[Kitguru ] AMD’s partners cannot get enough Radeon R9 Fury graphics cards –

[desprado]

Apparently, the problems of Advanced Micro Devices with its latest Radeon R9 Fury-series graphics cards may be significantly more serious than whining cooling system or limited overclockability. If a new media report is to be believed, then the company cannot supply enough boards to its partners.
It is not a secret that the launch of AMD’s latest Radeon R9 Fury-series graphics cards was not very easy for AMD. The company delayed the product for a number of times and while its performance is high, it cannot beat its direct rival in all applications. Moreover, while AMD’s decision to use innovative high-bandwidth memory allowed the company to make its new graphics cards significantly shorter, its choice to use liquid cooling did not really pay off and resulted in a rather noisy cooler. However, all the challenges that AMD faces today may be considered as insignificant as the chip designer simply cannot deliver enough Radeon R9 Fury-series graphics cards to all of its partners.
Advanced Micro Devices has over 10 add-in-card partners, who officially buy graphics processing units from the company. Virtually all allies of AMD currently ship top-of-the-range Radeon R9 Fury X graphics solutions. However, only Asustek Computer and Sapphire Technology will offer AMD Radeon R9 Fury products initially, reports Hardwareluxx. Eventually companies like Gigabyte Technology, MicroStar International, PowerColor and other will also offer AMD Radeon R9 Fury graphics cards, but at first, such products will only be available in limited quantities from two of AMD’s partners.
According to the report, yields of AMD’s code-named “Fiji” graphics processing unit are rather low. Insufficient yields are not something surprising: with 8.9 billion transistors inside, the “Fiji” is the most complex chip ever produced. While the IC [integrated circuit] is not as large as Nvidia Corp.’s GM200, it is considerably harder to produce because of higher transistor density. Moreover, since “Fiji” uses all-new high-bandwidth memory (HBM) as well as a special interposer to connect memory to the GPU, testing and packaging process of the chip is extremely complex.
The exact yield rate of AMD’s “Fiji” is uncertain and it is unclear how many chips Advanced Micro Devices can get from its partners at Taiwan Semiconductor Manufacturing Co. Moreover, since cycle times of TSMC’s 28nm fabrication process are over two months, it is clear that AMD cannot solve all of its problems quickly.
AMD and its partners did not comment on the news-story.

http://www.kitguru.net/components/g...-enough-radeon-r9-fury-graphics-cards-report/

Same as reported by Wccf.
http://wccftech.com/amd-shipped-lim...custom-boards-xfx-club3d-powercolor-arriving/

 

[werepossum]

Oddly, a quick check of NewEgg shows six different manufacturers of Fury X cards, five of which have reviews and none of which are in stock. So while we can say that AMD is having problems delivering Fury cards, it appears that is because of Fury X card sales. If one must have a problem, high non-defective yield and high sales are good problems to have.

 

[Keysplayr]

Oddly, a quick check of NewEgg shows six different manufacturers of Fury X cards, five of which have reviews and none of which are in stock. So while we can say that AMD is having problems delivering Fury cards, it appears that is because of Fury X card sales. If one must have a problem, high non-defective yield and high sales are good problems to have.

Can you say exactly how many FuryX cards Newegg sold across all manufacturers? Or are you assuming that since they are out of stock, it has to be because they are selling like hotcakes, and definitely not a supply problem?

 

[ShintaiDK]

We will see at the steam survey (If the Fury series can get over the minimum limit). Because like the marketshare it disagrees with the notions that people got on this forum about how things actually sells. Entire R9 200 series for once.

 

[NTMBK]

We saw exactly the same doom and gloom stories about the 970 and 980. http://wccftech.com/nvidia-may-be-facing-gtx-980-gtx-970-supply-issues/ We aren't in a position to know whether this shortage is supply or demand driven. Wait and see.

 

[Silverforce11]

At first it was claimed that Fury Pro won't be any good, hence, only two AIBs (ASUS & Sapphire) have SKUs. Reviews done and dusted, the vast majority are positive.

