AMD said to have secured priority access to SK Hynix's HBM2 chips

Quote:

---

Could it put the brakes on Nvidia's Pascal plans in 2016?

---

Read more.

This could be quite a coup for AMD. Combined with the experience their engineers will already have with working on HBM1 it could give them quite an advantage over Nvidia in the next gen of cards.

One can hope at least.

Sound like a double edged sword, does AMD hoover up as much HBM2 as they can in the hope that they can shift plenty of GPUs and hopefully limit Nvidia's supply, running the risk of having unsold stock.

Or is something like memory production not done on a pre-order basis?

They probably only have to take enough stock that nvidia couldn't launch a new gpu line using it, since they would never go to the expense of having 1 specific card using it and all others on the old ram system.

given that HBM is in limited supply - holding up the Fury range - contracting to buy everything from1 supplier until they dont need anymore would be the Apple way of doing it.

Could be AMDs final bad move if they try and leverage that to stifle competition. Do they have the funds to buy it all up and let it sit in a warehouse gathering dust?

After some of the comments about nVidias business practices recently, not one of them come close to trying to prevent a competitor from making a new product like this seems to imply.

I guess this could get interesting and may force nVidia to throw it's weight around a bit more.

Quote:

---

Originally Posted by Corky34

Sound like a double edged sword, does AMD hoover up as much HBM2 as they can in the hope that they can shift plenty of GPUs and hopefully limit Nvidia's supply, running the risk of having unsold stock.

Or is something like memory production not done on a pre-order basis?

---

If you want 10 chips for a prototype run then that is from stock, but 40000 of them to make a 10000 graphics cards productions run?

Most chip production is done on a "just in time" basis as no-one wants to hold large stocks. Factories order the chips something like 3 to 9 months in advance. If you demand chips right now, then that is someone else's allocation they are taking them from to try and get you supply.

Quote:

---

Originally Posted by shaithis

Could be AMDs final bad move if they try and leverage that to stifle competition. Do they have the funds to buy it all up and let it sit in a warehouse gathering dust?

After some of the comments about nVidias business practices recently, not one of them come close to trying to prevent a competitor from making a new product like this seems to imply.

I guess this could get interesting and may force nVidia to throw it's weight around a bit more.

---

If supply is constrained, then someone is going to miss out. If AMD are currently helping Hynix to develop and sell HBM1, then a bit of mutual back scratching seems reasonable when it comes to HBM2.

Or to invert it, I don't see why Nvidia should think it can turn up late to the party and help itself to the drinks cabinet.

It is in the interest of SK Hynix to make enough of these devices for anyone that wants them, so I don't see the situation lasting too long.

Yeah a lot of this seems to be speculation on top of rumour with a bit of misunderstanding thrown in.

It's possible that, since HBM is indeed co-developed by AMD and Hynix that AMD would have priority access in one way or another. It doesn't make sense that a party would put years of R&D into development only to have a competitor benefit equally or greater from it at only purchase cost.

It also wouldn't make business sense to just buy up capacity and not use it, but that's not what's being said by the source article anyway. As DanceswthUnix said, semiconductors are very long lead-time parts, you don't come along last minute and expect the manufacturer to have a few million in stock waiting for you. The manufacturer will need to know well in advance what you need so they can allocate capacity.

AMD wouldn't in any way be 'preventing' Nvidia from making anything. If HBM had not been developed by AMD/Hynix then I doubt Nvidia would have just given up producing GPUs. In fact, the pictures of 'Pascal' prototypes I've seen around the web are not HBM (note no interposer). Going back a few years Nvidia were planning to use HMC if memory serves correctly but decided to switch to HBM more recently - it would be pretty bad business to switch to a competitor's product and simply assume there would be capacity if they hadn't discussed it with the involved parties well in advance.

If anything the 'priority' is probably something like AMD is guaranteed capacity since they invested in it, similar to the WSA with GloFo.

Yeah, bandwidth is nothing without a nice GPU. We saw that with the Fiji. No gain whatsoever. Maybe next time.

