Question 'Ampere'/Next-gen gaming uarch speculation thread

[Ottonomous]

How much is the Samsung 7nm EUV process expected to provide in terms of gains?
How will the RTX components be scaled/developed?
Any major architectural enhancements expected?
Will VRAM be bumped to 16/12/12 for the top three?
Will there be further fragmentation in the lineup? (Keeping turing at cheaper prices, while offering 'beefed up RTX' options at the top?)
Will the top card be capable of >4K60, at least 90?
Would Nvidia ever consider an HBM implementation in the gaming lineup?
Will Nvidia introduce new proprietary technologies again?

Sorry if imprudent/uncalled for, just interested in the forum member's thoughts.

 

[Stuka87]

Samsung's 8LPP is a full node shrink from 14nm ( which has a slight density advantage over TSMC 16nm ).

Samsung 8LPP has a theorical density of ~62 MTr/mm^2 compared to ~29 MTr/mm^2 for 16nm of TSMC.

nVidia isn't moving from 14, they are moving from TSMC 12.

And as mentioned above, non-mobile GPU's never use max density, it kills performance. Its far better for performance and cooling to be less dense.

 

[Lodix]

nVidia isn't moving from 14, they are moving from TSMC 12.

And as mentioned above, non-mobile GPU's never use max density, it kills performance. Its far better for performance and cooling to be less dense.

The "12nm" that Nvidia is using is just a more efficient 16nm from TSMC with no improvements in density. So that is irrelevant.

The numbers that I posted are the maximal theorical densities of the nodes. Depending on the libraries you chose you trade performance for density. But that is a given to any process node and again irrelevant. It doesn't change the fact that 8LPP would bring a full node shrink (~2x density) from the previous generation of products.

 

[FaaR]

You do realize that if Nvidia will use high-density libraries, they will sacrifice the clock speeds and performance of their products?

Tone your expectations down.

Apple's A13 SoC clocks up to 2.66GHz it seems and it is denser than most anything else out there. So there seems to be potential for decent clocks even with a high density layout.

 

[sontin]

Even on 16 and 12 NM TSMC process was possible to pack 35 mln xTors/mm2. So why neither Nvidia, nor AMD achieved that transistor density, and the best they could come up with was 25 mln xTors/mm2?

I think you are smart enough to understand the difference between high density libraries that are used in ultra mobile chips, versus high-performance libraries used in high-performance chips.

You do realize that if Nvidia will use high-density libraries, they will sacrifice the clock speeds and performance of their products?

Tone your expectations down.

Apple's A10 had 26,4 mln xTors/mm2 on TSMC's 16nmFF: https://en.wikipedia.org/wiki/Apple_A10
GP104 had 22,9 which is 14% less dense.

So i guess nVidia will lose 14% again (compare A100 with Apple's SoC on TSMC 7nm DUV process).

 

[Hitman928]

The "12nm" that Nvidia is using is just a more efficient 16nm from TSMC with no improvements in density. So that is irrelevant.

The numbers that I posted are the maximal theorical densities of the nodes. Depending on the libraries you chose you trade performance for density. But that is a given to any process node and again irrelevant. It doesn't change the fact that 8LPP would bring a full node shrink (~2x density) from the previous generation of products.

According to TSMC, 12 nm has a tighter metal pitch which leads to a small density improvement. They don't give any hard numbers though so I can't say if it is of any real significance or not.

 

[Lodix]

According to TSMC, 12 nm has a tighter metal pitch which leads to a small density improvement. They don't give any hard numbers though so I can't say if it is of any real significance or not.

Yes, there was a 12nm "Compact" version dedicated for low end and cheaper products. As a low cost version of Finfet. But as far as I know Nvidia is not using this.

 

[Thala]

Lodix is right - you are getting about 2x density improvement when going from 12nm TSMC to 8nm Samsung.

