• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Pascal Refresh to 14nm in 2017

You know AMD are sucking hard when Nvidia can rehash Pascal from 16nm to 14nm and flog it off to be devoured by their customers lol...

Sad state of affairs if this is just a Pascal refresh, it would show just how far AMD have actually fallen than Nvidia could do this and shovel it out the door and still sell it enmasse.

If its Volta though this could be interesting.
 
SiDeards73 said:
You know AMD are sucking hard when Nvidia can rehash Pascal from 16nm to 14nm and flog it off to be devoured by their customers lol...

Sad state of affairs if this is just a Pascal refresh, it would show just how far AMD have actually fallen than Nvidia could do this and shovel it out the door and still sell it enmasse.

If its Volta though this could be interesting.
Click to expand...

LIterally what on earth are you talking about? Exactly how is this different from selling Pascal on 16nm? What has it got to do with AMD at all? How does it require AMD to have 'fallen' for Nvidia to use a different process? The logically equivalent arguments are either Nvidia had fallen because AMD were also using TSMC, or Nvidia had fallen because AMD were able to produce cards cards on 14nm to be 'devoured' by their customers.

Seriously, explain the logic or thinking here.


If Nvidia move Pascal to to Samsung, it's because they can't get volume at TSMC, it means AMD made the move to Samsung because they knew TSMC were Apple's ***** and didn't want to be stuck with TSMC giving up all their priority capacity to Apple. It means Nvidia made the same decision extremely late and is responsible for the shocking 1080 supply to date and likely responsible for Nvidia making Titan X only available through them in small quantities.
 
drunkenmaster said:
LIterally what on earth are you talking about? Exactly how is this different from selling Pascal on 16nm? What has it got to do with AMD at all? How does it require AMD to have 'fallen' for Nvidia to use a different process? The logically equivalent arguments are either Nvidia had fallen because AMD were also using TSMC, or Nvidia had fallen because AMD were able to produce cards cards on 14nm to be 'devoured' by their customers.

Seriously, explain the logic or thinking here.


If Nvidia move Pascal to to Samsung, it's because they can't get volume at TSMC, it means AMD made the move to Samsung because they knew TSMC were Apple's ***** and didn't want to be stuck with TSMC giving up all their priority capacity to Apple. It means Nvidia made the same decision extremely late and is responsible for the shocking 1080 supply to date and likely responsible for Nvidia making Titan X only available through them in small quantities.
Click to expand...

If its purely driven by supply from TSMC then thats a different matter, if its just Nvidia chancing their arm, refreshing Pascal onto a smaller node at Samsung to tide them over til Volta, then you gotta say bravo to Nvidia for being able to pull off such a move, and then look at why AMD has not been able to compete...

The way i see it is Nvidia have a free reign on the GPU market currently and are intent on rinsing as much as they can out of the Pascal Architechture to carry them over til they absolutely have to release something newer, its perfect business sense, why rush something newer out the door when your competitors cannot compete? you rehash what you already have, slap a shiney badge on it and tell everyone how wonderful it is and reap the rewards, much like they Pascal being pretty much a shrink of Maxwell...

Hell we all know AMD and Nvidia like a refresh, you only need to look at Hawaii for the truth in that lol
 
Nv's decision to exclusively punt TXP isn't down to yields it's down to maximising profit on a desirable low volume part-meanwhile AMD's still standing in the corner twiddling it's thumbs casting a jealous gaze in Nv's direction...
 
SiDeards73 said:
If its purely driven by supply from TSMC then thats a different matter, if its just Nvidia chancing their arm, refreshing Pascal onto a smaller node at Samsung to tide them over til Volta, then you gotta say bravo to Nvidia for being able to pull off such a move, and then look at why AMD has not been able to compete...
Click to expand...

Again, refreshing Pascal on an extremely similar node would be almost identical chips, according to some people they would undoubtedly have worse power and clock speeds than TSMC.

If it's a faster architecture then it's not Pascal, if it's literally Pascal, then there isn't going to be magical massive speed improvements because.... it's the same chips. A refresh is generally a 5-10% speed boost, if that, and only bigger if you completely screwed up the first one as with Fermi.

There is no reason to believe a refresh means dramatically faster chips and it has precisely nothing to do with AMD so I don't know why you're talking about how much AMD have fallen. AMD bring out Vega and Nvidia switch fabs, nothing more or less.

For Nvidia to switch to Samsung would be 99.9% likely about price of wafers, yields at TSMC being significantly worse(which is effectively wafer pricing anyway) or lack of capacity either due to real lack of capacity, or that all the capacity is being taken by Apple.

If these were 'just' faster chips, why wouldn't they do them at TSMC? This isn't about faster newer chips, it's about a reason to make them at Samsung rather than TSMC.
 