Quote:

---

Originally Posted by shaithis

Could be AMDs final bad move if they try and leverage that to stifle competition. Do they have the funds to buy it all up and let it sit in a warehouse gathering dust?

After some of the comments about nVidias business practices recently, not one of them come close to trying to prevent a competitor from making a new product like this seems to imply.

I guess this could get interesting and may force nVidia to throw it's weight around a bit more.

---

I'm hoping the AMD Arctic Islands GPUs will actually be good along with being paired with the HBM2.. I have the distinct feeling Nvidia's Pascal and Volta will be absolute monsters when paired with HBM2 in time.

Quote:

---

Originally Posted by watercooled

It also wouldn't make business sense to just buy up capacity and not use it, but that's not what's being said by the source article anyway. As DanceswthUnix said, semiconductors are very long lead-time parts, you don't come along last minute and expect the manufacturer to have a few million in stock waiting for you. The manufacturer will need to know well in advance what you need so they can allocate capacity.

---

But nVidia have been talking about Pascal with HBM for a while now.....and I doubt they aren't aware of lead-times......so something isn't right....Maybe it's the article itself?

Quote:

---

Originally Posted by shaithis

But nVidia have been talking about Pascal with HBM for a while now.....and I doubt they aren't aware of lead-times......so something isn't right....Maybe it's the article itself?

---

Talking things up is a skill of theirs. I expect to see Pascal boards held out for all to see which turn out to have wooden HBM chips screwed into place ;)

Seriously though, I think Fury proves that the current generation isn't generally bandwidth constrained. Perhaps the next high end (will that be Titan Y?) will need HBM to feed double the shaders of a Titan X, but Pascal mid range can probably stay on DDR5 without performance problems.

OTOH, if AMD are betting the farm by putting their whole range on HBM then things could get interesting.

Quote:

---

Originally Posted by shaithis

But nVidia have been talking about Pascal with HBM for a while now.....and I doubt they aren't aware of lead-times......so something isn't right....Maybe it's the article itself?

---

Most likely. I find it extremely hard to believe Nvidia would be caught off-guard by HBM volume available to them. Committing to use something without any knowledge of volume would be a catastrophically bad move. Moving away from HMC in the first place was likely to do with it not being ready in time (and/or at necessary volumes/costs) after all.

As I say I imagine any sort of priority is simply a reserved fab capacity like AMD also have with GloFo, or Samsung have with their own fabs.

Quote:

---

Originally Posted by shaithis

But nVidia have been talking about Pascal with HBM for a while now.....and I doubt they aren't aware of lead-times......

---

And before that they were talking about Pascal with HMC, and also generically about "3D" memory. And they're currently talking about Pascal in 2016, so that's at least 6 months away. So it's reasonable to think that supply contracts may not have been finalised yet.

AMD getting priority doesn't mean nvidia aren't allowed through the door. It may mean nvidia have to delay launch of Pascal, or launch in limited quantities. But nvidia are hardly going to start bad-mouthing SK Hynix's product allocation plans if they're also reliant on SK Hynix for the memory for their next generation of cards, are they...? Just because they aren't whinging doesn't mean they're happy with the deal they're being offered.

Quote:

---

Originally Posted by watercooled

Moving away from HMC in the first place was likely to do with it not being ready in time (and/or at necessary volumes/costs) after all.

---

I see Intel is using HMC parts for Xeon Phi.

http://www.anandtech.com/show/9436/q...h-100-isc-2015

Edit to add: Although I suppose Larrabee is well known for delays, so shouldn't read much into that.

Yeah I didn't originally include the bracketed part then I remembered Phi. But being a HPC part, Phi may be able to cope with higher costs than a consumer GPU. Or it could be for technical reasons i.e. HBM being better-suited to GPUs.

Edit: Oh and with regard to volumes, it's possible Intel have a reserved supply considering HMC is Micron-Intel developed. So at the end of the day Nvidia had to compete with either Intel or AMD for capacity, both of whom co-developed the technologies.

Quote:

---

Originally Posted by shaithis

Could be AMDs final bad move if they try and leverage that to stifle competition. Do they have the funds to buy it all up and let it sit in a warehouse gathering dust?

