The Rise and Fall of AMD.

[NTMBK]

AMD to Show the First Temash-Based Product Prototypes in CES Timeframe.

Interesting how they call 4 cores and cache one Jaguar Compute Unit. Seems like the graphics guys are rubbing off on them...

 

[Olikan]

Why not? With memory they rebranded and repackaged a patriot memory and it was considered a dumb move. Now they are rebranding and repackaging someone else' SSD (Patriot again?) and suddenly it isn't a dumb move?

ddr market is dead, some companyes already went bankrupt, the still alive ones aren't doing great and are selling other stuffs to survive
it's a dead end in the road
SSD market is new and increasing...the road is still beeing built

to make thing clear,
re-seling stuff is never a clever idea...but atleast amd is doing it in a fresh market, not a in a undead one

 

[NTMBK]

ddr market is dead, some companyes already went bankrupt, the still alive ones aren't doing great and are selling other stuffs to survive
it's a dead end in the road
SSD market is new and increasing...the road is still beeing built

to make thing clear,
re-seling stuff is never a clever idea...but atleast amd is doing it in a fresh market, not a in a undead one

I suspect the memory move was to support their APU. If I were AMD I'd be sick of seeing OEMs selling APUs with 1066 DDR3. Slapping the Radeon brand onto high end memory indicates that you will want this level of performance to get good graphics performance out of your APU.

 

[AtenRa]

Yes, we know. That didn't make sense either. This move is a death throw, AtenRa. Probably one of many because this is going to be a slow process to watch. And a sad one. Please stop ignoring logic here and understand why this is an unusually dumb move, even for AMD.

People condemned the RADEON SSD already without even have any inside of the actual deal between the two parties(AMD and the manufacturer) or any deals that AMD have made with OEMs or how much the actual product will cost in retail etc etc etc.

I have only linked the actual news and haven't made any comments because i dont now shit of any mentioned above. Ill wait to see the product first and then judge it.

 

[mrmt]

I suspect the memory move was to support their APU. If I were AMD I'd be sick of seeing OEMs selling APUs with 1066 DDR3. Slapping the Radeon brand onto high end memory indicates that you will want this level of performance to get good graphics performance out of your APU.

I see something along the same lines too, but it is still a dumb move. AMD doesn't have any kind of leverage on the memory market to make a difference or in the OEM market to coerce them to do anything. They have just 16% of the entire market, and of those Trinity/Llano APU are just 6-7%, and shrinking.

I understand that memory really impairs Trinity GPU, but that's what AMD gets for coupling a good GPU with Bulldozer. As a result AMD APUs are too low on the price ladder to OEMs to couple them with expensive parts, including high speed memory.

 

[Makaveli]

People condemned the RADEON SSD already without even have any inside of the actual deal between the two parties(AMD and the manufacturer) or any deals that AMD have made with OEMs or how much the actual product will cost in retail etc etc etc.

I have only linked the actual news and haven't made any comments because i dont now shit of any mentioned above. Ill wait to see the product first and then judge it.

I think most people are making these assumptions because.

1. AMD doesn't have their own controller and I doubt the they have the R&D budget to design there own.

2. Therefore they will be licensing.

3. They don't produce their own flash memory.

4. They are jumping into a market already full of established competitors.

5. They are going to dilute the Radeon brand by doing this.

6. People want to see innovation from AMD not re branding.

 

[AtenRa]

@Makaveli

1 and 2. Intel 330,335,510,520 and next year 5xx series dont have an Intel controller.

3. Only a handful of SSD makers produce NAND memory. OCZ, Patriot, Corsair and more DON'T produce NAND but they have nice SSDs and they make a profit out of them.

4. Yes but SSD market is doubling almost every two years.

5. We havent seen the product yet.

6. yes but they(AMD) need to make money too.

 

[georgec84]

Side note - with Intel potentially removing upgrade capabilities after Haswell, I wonder if this will swing a little financial support AMD's way?

 

[MaxPayne63]

Side note - with Intel potentially removing upgrade capabilities after Haswell, I wonder if this will swing a little financial support AMD's way?

No. The vast majority of users are never going to change an internal component in their PC. If anything I would say that it will erode AMD's market share further, as greater integration will over time lead to reduced costs and/or more features.

 

[Makaveli]

@Makaveli

1 and 2. Intel 330,335,510,520 and next year 5xx series dont have an Intel controller.

3. Only a handful of SSD makers produce NAND memory. OCZ, Patriot, Corsair and more DON'T produce NAND but they have nice SSDs and they make a profit out of them.

4. Yes but SSD market is doubling almost every two years.

5. We havent seen the product yet.

6. yes but they(AMD) need to make money too.

It doesn't really matter for Intel they aren't in AMD's position financially. So even an Intel SSD with a sandforce controller people will still buy. But would those same people buy an AMD SSD with a sandforce controller and sandforce firmware?? I know this would cause me to think twice.

can't speak for OCZ but Patriot, Corsair i'm pretty sure are not in the red as much as AMD. And they also have been in the rebranding business for much longer

The SSD market doubling in size is good but we don't really know what the profit margins are like.

And yes AMD needs to make money but you aren't gonna make much when you are late to the party and have to license other people's technology!

 

[mrmt]

can't speak for OCZ but Patriot, Corsair i'm pretty sure are not in the red as much as AMD. And they also have been in the rebranding business for much longer

IIRC OCZ has its own controller and develops their own firmware so I would not say that they are doing just re-branding, Corsair develops its own firmware too, I'm not really sure about Patriot.

In any case, at the very least those guys are testing and validating software and performing QA, not a trivial task on this market. It's a very different thing of what AMD is doing with RAM, and if the rumor is correct, very different from what AMD aims to do with their SSD.

 

[NostaSeronx]

Next thing you guys will notice is Next-gen Graphic Core Next has x87/ZMM registers and can compute x86.

 

[NTMBK]

Next thing you guys will notice is Next-gen Graphic Core Next has x87/ZMM registers and can compute x86.

ZMM? As in Xeon Phi registers?

 

[Makaveli]

IIRC OCZ has its own controller and develops their own firmware so I would not say that they are doing just re-branding, Corsair develops its own firmware too, I'm not really sure about Patriot.

In any case, at the very least those guys are testing and validating software and performing QA, not a trivial task on this market. It's a very different thing of what AMD is doing with RAM, and if the rumor is correct, very different from what AMD aims to do with their SSD.

We will have to wait and see.

And OCZ is in just as bad of shape as AMD right now so they are kind of a bad example

 

[NostaSeronx]

ZMM? As in Xeon Phi registers?

Correct.

Where Xeon Phi isn't really compatible with AMD64, I assume NGGCN will be. Each CU for GCN 1.0 has 1 Scalar ALU + 4 * 16-lane Vector ALUs -> 1 x87 unit + 4 512-bit vector units. Then, you have Steamroller and Graphic Core Next having the same cache design and the same memory design. It is only natural that AMD will improve on x86 integration with GCN.

"At the system level, GCN is the only discrete graphics architecture that is compatible with the x86 programming model, paving the way for future software
and hardware integration." - www.amd.com/jp/Documents/GCN_Architecture_whitepaper.pdf

 

[Haserath]

I'm wondering if the SSD move is to get more platforms with solid state performance.

Even delivering a crap SSD, it will still demolish a hard drive in 'platformance'.

I don't know how else they could possibly think this is a good idea. Unless they actually have secret access to some controller.

 

[Ferzerp]

The AMD way is to benchmark them vs $1,500 15k RPM FC drives and then claim massive victory

It makes as much sense as their GPU limited benchmarks comparing their junk to the top of the line Intel cpus.

 

[Idontcare]

AMD to launch Radeon Branded SSD's (S/A)

AMD Announces Global Distribution Agreement with Symmetry Electronics

I think this is a good thing for multiple reasons.

For one, it shows that AMD is not sitting around letting grass grow under its feet while its core revenue generator slowing erodes (x86 desktop units decrease). Yeah its not the perfect solution where they design everything in house top-to-bottom but these things are built in a day. This is just step one, if they gain traction then they'll invest more money into the effort to begin to develop products and so on.

I remember when Intel got into centrino branding, they didn't have the wifi capability inhouse so they bought what chips they needed from TI and Philips and repackaged/rebranded them as Intel's Centrino. When that early effort paid dividends and the a market opportunity began to show itself, only then did Intel take the development internal and design their own chips to enable the Centrino platform.

The second reason I like what I am seeing in this move is that the move itself could not have happened if the culture within AMD was unreceptive to the necessity of being nimble and trying things that are outside the bounds of what AMDers traditionally viewed their jobs being about.

Having a workforce that is willing to embrace change is key to enabling the success of the kind of change that is necessary for the company to navigate the forthcoming transition in its business model from being x86-centric to a more diverse IP design house.

The rebranded SSDs may be a flop, or a success, either way the move to sell rebranded SSDs will neither be the deathknell nor the savior of AMD. But the fact that such a shift in product offerings was possible, that it wasn't killed internally by a swarm of naysayers and can't-be-done'rs is a very good omen IMO.

 

[pablo87]

I think this is friggin' sad, in the old days AMD had their own flash technology going head to head with Intel, having major market share with OEMs all over the world, and shared a dedicated fab in Japan with Fujitsu.

Now they're relabelling someone else's products? Anybody and I mean anybody can do that.

Its embarassing.

 

[mrmt]

For one, it shows that AMD is not sitting around letting grass grow under its feet while its core revenue generator slowing erodes (x86 desktop units decrease).

(...)

The second reason I like what I am seeing in this move is that the move itself could not have happened if the culture within AMD was unreceptive to the necessity of being nimble and trying things that are outside the bounds of what AMDers traditionally viewed their jobs being about

I see your point here. From this POV, it is good news. It has the right mindset for AMD moment. I was really critical of the move because I wanted to see moves that would really address the issues with AMD core business problems, like more deals like the one with Symmetry Electronics but this time involving Kabini.

 

[Idontcare]

I think this is friggin' sad, in the old days AMD had their own flash technology going head to head with Intel, having major market share with OEMs all over the world, and shared a dedicated fab in Japan with Fujitsu.

Now they're relabelling someone else's products? Anybody and I mean anybody can do that.

Its embarassing.

There are two dominant types of Flash - NOR and NAND.

AMD was a leader in the field of NOR Flash which isn't the kind of flash that we think of when we think of flash used in storage applications like SSD's and thumbdrives.

The NAND type is primarily used in memory cards, USB flash drives, solid-state drives, and similar products, for general storage and transfer of data. The NOR type, which allows true random access and therefore direct code execution, is used as a replacement for the older EPROM and as an alternative to certain kinds of ROM applications.

NOR flash, the kind AMD made until they spun-off their flash division as Spansion (which then went bankrupt), is more akin to what HP is doing with their memristors that they are rebranding as "nanostores" (because hyping a new method of achieving the same thing that now-dead NOR flash achieved would be a deal killer in the press).

Ironically, HP calls their efforts to use memristors "Project Moonshot", which adequately captures their odds of succeeding where NOR Flash makers failed. :hmm:

Getting back to flash, NAND flash is what everyone wants now, and AMD never had much experience making that type of flash.

 

[pablo87]

There are two dominant types of Flash - NOR and NAND.

AMD was a leader in the field of NOR Flash which isn't the kind of flash that we think of when we think of flash used in storage applications like SSD's and thumbdrives.



NOR flash, the kind AMD made until they spun-off their flash division as Spansion (which then went bankrupt), is more akin to what HP is doing with their memristors that they are rebranding as "nanostores" (because hyping a new method of achieving the same thing that now-dead NOR flash achieved would be a deal killer in the press).

Ironically, HP calls their efforts to use memristors "Project Moonshot", which adequately captures their odds of succeeding where NOR Flash makers failed. :hmm:

Getting back to flash, NAND flash is what everyone wants now, and AMD never had much experience making that type of flash.

IDC, you are correct nonetheless the point I made still stands - AMD has no business being a NAND "broker" - it shows how clueless mgmt is.

 

[krumme]

I am stunned how stupid this rebrand move seems.

Even cowboy RR couldnt come up with something like this. Perhaps its the experts who was called in to help ? - it looks like a consultancy thing

I am trying to figure out the purpose. There must be something to it? - i mean like hiding some financial problems by keeping cash flow? Arg... Something we dont know??

 

[Idontcare]

I am stunned how stupid this rebrand move seems.

Even cowboy RR couldnt come up with something like this. Perhaps its the experts who was called in to help ? - it looks like a consultancy thing

I am trying to figure out the purpose. There must be something to it? - i mean like hiding some financial problems by keeping cash flow? Arg... Something we dont know??

What would you have them do? They need to take actions now that will have a bottom-line impact no farther out than 12 months from now.

Anything involving IC design won't pay dividends for 3-4 yrs. To be sure they are working on changes there too, but they've got to have something else happening in the nearer term too.

This is what happen when you have an "all hands on deck" crisis.

The guys that are trained and capable of doubling-down on stuff that will make a difference in 3 yrs are doing that, meanwhile you've got a marketing team that could add value to AMD in the short-term by structuring some rebranding projects that don't tie up engineering resources, don't require intensive capital outlays, and essentially repurpose idle hands into doing something the could possibly bring in a few more bucks and reduce the painful layoffs by a few more headcount in a year's time.

What would you have them do? Fire everyone who isn't working on a project that won't come to market in less than 3yrs, crater revenue for 36 months, and try to re-emerge out of the other side of the collapse with some hail-mary miracle product?

It may come to that, but in the meantime rebranding is but one of many ways you'd expect a company to explore to raise revenue and cash flow when things are coming down to the 11th hour.

 

[Centauri]

Agreed. It's hardly an ideal situation, but the very fact these actions are being taken should be building some confidence that people within are well aware of how bad things are at the moment.

With the roadmaps that have leaked recently, the world now knows that AMD has nothing substantial coming in 2013, and certainly nothing that will turn the entire company around. So in the meantime AMD has to do what it can to strengthen its position, to dig in, and spreading brand awareness is one very cheap and, more importantly, immediate way to do it.