>amd polaris NDA date confirmed

...

Other urls found in this thread:

techgrapple.com/apple-a9-14nm-vs-16nm-performance-test-with-iphone-6s/
tomshardware.com/news/iphone-6s-a9-samsung-vs-tsmc,30306.html
anandtech.com/show/9886/amd-reveals-polaris-gpu-architecture
videocardz.com/60686/watch-amd-computex-polaris-update-live-stream-here
amd.com/en-us/who-we-are/corporate-information/events/computex
twitter.com/SFWRedditImages

>June 29th
RIP in peperoni AMD

The only reason i can think of them keeping it under wraps until then is they have some kind of huge partner that greased their palms to keep it under wraps until their new product was out.

Apple maybe?

Polaris xfire would be perfect for the iMac, could be.

>Thinks Nvidia releasing a card earlier is better
>Doesn't realize that this lets AMD confirm they have better tech

>AMD: This is all under NDA
>Pajeet whips out DSLR with gigantic flash unit.

But they'd lose lots of potential customers who buy their cards on day 1?

NDA=/=no press taking down info/photos
they will have to hold it for the NDA though

I like how they left the password there too

I guarantee you that AMD saw final silicon of Pascal within weeks of the wafers being sent out to get packaged. Same with Nvidia getting AMD silicon.
There are no secrets between companies when it comes to finished products. They each know what the other has.

So they lose the people who don't deserve better value for dollar anyway. If you know the competitor is releasing a new tech soon, and your current tech isn't literally unusable, why not just wait another month to make sure you're making the best choice?

oh no they'll get to use amd's guest wifi at computex!

Journalists who got Pascal review samples weren't allowed to talk about it either before the official embargo ended.
This isn't unusual.

>I guarantee you that AMD saw final silicon of Pascal within weeks of the wafers being sent out to get packaged. Same with Nvidia getting AMD silicon.
>There are no secrets between companies when it comes to finished products. They each know what the other has.

Yeah, I bet it's really easy to smuggle out 300mm wafers from the fab.

Maybe, but if that gets unveiled at WWDC like you might expect then its June 13-17th and does not fit well with that 29th NDA date.

I do think they might be holding off for a volume launch though I question how much doing this really makes sense when it would mean NVIDIA beating them to the punch by a whole month.

It's fucking retarded to build a high end PC with the newest released graphics card to play video games when all you need to buy is a fucking PS4 and every game for it is optimized for the hardware.

According to rumours some OEMs are pissed at AMD because some of their products are packing Polaris and won't be able to be shown at computex.

The best time for AMD to do a reveal would be day 1 of Computex, meaning their partners can show off their hardware. (to be honest the OEMs might do backroom shows, that'll get us a lot of "leaks"), so...

You're suspiciously poor at reading.
Try again.

AMD said those rumors were false

>so they lose people who don't deserve
YOU ARE A FUCKING IDIOT HAHAHAHAHA FUCK YOU YOU HAVE NO IDEA HOW BUSINESS WORKS THEY NEED EVERY FUCKING SALE TO FINALLY TURN OVER A PROFIT YOU FUCKING AUTISTIC SACK OF SHIT WHO CARES IF ANYONE DESERVES AMD OR NVIDIA THEY SHOULD FOCUS ON THE FUCKING SALES

Erin

1. AMD was surprised about the very fast release of Nvidia Pascal.

2. The SAMSUNG 14nm is hotter and not mature like the TSMC 16 nm.

3. The Star Wars Battlefront 60fps 'polaris 10' demo was an unclocked Fury Nano.

?

So you're saying that patient informed consumers aren't a market that is desirable? You'd rather your company market to the impulse buyers and the "new is always better" type?

>unclocke
*under

>OEM leaks Pascal info to AMD
>Nvidia finds out and gets fucking pissed
>No more business for them

Honestly, I don't think it's worth it for them.

Why do you idiots just say shit like this?
Samsung had their 14nm lines up and producing Exynos 7420 chips at the end of 2014. They were in volume production before TSMC started risk production on 16nm FinFET.

The Battlefront demo was Polaris 11, drawing under 50w, and it was compared to a GTX 950.
If you think Fury at 850mhz~ only draws

>Rumors I believe because they make my favorite gpu sports team look good

we're talking about gaymers here user

He is right...

techgrapple.com/apple-a9-14nm-vs-16nm-performance-test-with-iphone-6s/

>The Samsung Made A9 Chip is hotter than TSMC made; Samsung at 39.3 degree temperature and TSMC at 35 degree. Chip based on 14nm FinFET tend to get heat quicker than 16nm

They each pay fab workers to sneak samples out to one another. Even Charlie from S|A has enough contacts and cash to pay for insider info, thats how he got so much info about AMD's console APUs a full year before launch.
Corporate espionage is so common place that its expected.

Where is Polaris?