SiDeards73 said:
If its purely driven by supply from TSMC then thats a different matter, if its just Nvidia chancing their arm, refreshing Pascal onto a smaller node at Samsung to tide them over til Volta, then you gotta say bravo to Nvidia for being able to pull off such a move, and then look at why AMD has not been able to compete...

The way i see it is Nvidia have a free reign on the GPU market currently and are intent on rinsing as much as they can out of the Pascal Architechture to carry them over til they absolutely have to release something newer, its perfect business sense, why rush something newer out the door when your competitors cannot compete? you rehash what you already have, slap a shiney badge on it and tell everyone how wonderful it is and reap the rewards, much like they Pascal being pretty much a shrink of Maxwell...

Hell we all know AMD and Nvidia like a refresh, you only need to look at Hawaii for the truth in that lol
Click to expand...




Samsung doesn't have a smaller node.
14 and 16nm are just marketing names. In fact when I last looked into this the TSMC 16nm process was smaller than the samsung 16nm.
 
drunkenmaster said:
If Nvidia move Pascal to to Samsung, it's because they can't get volume at TSMC, it means AMD made the move to Samsung because they knew TSMC were Apple's ***** and didn't want to be stuck with TSMC giving up all their priority capacity to Apple. It means Nvidia made the same decision extremely late and is responsible for the shocking 1080 supply to date and likely responsible for Nvidia making Titan X only available through them in small quantities.
Click to expand...

TSMC has been trying to entice certain customers with price drops on 16nm due to spare capacity going forward - while unlikely maybe as that discount doesn't apply to nVidia they are trying to twist TSMC's arm to keep them.

I think it more likely if there is any truth to it that it is a specific product for a specific market and not a move for future GeForce GPUs.
 
D.P. said:
Samsung doesn't have a smaller node.
14 and 16nm are just marketing names. In fact when I last looked into this the TSMC 16nm process was smaller than the samsung 16nm.
Click to expand...

Second gen 14nm is a bit smaller than 16nm, but based on AMD's (and Apple's) chips I'd still argue that it is worse than TSMC's process. They are both still the same as 20nm though: "The back-end of line (BEOL) processes are more or less the same for 20, 16nm and 14nm technologies provided by the foundries."(source)

Cell-SizeComparison.png
 
Samwell said:
It's neither Pascal refresh or Volta. GP107/GP108 are produced at Samsung.
Click to expand...


I didnt know this , all i did was post an article for people to put there opinions across . dont make any sense to me , the reason why i posted it as wanting to upgrade to 2x pascal titans and put them under water , with this article im thinking there will me more pascal card releases and more powerful that the Titan PX is. so i will now wait and save my money and wait it out .
 
killswitch73 said:
I didnt know this , all i did was post an article for people to put there opinions across . dont make any sense to me , the reason why i posted it as wanting to upgrade to 2x pascal titans and put them under water , with this article im thinking there will me more pascal card releases and more powerful that the Titan PX is. so i will now wait and save my money and wait it out .
Click to expand...

Keep on waiting and your'll never buy anything!

Enjoy the moment now. After all it's only money. When something better comes out I'll buy that as well.
 
SiDeards73 said:
.

The way i see it is Nvidia have a free reign on the GPU market currently and are intent on rinsing as much as they can out of the Pascal Architechture to carry them over til they absolutely have to release something newer, its perfect business sense,
Click to expand...

+1

If Nvidia do not need to do more than this to win out against whatever AMD have on offer then that is what they will do and nothing more, They will sit on the same tech for as long as possible, As written, it makes sense from a business perspective.
 
SiDeards73 said:
You know AMD are sucking hard when Nvidia can rehash Pascal from 16nm to 14nm and flog it off to be devoured by their customers lol...

Sad state of affairs if this is just a Pascal refresh, it would show just how far AMD have actually fallen than Nvidia could do this and shovel it out the door and still sell it enmasse.

If its Volta though this could be interesting.
Click to expand...

It's the same 14NM process as GF is using since it was licenced from Samsung and AMD has access to Samsung fabs too:

http://www.forbes.com/sites/patrick...14nm-manufacturing-with-samsung/#21a6587a50fb
 
Rroff said:
TSMC has been trying to entice certain customers with price drops on 16nm due to spare capacity going forward - while unlikely maybe as that discount doesn't apply to nVidia they are trying to twist TSMC's arm to keep them.

I think it more likely if there is any truth to it that it is a specific product for a specific market and not a move for future GeForce GPUs.
Click to expand...
You are forgetting the new consoles like the XBox One S are being made on TSMC 16NM so they are in line for some major contracts.

The article says Pascal GPUs this year - it's most likely to be the GP107 and GP108.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom