Samsung mass producing HBM memory in 2016

[Abwx]

Tential. He just claimed Semi-Accurate as a resource for his information.

He doesn't need education
He claims citations aren't needed
And when he is cornered and is pressed for a source, he says SA.

El fini

Yeah, you are right, HBM2 is not even validated and AMD/Hynix are working hard so Nvidia will be first to access to HBM2....

I guess that this wet dream started with this rumour :

http://www.kitguru.net/components/g...ly-taped-out-on-track-for-2016-launch-rumour/

And of course some fans promptly jumped on this badwaggon thinking that Nvidia would be first to access the RAM that AMD/Hynix are currently working on, that s telling about the desperation caused by AMD innovation and lead in this strategical area.

The only thing they could access if they release their next chip in H1 2016 is HBM1, and they know that it wouldnt bring them any advantage, likely that they ll wait that AMD/Hynix has validated HBM2, they are lagging and it s a fact, or do you need a source to support this.??.

Btw, where are your own sources..???.
Yeah, true that you said nothing relevant at this point, so it s right that you dont need any source..

 

[tential]

Oh you mean like how HBM 1 did for Fury X? :hmm:

If people want to use HBM2 as a main selling point, it's going to be VERY hard given Fury X. And I don't think Fury X is a bad launch or a bad product. I believe if the tables were turned (Nvidia + Freesync + Fury X vs AMD + 980Ti + Gsync), we'd have more Fury X owners here than 9809Ti owners. 80/20.
I'm not kidding either in this belief.

Nvidia knows how to sell something. They make things sound good to consumers.
AMD can take a perfectly good product, and kill it. Give AMD the 980Ti, and lets even ignore the fact that they'd screw up something like the cooler, or not have enough review samples, or have drivers that take years to reach maturity and show the true potential of the card, etc.
They'll push power consumption, FPS or some other metric. They'll draw straight comparisons to Fury X the whole time.

Give Nvidia Fury X, they'll know they're at an FPS disadvantage. They'll know they're at a power draw disadvantage. Nvidia doesn't CARE about what it's bad about. Nvidia always shapes the game to what they're winning at, which is what ANY intelligent company does.
What would Nvidia do in this position? They'd push Smooth gaming all day in our face. They would tell us how cheap Freesync monitors will be here at the end of the year. So what if you gain 5 FPS if you lose your soul with those evil AMD guys who pay devs off to gimp our gpus. At the end of the day, can you tell the difference between smooth gaming and smooth gaming? No... but your wallet can. Go with Fury X.

The reason I LOVE Nvidia as a COMPANY. I should have done a case study on them in business school haha.

Bumpgate? No problem for Nvidia.
GTX 970? "Show them the benches, they'll forget that we gave them segmented memory and love us for the fact that a GTX 970 came out SO CHEAP. SO CHEAP. Remember what we did for you... so forgive us this minor mistake that didn't effect your benches anyway and stop complaining! We know you love this card.
http://store.steampowered.com/hwsurvey/directx/
5.98% of cards.. It's the MOST POPULAR CARD, and I didn't even know that before I started typing this." - NV internal convo between employees [taken with hidden camera footage blocked for concealment of identities /sarcasm]

Nvidia plays off things with no issues.
Their launches? Done so well. How they time their product releases to get the MOST out of their customers? They can get a person to get a GTX 970, then a Titan X, then go "You know what? I'll just get GTX 980 Ti SLI, that's faster and only a little bit more extra". They give themselves THREE chances to hit a high end consumer. In one generation. That's how you launch a generation there boys. That's how you make money.

That's why Nvidia isn't going anywhere right now. Nvidia's shares, ALWAYS go up(The only downfalls coincide with instances where the whole market collapsed.....). I wonder why...

 

[tential]

I agree. Nothing to be won here.

A couple of years ago ATI pushed ahead with GDDR4. They didn't realize any significant gains with that either iirc. A year later though the fairly similar GDDR5 was released and AMDs transition to that standard was much smoother than Nvidias. Memory controllers certainly aren't easy to develop and a head start on that can be worth a lot. It took Nvidia until Kepler to catch up and that was one of the strongest market positions AMD has had.

Then you're talking about a different point entirely and one I will readily concede as it's one I don't know about.

My issue is comments like this in bold:

Still the possibility there was a contract between AMD and Hynix that AMD gets "premium access" because they helped them develop it. That contract probably was done half a decade ago.

Also why is there money to make? With the proclaimed death of dGPU and dGPU having low volume anyway compared to say mobile chips, I don't really see there being that much money in HBM. As far as we know it's expensive so already a no-go for mobile SOCs besides them not needing it anyway.

NV and AMD will be the only ones to use it in the foreseeable future. Intel/micron have their own thing so also not possible to get money from Intel.

And This:

There's a lot of silly speculation here, but lets put our common sense hat on. AMD funded the development of HBM. AMD is still a billion dollar corporation that is probably smarter as a whole with their fleet of lawyers smarter than any poster here. They did the earlier "Heavy lifting" for Hynix. It would be incredibly incompetent of them if there is nothing in the contract of Hynix that give them a competitive advantage.

 

[boozzer]

If people want to use HBM2 as a main selling point, it's going to be VERY hard given Fury X. And I don't think Fury X is a bad launch or a bad product. I believe if the tables were turned (Nvidia + Freesync + Fury X vs AMD + 980Ti + Gsync), we'd have more Fury X owners here than 9809Ti owners. 80/20.
I'm not kidding either in this belief.

Nvidia knows how to sell something. They make things sound good to consumers.
AMD can take a perfectly good product, and kill it. Give AMD the 980Ti, and lets even ignore the fact that they'd screw up something like the cooler, or not have enough review samples, or have drivers that take years to reach maturity and show the true potential of the card, etc.
They'll push power consumption, FPS or some other metric. They'll draw straight comparisons to Fury X the whole time.

Give Nvidia Fury X, they'll know they're at an FPS disadvantage. They'll know they're at a power draw disadvantage. Nvidia doesn't CARE about what it's bad about. Nvidia always shapes the game to what they're winning at, which is what ANY intelligent company does.
What would Nvidia do in this position? They'd push Smooth gaming all day in our face. They would tell us how cheap Freesync monitors will be here at the end of the year. So what if you gain 5 FPS if you lose your soul with those evil AMD guys who pay devs off to gimp our gpus. At the end of the day, can you tell the difference between smooth gaming and smooth gaming? No... but your wallet can. Go with Fury X.

The reason I LOVE Nvidia as a COMPANY. I should have done a case study on them in business school haha.

Bumpgate? No problem for Nvidia.
GTX 970? "Show them the benches, they'll forget that we gave them segmented memory and love us for the fact that a GTX 970 came out SO CHEAP. SO CHEAP. Remember what we did for you... so forgive us this minor mistake that didn't effect your benches anyway and stop complaining! We know you love this card.
http://store.steampowered.com/hwsurvey/directx/
5.98% of cards.. It's the MOST POPULAR CARD, and I didn't even know that before I started typing this." - NV internal convo between employees [taken with hidden camera footage blocked for concealment of identities /sarcasm]

Nvidia plays off things with no issues.
Their launches? Done so well. How they time their product releases to get the MOST out of their customers? They can get a person to get a GTX 970, then a Titan X, then go "You know what? I'll just get GTX 980 Ti SLI, that's faster and only a little bit more extra". They give themselves THREE chances to hit a high end consumer. In one generation. That's how you launch a generation there boys. That's how you make money.

That's why Nvidia isn't going anywhere right now. Nvidia's shares, ALWAYS go up(The only downfalls coincide with instances where the whole market collapsed.....). I wonder why...

that is marketing bro the power of marketing.

 

[Keysplayr]

Yeah, you are right, HBM2 is not even validated and AMD/Hynix are working hard so Nvidia will be first to access to HBM2....

I guess that this wet dream started with this rumour :

http://www.kitguru.net/components/g...ly-taped-out-on-track-for-2016-launch-rumour/

And of course some fans promptly jumped on this badwaggon thinking that Nvidia would be first to access the RAM that AMD/Hynix are currently working on, that s telling about the desperation caused by AMD innovation and lead in this strategical area.

