AMD HD7XXX Series

[DotA]

Yeah just reading up on it and primary focus for power is consolidation and refinement of current standards tho it doesn't rule out extending the spec.

Found the article, its from bit tech.

However, PCI-E Express 3.0 provides no more power than its predecesors - but that's actually a good thing. Having >300W drawn from each 16xPCI-E slot would drive up the cost of motherboards, as the copper traces on the motherboard would need to be thicker. Manufacturers might even need to add extra layers to try and route these high power traces and their associated electromagnetic interference away from sensitive data traces.

http://www.bit-tech.net/hardware/2010/11/27/pci-express-3-0-explained/1

 

[bru]

I knew id read it somewhere

http://www.pcisig.com/news_room/faqs/pcie3.0_faq/#EQ15

third question up from the bottom

Q: Does PCIe 3.0 enable greater power delivery to cards?
A: The PCIe Card Electromechanical (CEM) 3.0 specification consolidates all previous form factor power delivery specifications, including the 150W and the 300W specifications.

 

[Jay85]

You know if the next gen cards are only 20-25% faster i don't see why they don't just wait a year or more and wait till they can get greater more significant gains and slightly Lower temps to follow. Seems abit silly upgrading for the sake of 20% if you already have say a GTX 560-580 which are already silly fast cards.

 

[GJ02]

You know if the next gen cards are only 20-25% faster i don't see why they don't just wait a year or more and wait till they can get greater more significant gains and slightly Lower temps to follow. Seems abit silly upgrading for the sake of 20% if you already have say a GTX 560-580 which are already silly fast cards.

becuse they know silly people like me will buy them

and then buy the next gen cards too

 

[Jay85]

Damn it haha, it fustrates me because of how this market operates. I too upgraded from a 5870 which is a perfectly powerful and still very much maxes out every game I throw at it. We're all victims of the super corperations that govern our spending subconsciously.

 

[GJ02]

We're all victims of the super corperations that govern our spending subconsciously.

not all, just the ones who are the must have's

 

[Jay85]

not all, just the ones who are the must have's

Lol that's me and you then

 

[Rroff]

I knew id read it somewhere

http://www.pcisig.com/news_room/faqs/pcie3.0_faq/#EQ15

third question up from the bottom

I find this a bit odd as there isn't a 300watt specification despite it being mentioned a lot - AFAIK 150watt one only exists due to legacy issues.

 

[Shayper]

becuse they know silly people like me will buy them

and then buy the next gen cards too

This

When the 680 comes out, may try and mod my 580 block to fit it

 

[Freddie1980]

28nm Radeon Chips This year Says AMD

They already have working samples and their going to be producing these chips at TSMC and Global Foundries (although each factory will be producing different chips so you won't get 7900's from both TSMC and Goflo).

Its also interesting to note that in the last quarter their GPU business lost $7 million, it looks like this 28nm switch can't come fast enough.

 

[bru]

Its also interesting to note that in the last quarter their GPU business lost $7 million, it looks like this 28nm switch can't come fast enough.

that is quite surprising, i would have thought they would have done OK with their currant range being good and competitive in all the right places especially at the low end where the volume sales are.

Anyone have any idea how Nvidia did for the same quarter..?

 

[An Exception]

28nm Radeon Chips This year Says AMD

They already have working samples and their going to be producing these chips at TSMC and Global Foundries (although each factory will be producing different chips so you won't get 7900's from both TSMC and Goflo).

Its also interesting to note that in the last quarter their GPU business lost $7 million, it looks like this 28nm switch can't come fast enough.

The -7m is probably due to the costs of fusion GPU R&D being accounted to the GPU side with the sales revenue of the GPU portion of the chips being accounted to the CPU side of the company...

 

[shaffaaf27]

hmm why would they want to show in the books that the gpu section is loosing money though? Then again I assume most shareholders wants the CPU business to thrive, rather than the GPU.

 

[Freddie1980]

I doubt its due to accounting rules on AMD's APU's, its more to do with the fact 40nm is end of life now and more to do with shrinking margins. Its a safe bet to say a 6970 costs more to make then 5870, twice the memeory, a bigger die size sold at a lower price along with price cuts due to competition, apply that logic to the rest of line and it goes a long to explain the loss. Its natrual, profits will return once 28nm ramps up.

 

[An Exception]

hmm why would they want to show in the books that the gpu section is loosing money though? Then again I assume most shareholders wants the CPU business to thrive, rather than the GPU.

The problem is how do you account for Fusion chips? do you place their revenue in the CPU or GPU side of the business?

Also Freddie has a point, but yields and capacity are now better than when the 5000 series was around, so that should offset that, as seriously doubt AMD are actually selling the chips themselves at a loss, especially as they clearly don't have to, as stock levels are pretty low especially since the recent explosion in demand for AMD GPU's due to Bitcoin mining fast becoming the new craze.

Edit:

Maybe AMD have also sunk allot of cash in 28nm silicon allocation both at TSMC and Glofo as the last rumour I read indicate AMD will actually be using both fabs for 28nm GPU's...

 

[Freddie1980]

The problem is how do you account for Fusion chips? do you place their revenue in the CPU or GPU side of the business?

Also Freddie has a point, but yields and capacity are now better than when the 5000 series was around, so that should offset that, as seriously doubt AMD are actually selling the chips themselves at a loss, especially as they clearly don't have to, as stock levels are pretty low especially since the recent explosion in demand for AMD GPU's due to Bitcoin mining fast becoming the new craze.

Edit:

Maybe AMD have also sunk allot of cash in 28nm silicon allocation both at TSMC and Glofo as the last rumour I read indicate AMD will actually be using both fabs for 28nm GPU's...

And the fact they are using two fabs might also explain another problem AMD (as well as Nvidia) have - supply. If your a designer without a manufacturing arm your pretty much reliant on TSMC for production as there's not really anywhere else to go.

TSMC might have sorted out their 40nm yield issues long ago but given that Apple uses them to knock out all their ARM chips for the iphones, ipads and ipods I tend to think supply is limited (also TSMC supply's the likes of Qualcomm, Marvall and VIA). With a limited supply of waffers TSMC can afford to keep prices high whilst AMD has to cut the price periodically on its Radeon's to keep them competitive.

If AMD are losing $7 million this quarter I dread to think how much Nvidia are losing on its consumer video card line.