No, hes not, and you're just as stupid as he is.
tomshardware.com/news/iphone-6s-a9-samsung-vs-tsmc,30306.html

The only notable difference between the chips came from a single bench, Geekbench. It was quickly found to be an issue with Geekbench itself.

There is no real statistical difference between A9s from either vendor because they're each tuned to the same nominal power consumption and clocks. Neither is running at their max clock for their given binning. Apple as a manufacturer would never have two totally different performing models sold under the exact same SKU.

>Fury at 850mhz~
Lower.
Nice point user

Every consumer should be the market and they shouldn't keep their cards secret for risk of losing customers to nvidia. They've already lost a lot of market share

I'll be honest here.

I've only ever had a PC for gayming. I can't even code html. I have no idea what Linux even really is other than an OS. But I've always been someone who waits to see what the big picture is and compares it to my needs before making a decision.

Sure, I could have dropped ~2-3k on a maxed PC easily. But I only really play old school RPG (Divinity: Original sin, Baldurs Gate, Pillars of Eternity, etc...) and the occasional MOBA.

It would be like buying an F-450 V12 twin turbo super duty to haul your kids to soccer practice twice a week.

It's the 1st time samsung make big GPUs like that with their wafer, TSMC is more advanced and where is the Polaris PCB user? :3

Of people who feel that they need to upgrade this instant.

Anyone patient hasn't been lost.

Even had this been true, as points out, it still wouldn't be a good comparison.

The difference there is that was Samsung/GloFo's 14nm LPE or Low Power Early process vs TSMC's 16nm FinFET process. Polaris and later Zen are being made on the LPP or Low Power Plus process that according to the fabs uses 15% less power and gets 15% higher clocks than LPE.

>Overpaying for a card that will be obsolete in 3 months.
: ^ }

No, retard.
Polaris 11 in the demo was clocked at 850mhz, and the vcore was set at 0.8375v. Hardware.fr and Ryan from Anandtech both reported on this. They all saw the die that was being used in the GPU.

Words cannot describe just how dumb you are.

>June 29th

WHAHAHAHAHAHAHAHAHAHA LE OVERHYPPPPPPED GAMMMICK SHIT LIKE HBM1 AND FURY X

That's my point you fucking retard.

>850mhz, and the vcore was set at 0.8375v.
with a load cap via costum drivers

>They all saw the die
citation

>Overpaying for a card that will burn down your house.

>I was wrong about literally everything so far
>better keep talking out of my ass
>m-muh custom drivers!

Jesus you're pathetic.
anandtech.com/show/9886/amd-reveals-polaris-gpu-architecture
In any case, the GPU RTG showed off was a small GPU. And while Raja’s hand is hardly a scientifically accurate basis for size comparisons, if I had to guess I would wager it’s a bit smaller than RTG’s 28nm Cape Verde GPU or NVIDIA’s GK107 GPU, which is to say that it’s likely smaller than 120mm2

I'd like to see you try and run a nano at that frequency and voltage. It's not possible.

Trolling is a bannable offense

...

I can't even tell what genger this bean is supposed to be...

Jesus fucking christ you're retarded. Samsung still isn't making "big GPUs". Global Foundries is licensing tech from Samsung in order to make GPUs for AMD. Read that again since you're obviously too stupid to have understood it the first time.

Global Foundries is making the Polaris chips. GLOBAL FOUNDRIES.

As for the linked article, it clearly shows that Samsung's 14nm LPE FF is comparable or even slightly better than TSMC's 16nm FF.

The tech that GloFo is licensing from Samsung is 14nm LPP FF, which is better than the 14nm LPE that was already comparable to TSMC's 16nm being used in Nvidia's Pascal.

What this means is that TSMC 16nm FF used in the new Nvidia chips is cheap chinkshit. The Samsung 14nm LPP being used by GloFo to make AMD's new chips is slightly better gookshit. Meanwhile, Intel's 14nm being used on their most recent Skylake CPUs is still miles ahead of both of them.

> Meanwhile, Intel's 14nm being used on their most recent Skylake CPUs is still miles ahead of both of them.

Intel has superior density and area scaling. There is as of yet nothing to indicate they have superior electrostatic characteristics. Intel has had nothing but issues on their 14nm node, and it took an inordinately long period for them to build up enough inventory of high binned silicon to release desktop Skylake.

Broadwell CoreM was a disaster at launch.
Desktop Broadwell was delayed several times, only two SKUs ever launched, and they were pulled off the market quickly. Only extremely limited numbers were ever available.
The first Skylake chips out were only low TDP mobile parts
Desktop Skylake was limited availability for months after launch because intel still wasn't yielding enough high clocking chips to meet demand.

Samsung has been endlessly shitting out 14nm parts with die sizes rivaling CoreM since the end of 2014, and they never had a hiccup. Area scaling and density go out the window when it comes down to per transistor performance and process yield. Intel paid a heavy price for having the smallest BEOL.

It's over, AMD is finished.