---

Theoretically it would be the choice between your next generation card working like a next generation card, or not.

Fiji performance will be mid-range (we are dropping to 16nm from 28nm in one go), so HBM is a necessity even here.

So Nvidia need HBM to even have a chance of competing. If AMD is the only company with mid-range to high-end GPUs that are actually usable, then that is what people will have to buy.

However there is nothing in AMD's history that would suggest that they would try to do what is suggested here. And there is a chance that due to all the work they put into the technology, that they will earn from it selling, even if it to Nvidia.

Quote:

---

Originally Posted by yeeeeman

Yeah, bandwidth is nothing without a nice GPU. We saw that with the Fiji. No gain whatsoever. Maybe next time.

---

Well yes, it did help. The increased bandwidth meant that the 4GB VRAM could be more efficiently used, and is enough for 4K gaming, when in some cases 4GB of GDDR5 isn't enough.

Quote:

---

Originally Posted by DanceswithUnix

Talking things up is a skill of theirs.

---

It must be something in the air as it seems most companies are skilled at talking things up, or are the benchmarks in the AMD press release for the Fury X yesterdays news. ;)

Quote:

---

Originally Posted by CustardInc

Well yes, it did help. The increased bandwidth meant that the 4GB VRAM could be more efficiently used, and is enough for 4K gaming, when in some cases 4GB of GDDR5 isn't enough.

---

That difference was down to VRAM compression, not bandwidth AFAIK.

Quote:

---

Originally Posted by shaithis

Could be AMDs final bad move if they try and leverage that to stifle competition. Do they have the funds to buy it all up and let it sit in a warehouse gathering dust?

After some of the comments about nVidias business practices recently, not one of them come close to trying to prevent a competitor from making a new product like this seems to imply.

I guess this could get interesting and may force nVidia to throw it's weight around a bit more.

---

You are unbelievable. AMD buying up the memory technology they helped develop and according to you they are acting uncompetitively? Maybe Nvidia should spend more money on developing tech than on bribing the press and games devs?

Only fair really. Dumb people bash AMD because they have less money and praise nvidia for pushing "innovations' like gameworks that more than not have a negative impact on the industry. AMD actually HAS done massive things for us gamers, so why should this anti-competitive nothing doing company nvidia continue to get the praise of its followers by using tech developed by AMD while at the same time being unfair to AMD?

I would guess HBM2 production won't be as massive as gddr5. It would be enough to cover an entire range of GPUs (Top to bottom) and that seems to be AMDs intention. They will replace their entire line-up next year most likely. Serious trouble for nvidia if that comes to pass but their loyal deluded followers will still buy the less powerful more expensive hardware

Quote:

---

Originally Posted by shaithis

Could be AMDs final bad move if they try and leverage that to stifle competition. Do they have the funds to buy it all up and let it sit in a warehouse gathering dust?

After some of the comments about nVidias business practices recently, not one of them come close to trying to prevent a competitor from making a new product like this seems to imply.

I guess this could get interesting and may force nVidia to throw it's weight around a bit more.

---

Quote:

---

Originally Posted by shaithis

Could be AMDs final bad move if they try and leverage that to stifle competition. Do they have the funds to buy it all up and let it sit in a warehouse gathering dust?

After some of the comments about nVidias business practices recently, not one of them come close to trying to prevent a competitor from making a new product like this seems to imply.

I guess this could get interesting and may force nVidia to throw it's weight around a bit more.

---

Nvidia are more than entitled to partner with a memory company and develop their own advanced memory technology, at that point I'm sure they would get priority access to it too. Alternatively they could continue developing GDDR technology and find improvements in other areas.
No matter which way you look at this, there is nothing that's preventing Nvidia create a product, AMD are simply ensuring any demand that comes from companies that haven't had much/any involvement in the development of said technology doesn't impede their ability to get their product to market. This is a completely fair business approach IMO.

Given that AMD gets no access to Gameworks code at all which means the games using these 'plug-in APIs' cripple the game for AMD users untill they are disabled, I really don't understand how you came to your opinion that this is somehow worse.

Interesting times ahead.

