AMD Greenland Vega 10 said to have 4096 stream processors

[Corky34]

Some may say that, but the person who matters - NVidia's CEO - didn't: he said it was actual Pascal silicon. In the same way he said he was holding an actual Fermi card a few years earlier....

I guess selectively quoting only part of what i said serves what ever purpose you wanted, but i feel it remiss that you failed or choose to ignore that the rest of what i said as that would bring some context to what you chose to quote...

seeing as we're dealing with rumors and speculation though I'm not going to damn or praise any company based on conjecture.

And FYI NVidia's CEO is a she not a he.

[scaryjim]

... i feel it remiss that you failed or choose to ignore that the rest of what i said as that would bring some context to what you chose to quote...

If that was relevant I'd've quoted it. What Jen-Hsun Huang did in January is not rumour or speculation. It's been reported by the majority of the tech press as fact, and if they were wrong they would've been sued for libel by now. He showed Maxwell silicon and claimed it was Pascal. So I'm not damning nvidia based on conjecture, I'm damning them for lying to the press and investors. Again.

And FYI NVidia's CEO is a she not a he.

I can't help feeling that you're confusing Jen-Hsun Huang and Lisa Su.

I'd say it's an easy mistake to make, but, really, it isn't...

[Corky34]

If that was relevant I'd've quoted it. What Jen-Hsun Huang did in January is not rumour or speculation. It's been reported by the majority of the tech press as fact, and if they were wrong they would've been sued for libel by now. He showed Maxwell silicon and claimed it was Pascal. So I'm not damning nvidia based on conjecture, I'm damning them for lying to the press and investors. Again.

It was relevant you partially quoted what i said because it best served your attempt at trolling.

I can't help feeling that you're confusing Jen-Hsun Huang and Lisa Su.

I'd say it's an easy mistake to make, but, really, it isn't...

My apologies, although I'm not sure why I'm apologising as it seems your more interested in starting a flaming session than discussing rumors, speculation, and conjecture.

[scaryjim]

... I'm not sure why I'm apologising as it seems your more interested in starting a flaming session than discussing rumors, speculation, and conjecture.

I've yet to see a single flame in this thread (although you're getting close, IMSNHO). I also don't see how pointing out an established fact counts as trolling, unless you think reality is trying to troll you as well? The nearest this thread has come to a genuine troll is actually Jen-Hsun Huang's claim that he was waving Pascal on stage when it was actually Maxwell!

I've simply pointed out that some thing are rumour and speculation - like released dates for Pascal, chip layout for Vega, whether nvidia will seed sympathetic press with early Pascal samples and ask for particular benchmarks to be use - while others are established fact.

I did point out previously nvidia may have had a variety of reasons for showing non-Pascal boards: they might be concerned about revealing business-sensitive information, they might be trolling AMD by implying they're having problems with Pascal, or Pascal might genuinely not be ready yet. We don't know for sure what the reason behind using Maxwell on stage in January was. That might mean that Pascal is actually coming along well and will launch alongside, or even before, Polaris.

But NVidia could've said they were using Maxwell to demo the PX2 board. Instead that said they were using Pascal, when they clearly weren't. So whether some people might say it was just a mock-up so they had something to show is irrelevant. Their CEO got up on stage, in a press briefing, and lied about a product. In my book that's just out and out shady.

[CAT-THE-FIFTH]

These are the facts on the ground:
1.)AMD has showed off both Polaris 10 and 11 running games briefly. Polaris 11 die has been shown to journalists.
2.)Lisa Su is on record that AMD is aiming for a mid year launch in time for the back to school period. The most important months for back to school in the US is July and August according to market research carried out by Google and IPSOS MORI.
3.)JHH showed of a Drive PX2 module which was meant to have Pascal chips but people found out that it only had placeholder chips.
4.)Drive PX2 will launch at some time in 2016.

That is all the official stuff we know now.

[DanceswithUnix]

Interesting take on it i guess, some may say it was just a mockup so they had something to wave at the crowd, seeing as we're dealing with rumors and speculation though I'm not going to damn or praise any company based on conjecture.