The only thing they could access if they release their next chip in H1 2016 is HBM1, and they know that it wouldnt bring them any advantage, likely that they ll wait that AMD/Hynix has validated HBM2, they are lagging and it s a fact, or do you need a source to support this.??.

Btw, where are your own sources..???.
Yeah, true that you said nothing relevant at this point, so it s right that you dont need any source..

Sources for...... what? My claim is that you didn't know anything. So, I guess you are my source for my claim incarnate.

What I CAN say without any fear of being wrong, is that both AMD and Nvidia will have ample supply of HBM/HBM2 to meet their needs when the time comes.
That's all we can really say with confidence right now.

 

[tential]

that is marketing bro the power of marketing.

AMD is an easy sell.
It really actually is when you think about it.

It's AMD that can't sell what it has.

I think the GTX 980Ti is the best card by far. I just realized that the Fury X gives me what I want. A smooth gaming experience at 4K at a reasonable price while being quiet. Gsync? Nothing above 30 inches out and like 34 is the biggest maybe 35? Nothing above 40 inches for a monitor. Freesync? Up to 65 inches coming out. 55 available, a couple 40, you have options at ALL price points. Talk about convenience for any person who wants to buy ANY amd GPU and get a good gaming experience.
Then we talk Sapphire Fury? It's almost Fury X performance, like 97% for over $100 less, it's actually in stock, and it's QUIETER than Fury X?
Wow.... and I still get a great gaming experience that's cheaper than the GTX 980 Ti....

It's not a hard sell....

AMD just has to get people to play their game. If you give people the Fury X, and you don't give them a FreeSynx Monitor as well to do a GPU review as AMD. Then you should quit business forever. You should be allowed to leach off the tax payer, because you're clearly incompetent and making people see your product in the best light and attempt to downplay all potential pitfalls of your product like any intelligent human being would when selling something.

 

[Azix]

the 980ti probably would not meet AMDs standards. A cut down chip that is still larger than their own. They don't even have a hardware scheduler in there and it still is larger, similar in power consumption, less compute. Its a rubbish chip by AMD standards, but it does push the pixels adequately (for now). AMDs problem is not having a fab (at least not going below 28nm soon enough). I am not sure it would matter if they switched. AMD will sell as many fury x chips as they make I think. Unlikely they intend to make a bazillion since the high end is not the main income.

HBM2 will be great for AMD and Nvidia. its a big deal. For AMD at least we see decent gains on fury x with a little HBM oc. getting to 1TB/s will solve most bandwidth problems and the space boost doesn't hurt (8GB per stack)

I dont expect HBM to be the huge advantage in itself. Even if AMD gets some benefit nvidia will have hbm one way or the other. The difference will be in engineering and AMDs superior engineering might put them ahead in performance. It's quite possible nvidia has been caught with their pants down regarding dx12 or they could screw up implementing HBM. It's down to what each party brings next year in their GPUs. Even if AMD releases first, people will wait.

I don't expect the same situation that we had with the 290x and its reference cooler next year. Though that will depend on how the nano performs on heat and noise. Next year reference GPUs are very likely to take that form with non-reference looking like the Fury

 

[tential]

the 980ti probably would not meet AMDs standards. A cut down chip that is still larger than their own. They don't even have a hardware scheduler in there and it still is larger, similar in power consumption, less compute. Its a rubbish chip by AMD standards, but it does push the pixels adequately (for now). AMDs problem is not having a fab (at least not going below 28nm soon enough). I am not sure it would matter if they switched. AMD will sell as many fury x chips as they make I think. Unlikely they intend to make a bazillion since the high end is not the main income.

HBM2 will be great for AMD and Nvidia. its a big deal. For AMD at least we see decent gains on fury x with a little HBM oc. getting to 1TB/s will solve most bandwidth problems and the space boost doesn't hurt (8GB per stack)

I dont expect HBM to be the huge advantage in itself. Even if AMD gets some benefit nvidia will have hbm one way or the other. The difference will be in engineering and AMDs superior engineering might put them ahead in performance. It's quite possible nvidia has been caught with their pants down regarding dx12 or they could screw up implementing HBM. It's down to what each party brings next year in their GPUs. Even if AMD releases first, people will wait.

