NDAs, just give us the bloody card for christ sake!

Not only had they best hurry up, but it also be fairly priced and live up to performance expectation or I'm buying one of those new consoles and a nice new OLED TV to tide me over, and I don't care what they do if I'm forced to go down that road.

Fed up with the GPU / monitor market.

I don't think Nvidia are going to rush-release their cards because of angry internet forumites. Consoles have always been direct competition and the new cards are still going to sell like hot cakes.

 

[Calin Banc]

Thats something raised on the other discussion. Everyone expected at GDC that the keynote from Jensen to make some big announcements where the company is heading this year with the new gen. Nobody expected specs etc but architecture and the vision. However while everyone else published their speeches & presentations on video through YT & Twitch for the GDC, AMD had to transform the 24fps RDNA2 RT video to 60fps for YT, others made the videos from their bedrooms, Nvidia hasn't uploaded anything in relation to the GDC. On the NV website states that they will publish something but nothing 12 days later.

Someone said in here that I read between the lines, however if NV had something impressive surely there would be a hint other than they support DXR1.1 (a half arsd statement) with RTX2000 series. Seriously where is Jensen's keynote? They could have published something, even in text format. Or he was going up there on stage without anything planned to say to the developers? But even standup comedians have some generic plan & script of what to say.....

Well, people were saying that to shrink something like the 2080ti on the new process, will still be around 4-500mm^2 (?), which is quite a lot still and will be pricey and probably the same will be with the rest. Considering the Corona situation and the new consoles, most likely nVIDIA and AMD are looking more into 149-299$ price range to have good, competitive offers. Some will still buy crazy expensive hardware, but I doubt it will be as many as they're were doing it until now.

Also these 2 (and you can include Intel), can't sell at a loss and have profits based on games sales or subscription models. Interesting times indeed.

 

[Grim5]

Well, people were saying that to shrink something like the 2080ti on the new process, will still be around 4-500mm^2 (?), which is quite a lot still and will be pricey and probably the same will be with the rest. Considering the Corona situation and the new consoles, most likely nVIDIA and AMD are looking more into 149-299$ price range to have good, competitive offers. Some will still buy crazy expensive hardware, but I doubt it will be as many as they're were doing it until now.

Also these 2 (and you can include Intel), can't sell at a loss and have profits based on games sales or subscription models. Interesting times indeed.

Who's still pandering that ridiculous rumor anyway. I know there used to be people saying the 3080ti will suck because a 7nm shrunk 2080ti is at the max size of 7nm die. We have actual Turing cards in actual benchmarks recording over 8100 Cuda cores.
We're nearly double the 2080ti core count. Either 1600m2 12nm dies are a thing or 7nm doesn't stop at 500nm like the haters believe or their maths is just wrong and the die on 7nm is a lot smaller.

 

[GAC]

Are they really going to wait until the new consoles have stolen all the limelight before they release any juicy specs / benchmarks of the new card? I don't care about the frigging release schedule / trade show promotion / marketing / competition / NDAs, just give us the bloody card for christ sake!

Not only had they best hurry up, but it also be fairly priced and live up to performance expectation or I'm buying one of those new consoles and a nice new OLED TV to tide me over, and I don't care what they do if I'm forced to go down that road.

Fed up with the GPU / monitor market.

they will be waiting for a few reasons basically why announce stuff when you wont have stock for them or even have the ability for people to buy them until the end of the year at best, what they will do once they have production up and running and stores are open to sell them they will get announced, fresh new news not 6-12 month old news that will be lost in a sea of other news.

is it annoying YES, but i understand why they are doing it.

now if they havent announced anything because amperes a damp fart again with very little performance gains then yeah they will be scrambling to move cards down the stack, rebadging stuff down a level guess we will see if the titan ampere ever comes out.

 

[D.P.]

Who's still pandering that ridiculous rumor anyway. I know there used to be people saying the 3080ti will suck because a 7nm shrunk 2080ti is at the max size of 7nm die. We have actual Turing cards in actual benchmarks recording over 8100 Cuda cores.
We're nearly double the 2080ti core count. Either 1600m2 12nm dies are a thing or 7nm doesn't stop at 500nm like the haters believe or their maths is just wrong and the die on 7nm is a lot smaller.

There maths is very wrong. Ampere is like quite big but a 2080ti on 7nmEUV is likely around 330mm^2.

Contrary to the anti-nvidia lies, the RT cores only take up about 4-5% of the die. Doubling RT performance is very straightforward. Moreover, many of the changes to Turing were in the backend scheduler in order to aid scaling up to to many more cores, but staying on 12nm there wasn't the die area. Therefore, to scale Turing to more cores takes less die area than one might expect.

 

[Panos]

Did Panos really say PC gamers are peasants

hahaha this guy is ridiculous

Open Steam survey..... Weep

 

[Panos]

Who's still pandering that ridiculous rumor anyway. I know there used to be people saying the 3080ti will suck because a 7nm shrunk 2080ti is at the max size of 7nm die. We have actual Turing cards in actual benchmarks recording over 8100 Cuda cores.
We're nearly double the 2080ti core count. Either 1600m2 12nm dies are a thing or 7nm doesn't stop at 500nm like the haters believe or their maths is just wrong and the die on 7nm is a lot smaller.

Find me which Turing card has 8100 Cuda cores in a single die? I cannot see any on the Nvidia stack doing over 4608 (RTX Titan/RTX Quadro 8000/6000)

You see haters all around but you are the one who ignore maths, reality and cannot make a single logical argument before you result to ad-hominem attacks.

 

[Rroff]

Did Panos really say PC gamers are peasants

hahaha this guy is ridiculous

What is ridiculous is when he trots out the same anti-nVidia spin every single generation redressed for that generation and is wrong every time - too hot, too big, unworkable, etc. yet it happens yet he comes out with the same lines next time

 

[tyler_jrb]

Open Steam survey..... Weep

We here are quite the minority. But that doesn’t mean that lots of people own decent hardware because they do. Nvidia/AMD sales are proof of that.

Not everyone wants to go out and buy a £600+ graphics card every generation. Same reason not everyone will want to spend £600 or so for a new console some will stick with what they have.

 

[scubes]

To me I want to see games tech demos are crap get a game on the console next to a high end pc.anyway new card are coming soon and they will be faster again it’s neve gonna end lol.

 

[Panos]

What is ridiculous is when he trots out the same anti-nVidia spin every single generation redressed for that generation and is wrong every time - too hot, too big, unworkable, etc. yet it happens yet he comes out with the same lines next time

Where is my spin for the Pascal generation? I owned both the 1080 & 1080Ti and you will find I actively promoted the Aorus 1080Ti Xtreme
Find where I was anti-NV on Maxwell or Kepler also.

Sure was critical about the NV drivers & bugs but who wasnt especially during Sept-Oct 2016? Or we brush under the carpet the Nvidia driver issues in here blaming everyone as anti-Nvidia if points to issues or god forbid the official forums?

Come back with constructive arguments to what I write. But the likes of you, Grim5 and Gregster cannot make an argument and respond. Instead all 3 repeatedly attack me in person, like I stole your lunch at school.

 

[Rroff]

Where is my spin for the Pascal generation? I owned both the 1080 & 1080Ti and you will find I actively promoted the Aorus 1080Ti Xtreme
Find where I was anti-NV on Maxwell or Kepler also.

Sure was critical about the NV drivers & bugs but who wasnt especially during Sept-Oct 2016? Or we brush under the carpet the Nvidia driver issues in here blaming everyone as anti-Nvidia if points to issues or god forbid the official forums?

Come back with constructive arguments to what I write. But the likes of you, Grim5 and Gregster cannot make an argument and respond. Instead all 3 repeatedly attack me in person, like I stole your lunch at school.

I seem to remember you weren't happy with the 1080 at all - lot of gripes about the boost behaviour and saying you'd never own nVidia again... but that is beside the point I'm talking about the pre-launch stuff where you persistently talk down anything nVidia have to offer regurgitating the same anti-nVidia rhetoric that is going around on a certain other forum basically verbatim even though it has been wrong year after year.

 

[pingu666]

Find me which Turing card has 8100 Cuda cores in a single die? I cannot see any on the Nvidia stack doing over 4608 (RTX Titan/RTX Quadro 8000/6000)

You see haters all around but you are the one who ignore maths, reality and cannot make a single logical argument before you result to ad-hominem attacks.

they where engineering samples for some server gpu chips, low clock but huuuge cuda core count

wouldnt surprise me if nvidia roll out a console killer, but may well want your kidneys

 

[CAT-THE-FIFTH]

The leaked commercial GPU might be an MCM of some kind,and its much easier to do with that with a non-gaming GPU,unless there is a way to present an MCM as a single one large GPU.

 

[doody]

@ panos ignore em panos your tech savvy brain and knowledge leaves them for dust

 

[FoxEye]

@ panos ignore em panos your tech savvy brain and knowledge leaves them for dust

AMD Superfriends Alt Accounts - assemble!