Charlie hates Nvidia, so ignore the tone of his articles. He does however actually bother to check his facts before publishing, so you can put more faith in stated facts than on most internet sites (though it is the internet, so always apply at least one pinch of salt).

What Jen was waving looked like something rather special in its own right, which makes lying about it all the sillier. But it wasn't Pascal, that was a fib, plain and simple.

[CAT-THE-FIFTH]

If you watch the video he mentions the board as having "two next generation Pascal GPUs" twice. I made sure the video starts from the point he holds the board and shows it to the tech press.

[Biscuit]

It just started at the beginning for me but Im guessing it should have been ~@20 mins in

[CAT-THE-FIFTH]

It just started at the beginning for me but Im guessing it should have been ~@20 mins in

It should be timestamped from 22 minutes onwards.

[kalniel]

It is timestamped but that's not supported by the embed coding. Just follow the link:

https://www.youtube.com/watch?v=Md6K...outu.be&t=1310

[scaryjim]

Makes you wonder what silicon they used for the other processors...

[Corky34]

I've yet to see a single flame in this thread (although you're getting close, IMSNHO). I also don't see how pointing out an established fact counts as trolling, unless you think reality is trying to troll you as well? The nearest this thread has come to a genuine troll is actually Jen-Hsun Huang's claim that he was waving Pascal on stage when it was actually Maxwell!

Did i say you flamed anyone? No, i said "you seem more interested in starting a flaming session", also what does pointing out a "fact" have to do with future rumors? Other than giving you a platform to launch an attack on someone.

TBH i don't care what's waved around on stage, if something is a mockup, if something is show playing a game or any of the other rubbish you're probably going to drag up in some vain attempt to win the Internet, what i care about is what each company has said, the rumors, speculation, and conjecture, you can spend all day arguing how X or Y isn't to be believed, or trying to prove and disprove whatever you like, personally though I'm more interested in discussing the possible implications of opinion or theories based on rumors and speculation.

I've simply pointed out that some thing are rumour and speculation - like released dates for Pascal, chip layout for Vega, whether nvidia will seed sympathetic press with early Pascal samples and ask for particular benchmarks to be use - while others are established fact.

I did point out previously nvidia may have had a variety of reasons for showing non-Pascal boards: they might be concerned about revealing business-sensitive information, they might be trolling AMD by implying they're having problems with Pascal, or Pascal might genuinely not be ready yet. We don't know for sure what the reason behind using Maxwell on stage in January was. That might mean that Pascal is actually coming along well and will launch alongside, or even before, Polaris.

But NVidia could've said they were using Maxwell to demo the PX2 board. Instead that said they were using Pascal, when they clearly weren't. So whether some people might say it was just a mock-up so they had something to show is irrelevant. Their CEO got up on stage, in a press briefing, and lied about a product. In my book that's just out and out shady.

Yea because it's not like any other company has ever "lied" to the press is it.

Charlie hates Nvidia, so ignore the tone of his articles. He does however actually bother to check his facts before publishing, so you can put more faith in stated facts than on most internet sites (though it is the internet, so always apply at least one pinch of salt).

What Jen was waving looked like something rather special in its own right, which makes lying about it all the sillier. But it wasn't Pascal, that was a fib, plain and simple.

I'm not calling his facts into question, what I'm saying is that just because a company said this is an apple and it turns out to be an orange doesn't mean their not going to have actual apples come harvest time, that at the end of the day has the apple farmers managed to provide me with the apples i wanted when he said he would.

[kalniel]

That's fine - you might be fine with it, but other people may think differently and take subsequent claims from a dishonest individual with less sincerity than from someone call calls oranges oranges rather than apples. You have every right to believe them next time though.

[DanceswithUnix]

That's fine - you might be fine with it, but other people may think differently and take subsequent claims from a dishonest individual with less sincerity than from someone call calls oranges oranges rather than apples. You have every right to believe them next time though.

Indeed, I like a bit of razzmatazz and showmanship, but being just plain lied to is mildly insulting.

Just not necessary either. He knows there will be apples, we know there will be apples, so why paint an orange green and poke a stick in the top? Nvidia is an impressive company full of impressive people doing impressive stuff, if the CEO can't make a good show out of that then that's just sad.