I don't expect the same situation that we had with the 290x and its reference cooler next year. Though that will depend on how the nano performs on heat and noise. Next year reference GPUs are very likely to take that form with non-reference looking like the Fury

Worst ROI ever.
Denied. Rewrite the plan again.

That's the correct analysis of the situation. If AMD looked down the road, and saw this situation during the strategic planning phase, this is the CLEAR outcome.

Too much risk, too little reward. Horrendous ROI.

No go plan.

Do you see why AMD should never have engaged in this?

Three options.

A) Ensure HBM is exclusively yours by patent (I'm sure someone can explain the legality of it blah blah)

B) Ensure you have all production to yourself by buying all available stock
(Extremely not viable for AMD given the other players).

C) Never engage in the creation of HBM and let another company do it first instead, then copy off them.
The risks in this option are as follows:
-if any other company creates HBM you will have to license it. So will Nvidia though.
-if any other company creates HBM as a free standard, you'll have to learn to implement it. So will nvidia though.
-if Nvidia creates HBM as an open standard, they'll have an advantage in learning to use it, but you'll still get to use it without additional cost.
-if nvidia creates hbm and AMD has to license it, and has a disadvantage coming to the party.

AMD decided to instead of taking a viable strategy, took an extremely risky one for a very small reward. At WORST, Nvidia screws up 1 generation of HBM implementation. If that's what AMD is banking on with this investment in HBM, then this is a lost cause, because Nvidia can survive 1 bad generation easily...

 

[Azix]

Hynix will be selling HBM. They aren't going to make it for free. The investment will be recovered one way or the other. AMD could get a portion of those sales as well as licensing.

As for AMD inventing HBM, they had to. Just like all the older VRAM tech they've upgraded from, GDDR5 was not going to last forever and nvidia sure as heck was not going to solve the situation.

 

[tential]

Hynix will be selling HBM. They aren't going to make it for free. The investment will be recovered one way or the other. AMD could get a portion of those sales as well as licensing.

As for AMD inventing HBM, they had to. Just like all the older VRAM tech they've upgraded from, GDDR5 was not going to last forever and nvidia sure as heck was not going to solve the situation.

AMD is a company run for profit. They should not care if RAM technology progresses. They should care if there is a GREAT ROI in inventing HBM. If not, you don't do it.

End of story.

 

[Abwx]

The reason I LOVE Nvidia as a COMPANY. I should have done a case study on them in business school haha.

Bumpgate? No problem for Nvidia.
.

Lol, so the guy that punch you, throw you to the ground and stomp you before taking you wallet, this is the guy that deserve your respect according to your post...

No need of business school to learn this kind of marketing, a gettho is a better place....:biggrin:

 

[Azix]

AMD is a company run for profit. They should not care if RAM technology progresses. They should care if there is a GREAT ROI in inventing HBM. If not, you don't do it.

End of story.

we don't know what their arrangement is on the technology.

But as far as their GPUs (and most of their segments actually) there was the need.

 

[AnandThenMan]

The reason I LOVE Nvidia as a COMPANY. I should have done a case study on them in business school haha.

You are essentially saying you love Nvidia as a company because they know how to pull the wool over their loyal customers eyes. Why this should garner respect and admiration escapes me.

Myself I respect engineering prowess and innovation, and especially when a company comes up with radical and better ways of doing things. Is see HBM as one of those things, it is simple on the surface but when you really look at what is being done it is pure genius and is one of those things where you wonder to yourself, why didn't I think of that?

In a nutshell engineering>marketing but the mass market sheep care mostly about perception and feeling warm and fuzzy about their purchases.

 

[Abwx]

Sources for...... what? My claim is that you didn't know anything. So, I guess you are my source for my claim incarnate.