AMD had this in the contract when it spent millions with Hynix helping to develop HBM, it was part of the deal and Nvidia will just have to wait until AMD's production demands slow down. That's what 7 years of working on HBM with Hynix gets you. HBM1 was a gap filler, a chance to showcase and make a few million from years of work, HBM2 will give AMD and its partners the chance to make huge money before surplus stock can be purchased by Nvidia.

Considering AMD co-developed HBM with Hynix since 2008, I'm really surprised they didn't make it proprietary and their own intellectual property and not release it for anyone else to use it except under huge licensing payments.

The fact that they are allowing other manufacturers, notably Nvidia, to use it is interesting and a nice stroke.

But not many of the bashers are looking at that. They just want a nail in the coffin for AMD for some weird variety of reasons so they can be happy to pay £600 for an equivalent performance of GTX 950Ti 5 years down the line.

I mean, call me a naive nelly but this is good stuff going down at the moment but the dregs and tosh have to be dredged up, come on and lets go back to playing with each other...in games...yeah..?

Quote:

---

Originally Posted by Corky34

It must be something in the air as it seems most companies are skilled at talking things up, or are the benchmarks in the AMD press release for the Fury X yesterdays news. ;)

---

lol, I think Fury X shows if anything that AMD are a bit rubbish at talking things up, or rubbish at rumour control at least.

Quote:

---

Originally Posted by Tabbykatze

Considering AMD co-developed HBM with Hynix since 2008, I'm really surprised they didn't make it proprietary and their own intellectual property and not release it for anyone else to use it except under huge licensing payments.

---

Could that even be done? Hynix probably holds the copyright and can prevent other companies making HBM, but can/could that be extended to preventing it being used in non-AMD products, I know the American copyright/patent laws are pretty messed up but that seems a stretch even for them. :)

Quote:

---

Originally Posted by watercooled

But being a HPC part, Phi may be able to cope with higher costs than a consumer GPU.

---

It is an expensive chip (I don't expect to see a Raspberry Phi any time soon lol), but I'm sure I read that they were basically giving them away to win prestigious builds in the last generation. I think it is more about denying Nvidia high end HPC sales than making any profit themselves.

Quote:

---

Originally Posted by Corky34

Could that even be done? Hynix probably holds the copyright and can prevent other companies making HBM, but can/could that be extended to preventing it being used in non-AMD products, I know the American copyright/patent laws are pretty messed up but that seems a stretch even for them. :)

---

I expect it could be done, but standards drive volume and volume is what wins in pricing in the long run so restricting use would probably hand the eventual crown to HMC.

Quote:

---

Originally Posted by Corky34

Could that even be done? Hynix probably holds the copyright and can prevent other companies making HBM, but can/could that be extended to preventing it being used in non-AMD products, I know the American copyright/patent laws are pretty messed up but that seems a stretch even for them. :)

---

Just to re-iterate, SK Hynix and AMD developed this and both likely hold the rights to it. AMD didn't just walk up and ask to buy some pre-made from SK Hynix. Charlie at Semiaccurate had pictures of prototype silicon back in 2011: http://semiaccurate.com/2011/10/27/a...-gpu-pictured/

And it would have been in development years before that of course. I'm not aware of any laws that force you to sell a product to your competitor, especially when there are alternatives available. I don't see anyone else using Intel's eDRAM.

But I agree with DanceswithUnix, it's not in anyone's interest to artificially strangle supply or lock out competitors if you want it to gain widespread adoption and likely more licensing fees in the long term.

Yea I know they both worked on the development, AMD as far back as 2008 (iirc), It was more a question on how such a joint venture pans out when it comes to patent restrictions if more than one company owns the IP, I'm guessing by the sounds of what yourself and DanceswithUnix said it could be something that's negotiated between the two and every tiny little detail is written up in a contact, although I'm still unsure if more than one company can own a patent. :undecided

It's not just patents that matter though, look at the x86/AMD64 cross-licensing between Intel and AMD64 - both are licensed to produce x86 and AMD64 compatible processors and besides VIA (I'm not exactly sure of the terms with VIA but they've been in the game a long time and don't directly compete with Intel), no-one else. Nvidia tried to get a license from Intel a few years ago, likely for Denver, but wasn't able to.