 

[BenSkywalker]

I personally expect anywhere between 36-38 mln xTors/mm2

If we assume you are correct, the 2060 is actively using about 8B transistors which with that density level would be fractionally larger than the 1650 is right now. Honestly I'm expecting a bit better density then that even on the low end parts on SS 8nm, but even assuming you are right that still puts a 2060 level 3050 as the logical landing spot. I'm thinking it should be at least a bit faster, more than a bit with RT.

 

[maddie]

If we assume you are correct, the 2060 is actively using about 8B transistors which with that density level would be fractionally larger than the 1650 is right now. Honestly I'm expecting a bit better density then that even on the low end parts on SS 8nm, but even assuming you are right that still puts a 2060 level 3050 as the logical landing spot. I'm thinking it should be at least a bit faster, more than a bit with RT.

Concerning the part you are responding to. Who wrote that exactly? I think I speak for most in saying that we don't have the time to search for snippets of posts when the solution is readily available AND easy to use.

I know your stated reason for this. BUT. Is it possible to be less self centered and just use the reply function? We will all be more easily able to follow threads.

Of course, I'm assuming that you don't want to stuff up things. Then also, who knows, maybe you're leading a 1 man revolution to change our habits and protocols.

 

[nurturedhate]

Concerning the part you are responding to. Who wrote that exactly? I think I speak for most in saying that we don't have the time to search for snippets of posts when the solution is readily available AND easy to use.

I know your stated reason for this. BUT. Is it possible to be less self centered and just use the reply function? We will all be more easily able to follow threads.

Of course, I'm assuming that you don't want to stuff up things. Then also, who knows, maybe you're leading a 1 man revolution to change our habits and protocols.

They stated they are doing it on purpose. The end result being obfuscation and removal of context. We can all make of that what we will.

 

[Kenmitch]

I know your stated reason for this. BUT. Is it possible to be less self centered and just use the reply function? We will all be more easily able to follow threads.

Click > highlight>choose quote is all it takes. Most likely less effort then typing it in. Maybe his fingers need the exercise is why he takes the long way around?

 

[Stuka87]

Concerning the part you are responding to. Who wrote that exactly? I think I speak for most in saying that we don't have the time to search for snippets of posts when the solution is readily available AND easy to use.

I know your stated reason for this. BUT. Is it possible to be less self centered and just use the reply function? We will all be more easily able to follow threads.

Of course, I'm assuming that you don't want to stuff up things. Then also, who knows, maybe you're leading a 1 man revolution to change our habits and protocols.

I actually wish it was against the rules. Not only does it remove any context and make it harder to follow, but the person being quoted it not notified that they have been quoted, so they are unlikely to give a rebuttal.

 

[BenSkywalker]

This is a tech forum, not Twitter. If you read the actual discussion, why do you need headers for who said what? Furthermore, why should it EVER matter who made the statement?

The contrarian tribalistic mentality so frequently displayed on these forums would be significantly abated if every post was anonymous.

Said another way, if you care more about the poster then the post and don't want to bother reading the thread put me on ignore, you have nothing to contribute to the conversation at hand anyway.

 

[maddie]

No wonder the elites desire AI so badly.

 

[Bouowmx]

A100 tested in MLPerf 0.7. See 8xA100 results 0.7-18, 19, and 20, and compare with 8xV100 results 39, 40, 41. In general, about half the time, or twice the speed.

Of course, not implying raster or standard compute (CUDA cores only) performance, but I do wish for ~1.5-1.7x performance in the same tier (GeForce RTX 2070 -> 3070).

 

[jpiniero]

Of course, not implying raster or standard compute (CUDA cores only) performance, but I do wish for ~1.5-1.7x performance in the same tier (GeForce RTX 2070 -> 3070).

At 2944, the 3070 probably won't be 50% faster than the 2070.

 

[Tup3x]

This is a tech forum, not Twitter. If you read the actual discussion, why do you need headers for who said what? Furthermore, why should it EVER matter who made the statement.