Now its only 2 AIBs have SKUs due to lack of yields. On a cut-down part. Lack of yields on a cut down part.

It would make more sense if they speculate that HBM yields are rubbish and is limiting the entire chain of supply so we don't get good volume of SKUs using it.

While we're on this trend of speculative random stuff, let's throw this one to the mix.

According the usual rumor click bait:
http://wccftech.com/amd-working-entire-range-hbm-gpus-follow-fiji-fury-lineup/

They are suggesting that AMD is going to play very dirty with HBM2. They will leverage their close relationship with SK Hynix to get first dibs on HBM2 chips, all of it, before NV is allowed to get their hands on them. This means they will have first to market next-gen Artic Islands + HBM2 GPUs.

 

[amenx]

Oddly, a quick check of NewEgg shows six different manufacturers of Fury X cards, five of which have reviews and none of which are in stock. So while we can say that AMD is having problems delivering Fury cards, it appears that is because of Fury X card sales. If one must have a problem, high non-defective yield and high sales are good problems to have.

Its all relative. If they could only produce 20 cards which were sold out, I dont think they would be smiling. "Sold out" can only have positive significance if they were produced in reasonable quantity. To be fair, they need a bit more time.

 

[Silverforce11]

If it helps the discussion, retailers here that I've talked to, said they got a single shipment of Fury X which was sold out instantly but now their distributors cannot even confirm an ETA on the next shipments. A few of them halted pre-orders completely because they have a lot on the list and they have no clue when its going to be in stock.

What this says to me is that Fury X volume is indeed low, because the market for top-grade $$ GPUs simply isn't that big. Lots of 980Ti/Titan X are in stock ever since launch here & we know that it sells very well.

 

[n0x1ous]

The good news is all the bad pumps should be out of the channel by the time more stock comes in for the X

 

[werepossum]

Can you say exactly how many FuryX cards Newegg sold across all manufacturers? Or are you assuming that since they are out of stock, it has to be because they are selling like hotcakes, and definitely not a supply problem?

They are definitely selling like hotcakes, but my impression is that there are very few to be sold. So some of both, but then, doesn't AMD usually have supply issues with new releases?

Its all relative. If they could only produce 20 cards which were sold out, I dont think they would be smiling. "Sold out" can only have positive significance if they were produced in reasonable quantity. To be fair, they need a bit more time.

True, on both accounts.

My point here was that whatever chips they have, it's better to sell them for the higher priced unit than for the lower, assuming they make the grade. Until and unless supply for the higher priced unit outstrips demand, there is little point in supplying the cheaper unit, unless of course there are significant quantities that have to partially disabled.

 

[n0x1ous]

Looks like Powercolor has one coming

http://www.guru3d.com/news-story/powercolor-r9-fury-4gb-hbm-announced.html

 

[DarkKnightDude]

Maybe HBM is in limited supply?

 

[Stuka87]

Just a new site invented news most likely. Every time a new card comes out, and there is no stock, they go on about how yields are terrible resulting in so production numbers. Same story, different GPU.

 

[Kenmitch]

While we're on this trend of speculative random stuff, let's throw this one to the mix.

According the usual rumor click bait:
http://wccftech.com/amd-working-entire-range-hbm-gpus-follow-fiji-fury-lineup/

They are suggesting that AMD is going to play very dirty with HBM2. They will leverage their close relationship with SK Hynix to get first dibs on HBM2 chips, all of it, before NV is allowed to get their hands on them. This means they will have first to market next-gen Artic Islands + HBM2 GPUs.

They should get 1st dibs. If they can suck it all up then so be it. Would be shocking to say the least.

As far as supply goes....The pump issue could be a factor as far as the supply chain inventory goes.

 

[Stuka87]

They should get 1st dibs. If they can suck it all up then so be it. Would be shocking to say the least

Exactly, they helped develop it, why give NV early access to AMD's technology. Just like GDDR3 and GDDR5.

 

[Azix]

We saw exactly the same doom and gloom stories about the 970 and 980. http://wccftech.com/nvidia-may-be-facing-gtx-980-gtx-970-supply-issues/ We aren't in a position to know whether this shortage is supply or demand driven. Wait and see.