What I CAN say without any fear of being wrong, is that both AMD and Nvidia will have ample supply of HBM/HBM2 to meet their needs when the time comes.
That's all we can really say with confidence right now.

Re lol, where is your source..?..

As for being confident and not wrong i wouldnt be that sure if i were you..

You think that Samsung need one year to be capable to manufacture HBM1..??.

Did you notice that they ll start production exactly 12 months after Hynix..?.

That s a smoking gun that say that Hynix will be the only one producing HBM2 for one year before licencees have the right to start their own production, so Nvidia is in deep trouble as they will have to live with HBM1 untill late 2017 while AMD will be using HBM2 exclusively for a full year, exactly like they are doing for HBM1.

And dont think that Hynix will lose some money, actualy this will allow them to have a substancialy lower cost than their licencees for a year or two, for both Hynix and AMD this is a win win situation.

Now i guess that you ll say that i know nothing as usual "argument" of yours...

 

[tential]

You are essentially saying you love Nvidia as a company because they know how to pull the wool over their loyal customers eyes. Why this should garner respect and admiration escapes me.

Myself I respect engineering prowess and innovation, and especially when a company comes up with radical and better ways of doing things. Is see HBM as one of those things, it is simple on the surface but when you really look at what is being done it is pure genius and is one of those things where you wonder to yourself, why didn't I think of that?

In a nutshell engineering>marketing but the mass market sheep care mostly about perception and feeling warm and fuzzy about their purchases.

Good for you on a tech forum in which we can have a conversation between 25 people easily.

The rest of the market isn't concerned with engineering feats.

http://www.neogaf.com/forum/showthread.php?t=1042015
Fury X release

http://www.neogaf.com/forum/showthread.php?p=176742481
R9 Nano release (Price the whole way in that thread... the only R9 nano release thread on a massive gaming forum, before dying out vs Fury X)

Notice the enthusiasm with both....
People aren't interested anymore.

Engineering feats don't sell anything. You can like it, but it won't keep AMD around.

I'm surprised you believe engineering feats will ever impress the general market to sell graphics chips... did you seriously think that was a viable way to attempt to sell graphics chips to gamers?

I'm also surprised you didn't buy a Fury just because of HBM alone since you're in admiration of it.
Fury X was my Day 1 purchase until the Overclocker's dream turned out to be bold lie like no other....

 

[AnandThenMan]

I'm surprised you believe engineering feats will ever impress the general market to sell graphics chips... did you seriously think that was a viable way to attempt to sell graphics chips to gamers?

I said nothing of the sort you just make that all up in your head.

 

[Keysplayr]

Re lol, where is your source..?..

As for being confident and not wrong i wouldnt be that sure if i were you..

You think that Samsung need one year to be capable to manufacture HBM1..??.

Did you notice that they ll start production exactly 12 months after Hynix..?.

That s a smoking gun that say that Hynix will be the only one producing HBM2 for one year before licencees have the right to start their own production, so Nvidia is in deep trouble as they will have to live with HBM1 untill late 2017 while AMD will be using HBM2 exclusively for a full year, exactly like they are doing for HBM1.

And dont think that Hynix will lose some money, actualy this will allow them to have a substancialy lower cost than their licencees for a year or two, for both Hynix and AMD this is a win win situation.

Now i guess that you ll say that i know nothing as usual "argument" of yours...

Common sense. Eventually, both AMD and Nvidia will be using HBM and HBM2...3.....4.....11.

I don't recall saying anything like Samsung needs one year to be capable to manufacture HBM1.

12 Months after Hynix? Well, when did Hynix start? Day before Fury X launch?

The smoking gun is all in your head and in your dreams my friend.

Yes....... sigh.... Hynix will lose money if they refuse to sell to Nvidia.

Don't be silly. I'm sure you know..... something..

 

[Ma_Deuce]

AMD is a company run for profit. They should not care if RAM technology progresses. They should care if there is a GREAT ROI in inventing HBM. If not, you don't do it.

End of story.

You seem pretty confident of the horrible ROI of developing developing HBM. Do you have anything to back this up or just gut feeling? Apologies if there was somethibg buried in a post that I missed.