They don't provide these new open source linux drivers for my 7870, so I'm just going to jump on the next Intel GPU with the best performance soon.

Will be more efficient, too.

Fake news, it conveniently came out right as Pascal was released but AMD said it's not true

I love when Americucks talk about cars and they always use a shit car as an example, muh corvette muh dodge muh big pickup.

Americans only buy shit cars though, it's all we know.

when a ps4 can hit 60fps on everything or let me turn graphics down to hit 1080 60, then we are talking.

once they let me use a mouse keyboard with an fps, then ill fully move to consoles again.

they lost sub 5% of potential sales, and with how little 1080's and 1070s will be out for a long time, probably sub 1%

no, its all we have that are cheap.
remember there is a massive fuck off import tax on cars, it use to be FAR cheaper to buy a car from overseas and have it delivered to you then going to a dealership.

then the dealerships and auto industry lobbied/bitched till laws were put in place to make these cars illegal.

more like what googled did for mozilla in the day, they send them a cake every new release.

i imagine that amd has sent intel a dev version of their zen just as a fuck you, along with nvidia a polaris just so they know how much they will get stomped.

when vega is tapped, they will be the second to get it.

Wow, those brush strokes are huge.

if they have cards, the embargo is for benching purposes, so everyone is as thorough as they can be and no one is rushing to be first.

Spoken like a true gamer.

TSMC has been sending 10nm units to Nvidia for quite a while now user.

LPP fixed that

Tell that to Sony. Xbox One released early and it was a catastrophic fucking failure

>founder's edition
kek

What new linux drivers are you talking about? I haven't been followig AMD news lately.

Tesla senpai.

Tesla.

Each time AMD blew hot air before launch it always ended in failure.
I prefer them keeping silent, shows more promise.

Not really, Llano was fantastic when it launched. They were talking about it nonstop well over a year before launch too.

There was a lot of hype surrounding the original athlon 64s, a few of my friends and I waited up all night to go out and buy our chips in the morning, I still keep my 3500+ in my desk drawer

It's the NDA for the tech day and not the NDA for Polaris.

Intel and Samsung are rumored to have working 7nm silicon. What does any of this have to do with Polaris?

The Athlon 64s are slower than Bulldozers.

Consoles at e3

>AMD is the one to stop consoles from holding back gaming.
They always seem to be the company forcing the others to get off their asses.

>polaris not competing with 1080
>not even competing with 980
>m-muh power efficiency

...

The fact that they are going to be trashed by Nvidia, the 1080 is just a 1/3 of the big Pascal chipset.

It's actually 52% of GP100

this and regarding polaris itself it's also either/or - the 29th or official launch date.

700 bucks for half a piece of chink shit.
we're doomed.

When is the Computex event for AMD? In a few hours?

videocardz.com/60686/watch-amd-computex-polaris-update-live-stream-here

there's a counter

No point in watching, they're only showing desktop Carrizo(Bristol Ridge) and they're barely teasing Polaris.

The webcast is 10pm est today.
amd.com/en-us/who-we-are/corporate-information/events/computex

So when is it?

>deserve
They lose money you fucking idiot
It's not about which customer 'deserves' what you fucking moron

Jesus crust

Do we have any idea at all of how fast Polaris 11 is? I've heard a ton about Polaris 10's performance but nothing on the 11.

5 times slower than a gtx950

What bugs me most is that we don't have transistor count for either Polaris yet.

They gave a live demo of Polaris 11 months ago. It was mentioned just a bit higher up in the thread.

AMD showed a PC with a Polaris 11 GPU running Star Wars Battlefront. It was running at 1920x1080, medium settings with the frame rate pinned at 60fps, and total system power consumption was under 90w. The same system with a GTX 950 was drawing 140w for the same frame rate.

anandtech.com/show/9886/amd-reveals-polaris-gpu-architecture

They'll have superior transistor density since since GF's 14nm gate pitch (I think that's the right measurement) is very slightly shorter than TSMC's 16, but that doesn't necessarily mean the process is overall better.

So the GPU uses 30 fucking watts? What the fuck?
CPU, mobo, fans, HDD at least would take up 60 watts.

Foundry characteristics aside, from transistor count you can make a good guess on the performance of the chip.

An i3-6100+DDR4+SSD system would sip power.

AMD had a quad core i5 in there IIRC, so the TDP of the CPU isn't under 60.

The CPU wasn't anywhere near fully loaded, so the system sans GPU was probably about 40w.
The Polaris 11 GPU shown is probably a 50w~ card, the frame rate being capped was likely saving each card a little power too.

Connected poly pitch, SRAM cell size, and a few other things are in Samsung's advantage for area scaling. Dense logic can have up to 15% area advantage vs TSMC's 16nm process.

TDP is not power consumption.

TDP != Power consumption.

It's not, a 60W TDP CPU won't be using 20 watts when all its cores are running at a decent clock.

>FE