Anyway, back to what Corky was asking: The rumours are that HBM2 isn't around until very late this year, which implies that stuff coming out in the summer will be not much faster than Fury/980ti if at all. That gets some time and experience on the 14/16nm process, which makes the big dies more cost effective later, and that is when we should see the monster graphics cards turning up.

[scaryjim]

... what I'm saying is that just because a company said this is an apple and it turns out to be an orange doesn't mean their not going to have actual apples come harvest time, that at the end of the day has the apple farmers managed to provide me with the apples i wanted when he said he would.

That's what you're saying now. And I've agreed with you. You've even quoted me agreeing with you. We don't know what nvidia will turn out or when it'll be released. I'm genuinely looking forward to finding out.

But let's go back to where this exchange started shall we? You pointed to rumours that Pascal will be ready in May, which were countered by the fact that nvidia lied* about having Pascal silicon on a Drive PX2 module in January.

Here's the thing with rumour and speculation: you can subject them to test by known facts and see if they ring true. Known fact: nvidia didn't show any real Pascal silicon at CES. There's two possible reasons for that: that they didn't have any to show, or that they chose not to show their new architecture to hide their progress from their rivals.

The problem for nvidia is that a May launch - a proper hard launch, not a paper "here's the specs of our new card" launch - would require products in volume shipping around the world pretty much now. They'd've shipped cards to partners for branding. They'd've had to ship GPUs to card factories for assembly. There's all sorts of evidence that would've been out there that nvidia were mobilising for a big product launch.

There's been nothing.

Of course, you can't prove a negative, and it's just possible that nvidia are currently running the best undercover operation to launch a new GPU series that we've ever seen. But I'm an Occam's razor guy; between competing theories, the one that requires the least assumptions should be preferred. Nvidia showed no Pascal silicon at CES, and there's been no evidence of anything Pascal based being anywhere near ready in any of the usual leek sources. Least assumptions? Pascal isn't even nearly ready yet, and a May hard launch shifts beyond speculation into the realm of fantasy.

But perhaps nvidia will prove me wrong.

[Corky34]

That's fine - you might be fine with it, but other people may think differently and take subsequent claims from a dishonest individual with less sincerity than from someone call calls oranges oranges rather than apples. You have every right to believe them next time though.

But whether they did or didn't wave around a real Pascal GPU on stage, or whether you're shown an orange instead of an apple, none of that changes what the company, rumors, or speculations indicate, it hasn't changed anything, sure there could be many reason why they done it but those reasons would be worse than guesswork.

Anyway, back to what Corky was asking: The rumours are that HBM2 isn't around until very late this year, which implies that stuff coming out in the summer will be not much faster than Fury/980ti if at all. That gets some time and experience on the 14/16nm process, which makes the big dies more cost effective later, and that is when we should see the monster graphics cards turning up.

That's not what i was expecting, i expected AMD's head start with HBM1 to give them a leg up with a HBM2 GPU, never know maybe it will turn out that way but going on the info around ATM it seems they may both release HBM2 cards around the same time.

That's what you're saying now. And I've agreed with you. You've even quoted me agreeing with you. We don't know what nvidia will turn out or when it'll be released. I'm genuinely looking forward to finding out.

Well the rumors seem to say both will hit the market with HBM2 cards around the same time.

But let's go back to where this exchange started shall we? You pointed to rumours that Pascal will be ready in May, which were countered by the fact that nvidia lied* about having Pascal silicon on a Drive PX2 module in January.

But that lie hasn't changed the rumored May release date has it.

The problem for nvidia is that a May launch - a proper hard launch, not a paper "here's the specs of our new card" launch - would require products in volume shipping around the world pretty much now. They'd've shipped cards to partners for branding. They'd've had to ship GPUs to card factories for assembly. There's all sorts of evidence that would've been out there that nvidia were mobilising for a big product launch.

Well based on the link i posted on the other page it probably is going to be a paper launch as in the article they say...

NVIDIA GeForce GTX 1080 (GTX 1800) would launch on May 27th, which is just 4 days before Computex, allowing AIB partners to show their cards at the event.

Can't see AIB partners doing much in the way of shipping products in the space of 4 days.