It makes discussion hard. Why don't you see that? On top of that one cannot easily check the context because quote hyperlink is not working. That too makes discussion harder. You are the first person that I've seen who does something as ridiculous as this. Even puts effort to make it harder for others to continue the discussion... Come on man.

You are replying to people not some random quotes. Your reasoning doesn't make any sense.

Read the thread.

Anyone who is too lazy to read the thread they are posting in please put me on ignore. If I wanted to partake in sound bite stupidity I'd go the Twitter route.

Reading a thread and having a decent conversation isn't done by focusing on the person talking, but the merit of the idea and the perspective it brings, then counter perspective it offers from others.

You want a personal discussion, there are forums for that, this is a tech forum.

What kind on nonsense is this? Here especially one should write the source of news piece or person they are quoting/referring.

The problem is that we have no idea if you just made up those quotes or quote something that doesn't even exist on these forums. It's a mess if the quoted message is not on the same page. I don't like to waste my time to play Sherlock Holmes trying. Quoting post properly is so ridiculously easy that no amount of excuses can justify not using those quoting features.

 

[BenSkywalker]

Read the thread.

Anyone who is too lazy to read the thread they are posting in please put me on ignore. If I wanted to partake in sound bite stupidity I'd go the Twitter route.

Reading a thread and having a decent conversation isn't done by focusing on the person talking, but the merit of the idea and the perspective it brings, then counter perspective it offers from others.

You want a personal discussion, there are forums for that, this is a tech forum.

 

[FaaR]

This is a tech forum, not Twitter. If you read the actual discussion, why do you need headers for who said what? Furthermore, why should it EVER matter who made the statement?

Not everyone lives on forums and hold entire discussion threads dozens of pages long in their heads. Proper quoting (which is NOT difficult to do) shows the flow of replies and provides a link back to the full reply, for greater context. It's stupid this should even have to be explained to you, since it's friggin' self-evident.

And, of course it matters who made a statement. People are individuals, not a faceless mass of thoughts and ideas. What an utterly ridiculous thing to say.

Now stop being obstructionist on purpose. Just show people some common courtesy and do what they ask you.

 

[BenSkywalker]

9,000 posts over the course of 21 years is living on a forum....

This is a tech forum, why would it matter who posted areal density figures as long as they were accurate? Explain why that would matter.

If I quoted someone out of context, which I haven't seen so much as implied, I'll correct it. If people are too lazy to read a thread and then expect a recap because they are oh so special, not my problem.

There's an ignore function, why don't you do as you're told and use it?

 

[JujuFish]

9,000 posts over the course of 21 years is living on a forum....

This is a tech forum, why would it matter who posted areal density figures as long as they were accurate? Explain why that would matter.

If I quoted someone out of context, which I haven't seen so much as implied, I'll correct it. If people are too lazy to read a thread and then expect a recap because they are oh so special, not my problem.

There's an ignore function, why don't you do as you're told and use it?

You couldn't even follow the single line of conversation we were having, and you're complaining about other people not reading the thread? GTFO.

 

[BenSkywalker]

What claim did I make?

I'm asking what piece of information did I offer up that was not linked to in this thread? I'm not even talking about some silly Twitter user posting some story they thought up, only links from valid news sources of course.

You never replied. There is an ignore feature for everyone saying reading is too hard.

Edit- Dropping further "reading is too hard" commenters into ignore for a few years.

 

[Krteq]

Stop being ***** guys and stop this nonsense.

Back to Ampere discussion, shall we?

 

[BenSkywalker]

Certainly.

The latest rumor about RDNA2 has to be some ploy, or just flat out BS, to get nVidia to sleep at the wheel, 20% faster than 2080Ti would be a catastrophic failure. I'm not nearly as high on AMD's prospects as others but fourth tier performance after all this hype? Can't be that bad.

 

[itsmydamnation]

You never replied. There is an ignore feature for everyone saying reading is too hard.

Are you talking to yourself?

Seriously is it that hard to quote properly?

How do you expect to be taken seriously when you cant even quote?