Edit: your statement is pretty simple and 100% correct, but I would be willing to bet that on any kind of deal like this with AMD you have plenty if people scrutinizing the return on the investment here.

 

[Ma_Deuce]

Hynix will lose money if they refuse to sell to Nvidia.

As long as Hynix can sell everything they can produce right now/future orders I don't think that nvidia sales are going to matter too much. You make nvidia sound like the biggest fish after HBM, but depending on who else jumped in the pond they can get pushed aside easily. If samsung is throwing some real muscle into this, there had to be more interest thatln just AMD and Nvidia discrete GPU.

 

[Keysplayr]

You heard it here folks. Nvidia sales don't matter to Hynix. They are only in the business of selling memory to companies OTHER than Nvidia. They don't really want money. They want the admiration of all the gamers out there supporting a dying underdog company.

All kidding aside, what planet are you people from anyway? LOL.

 

[ShintaiDK]

Its always funny when company loyalty changes the laws of capitalism.

 

[Abwx]

I don't recall saying anything like Samsung needs one year to be capable to manufacture HBM1.c

That s not what i said and wasnt the point, wich is that with the process methodology in hand they dont need that much time to be production ready.

12 Months after Hynix? Well, when did Hynix start? Day before Fury X launch? [/QUOTE]

A few months before its was launched, about Q1 2015..

The smoking gun is all in your head and in your dreams my friend.

Yes....... sigh.... Hynix will lose money if they refuse to sell to Nvidia.

Don't be silly. I'm sure you know..... something..

They wont lose money at all, and they can sell the gear to Nvidia but not before Q1 2016, and not before Q4 2017 for HBM2.

AMD will have a full year advantage and it s fair, they innovated and they deserve to get the fruits of their efforts, they will regain substancial marketshare and they aleady are, otherwise we wouldnt have Nvidia publishing slides with partial infos about marketshares, that s also a smoking gun as they are not acting randomly...

 

[ShintaiDK]

AMD will have a full year advantage and it s fair, they innovated and they deserve to get the fruits of their efforts, they will regain substancial marketshare and they aleady are, otherwise we wouldnt have Nvidia publishing slides with partial infos about marketshares, that s also a smoking gun as they are not acting randomly...

No, there is absolutely nothing that supports this. Only the opposite. Even AMDs financials are against you.

Please show your documentation.

 

[Keysplayr]

12 Months after Hynix? Well, when did Hynix start? Day before Fury X launch?

A few months before its was launched, about Q1 2015..




They wont lose money at all, and they can sell the gear to Nvidia but not before Q1 2016, and not before Q4 2017 for HBM2.

AMD will have a full year advantage and it s fair, they innovated and they deserve to get the fruits of their efforts, they will regain substancial marketshare and they aleady are, otherwise we wouldnt have Nvidia publishing slides with partial infos about marketshares, that s also a smoking gun as they are not acting randomly...

Ah you mean HBM1. Irrelevant. Surprised you didn't know that.

Show us the contract brudda.

Everything seems to be a smoking gun with you. LOL.

Know what? Nevermind. You are utterly brainwashed.

/out.

 

[Abwx]

Ah you mean HBM1. Irrelevant. Surprised you didn't know that.

Show us the contract brudda.

Everything seems to be a smoking gun with you. LOL.

Know what? Nevermind. You are utterly brainwashed.

/out.

Ad hominem is an aknowledgment that you have no argument..

As for being utterly brainwashed, lol, dunno but certainly that i m not gearwashed by some items that would be offered to me by AMD to promote them and bash their competitor.

No, there is absolutely nothing that supports this. Only the opposite. Even AMDs financials are against you.

Please show your documentation.

And where is your own documentation, that s no all to negate what others are saying if your have even less arguments at hand.

You think that since they updated their GPU line they did lose more marketshare, or that they did gain nothing..?.

That Fury and the refreshed Hawais didnt entail Nvidia s sales in thoses segments.?.

To prove this you ll have to prove that not a single Fury was sold, or that this segment is currently enjoying a growth that benefit mainly to Nvidia...