I was thinking about that. These shortages seem to always happen and this time AMD has three cards on the same chip. Nvidia only ever had 2 at most each time.

But because of the double standards, nvidia running out of stock is due to sales, AMD running out of stock is yield issues.

AMD should get the benefit from HBM2 first. Its the positive side of hardware innovation. More so I want that to happen so those who say AMD does nothing can continue "enjoying" failworks while AMD customers buy newer faster chips.

 

[Dribble]

Exactly, they helped develop it, why give NV early access to AMD's technology. Just like GDDR3 and GDDR5.

Only it's not AMD's technology, it's SK Hynix's. If AMD has a contract that guarantees them better treatment of some form which they got due to their earily adoption of HBM 1 then yes they will get that better treatment. If not then all SK Hynix care about is making money by selling as many HBM chips for as much as possible. Those chips will go first to those making the biggest orders at the highest prices.

 

[Meekers]

Only it's not AMD's technology, it's SK Hynix's. If AMD has a contract that guarantees them better treatment of some form which they got due to their earily adoption of HBM 1 then yes they will get that better treatment. If not then all SK Hynix care about is making money by selling as many HBM chips for as much as possible. Those chips will go first to those making the biggest orders at the highest prices.

It was developed jointly between AMD and Hynix from all the information that I have seen. I have no idea what sort of arrangement they have as far as ownership of IP, but from a couple google searches it is speculated that AMD owns some amount of the IP.

 

[Atreidin]

Only it's not AMD's technology, it's SK Hynix's. If AMD has a contract that guarantees them better treatment of some form which they got due to their earily adoption of HBM 1 then yes they will get that better treatment. If not then all SK Hynix care about is making money by selling as many HBM chips for as much as possible. Those chips will go first to those making the biggest orders at the highest prices.

From everywhere I have read, HBM was co-developed with AMD, as in, AMD was directly involved in the design and specifications.

 

[Sabrewings]

From everywhere I have read, HBM was co-developed with AMD, as in, AMD was directly involved in the design and specifications.

That doesn't mean AMD owns IP in it. Nvidia has already said Pascal will use HBM2 and with development time they have to have it worked out by now at the silicon level. If there was an IP issue they would have hit it by now.

SK Hynix will probably sell to who pays the most.

 

[Meekers]

That doesn't mean AMD owns IP in it. Nvidia has already said Pascal will use HBM2 and with development time they have to have it worked out by now at the silicon level. If there was an IP issue they would have hit it by now.

SK Hynix will probably sell to who pays the most.

Except that they have allegedly hit an issue. It is being reported that AMD can get all the HBM2 they want before NVIDIA. We also do not know if NVIDIA is paying any licensing costs.


Here is a video by AMD, so take it with a huge grain of salt. In it they say they started work on HBM in 2008 and only a couple years later brought in Hynix. https://www.youtube.com/watch?v=se9TSUfZ6i0

 

[Sabrewings]

It is being reported that AMD can get all the HBM2 they want before NVIDIA.

Am I missing the reliable source (not WCCFTech)?

 

[Stuka87]

Only it's not AMD's technology, it's SK Hynix's. If AMD has a contract that guarantees them better treatment of some form which they got due to their earily adoption of HBM 1 then yes they will get that better treatment. If not then all SK Hynix care about is making money by selling as many HBM chips for as much as possible. Those chips will go first to those making the biggest orders at the highest prices.

AMD has been very clear about the fact that they invented HBM. Just like they invented GDDR3 and GDDR5. This is not hearsay.

Quote from AMD's site: "GDDR and now HBM: Pervasive industry standards for high-performance memory, invented by AMD with contributions from the Joint Electron Device Engineering Council (JEDEC) and industry partners"

They were the ones that brought it into the standards committee.

 

[Sabrewings]

Yep, and Nvidia is definitely having a hard time getting GDDR5. Oh, wait.

 

[sm625]

I'd be surprised if anyone was surprised by this. I would be even more surprised if they've even produced more than 10000 Fiji chips as of today.