Unless patents are considered critical and hence given FRAND status, companies aren't necessarily obliged to license technology.

It's a distinct possibility, Hynix could hold the patent of which AMD has sole usage of through co-development and shares the IP (not the Patent). Allowing them to be the only company on the market with HBM until Nvidia go "well wait a frackin minute" and then it has to become FRAND. If it even extends that far, track record of patents being forced to be released on FRAND to the public (people who are better qualified and informed, please correct me if i'm wrong) are because they're a global enhancement in the technological sector in orientation to that area. Is HBM a global enhancement? Could be, is GDDR5 still sufficient? Maybe.

But AMD and Hynix have made a deal for priority so they've gone "hey Nvidia, we've advanced and evolved down the GPU evolutionary line and we will let you reap the benefits as well, but you'll have to wait."

In a business sector and R&D competition, that's like me sharing my bank account with my next door neighbour who's only on 20k and i'm on 120k and he gets to play around with what's left in it after i've bought my Pettabyte of pornography. This is a bold move cotton, lets see if it works out for them!

Quote:

---

Originally Posted by shaithis

Could be AMDs final bad move if they try and leverage that to stifle competition. Do they have the funds to buy it all up and let it sit in a warehouse gathering dust?

After some of the comments about nVidias business practices recently, not one of them come close to trying to prevent a competitor from making a new product like this seems to imply.

I guess this could get interesting and may force nVidia to throw it's weight around a bit more.

---

Frankly, I'd enjoy seeing AMD stick it to nVidia for a change by reserving all the HBM2 memory. They worked with Hynix, shared some of the development costs, so why shouldn't they make nVidia wait in line. It would serve them and all their fanboys right for always justifying nVidia's cheating tactics as 'just business.' Well, starving nVidia of HMB2 memory would also then be 'just business.'

Did not read anywhere that AMD were going to buy up everything and stash it away.

Sounds more like two companies willing to get behind a new tech, (as nothing has changed in memory for years, up clocking the same types is nothing big) and allowing AMD first pic of ..."initial production capacity as possible,"

Quote:

---

Originally Posted by shaithis

Could be AMDs final bad move if they try and leverage that to stifle competition. Do they have the funds to buy it all up and let it sit in a warehouse gathering dust?

After some of the comments about nVidias business practices recently, not one of them come close to trying to prevent a competitor from making a new product like this seems to imply.

I guess this could get interesting and may force nVidia to throw it's weight around a bit more.

---

AMD developed HBM together with Hynix. They have a much better track record than NVidia in both business practices and open technologies but even AMD can't be expected to immediately deliver their advantages to NVidia because somebody on the internet isn't happy.

I am not sure we will get it next year though

Can't understand what all the grumbles are about when it comes to AMD getting first priority. They spent the time and money to help develop it, so it is only fair that they should have first pick, I am just happy to see forward progress.
My problem is that when tech and software is developed to improve things for the consumer, some companies take their bat and ball and go home (they try to totally restrict others from ever doing similar things or try to sue everyone like a fruity product we all know - And they still use a ******* proprietary charging cable??!!??). It always comes as a disadvantage to the consumer, who looses out with cost and reduced usability.
If there was only one video card producer, there would be no competition and not a lot of impetus for major innovations. Just look at how slow car engine development went while fossil fuel was the only major player, until the recent push to find alternatives. Example- I once owned (until I was T-boned by a truck) a 4 cylinder, normally aspirated, car which was developed in the late 1950's and first produced in the early 1960's, capable of over 40MPG when cruising, had a eight speed gear box (plus a 2 speed reverse) and did over 150mph when pushed.
Besides, why badmouth one company for (possibly) doing what it's competition already does.
In the end, I hope that no matter what the outcome or brand you prefer, that we consumers are the final winners.
Personally, I just look for whoever has the best value for the money for what I require - no real preference for either brand.

I wish all the best to AMD, I hope they buy the entire stock of HBM2 and let NShidia starve, they deserve, at least they are playing dirty more than a decade. I am AMD fan and would like to see amd shine.