Thank you, dunno why people read so much into it.
Chances are wccf are faking it all anyway.
-
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.
AMD Polaris architecture – GCN 4.0
- Thread starter Boomstick777
- Start date
Thank you, dunno why people read so much into it.
Chances are wccf are faking it all anyway.
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
I have a feeling AMD held back the FUry X2 because it makes more sense to release it as the fastest card of the new gen than release it now and have the new gen fastest part be slower than it.
IE they release Fury X2 in December, then the medium Polaris part comes this year and is a good 30% slower than it... looks bad. Or you hold back Fury X2 and this year you release low end Polaris, medium Polaris maybe 20% faster than Fury at half the size and then Fury X2, another 30% faster than medium Polaris.
It's also quite possible that Fury has a fair amount of hidden DP performance and the Fury X2 gets released as a firepro card with epic DP performance as AMD continue to make fairly significant market share gain in the professional market. Then this is the expensive card we see allocated as part of the new range in terms of number part.
IE they release Fury X2 in December, then the medium Polaris part comes this year and is a good 30% slower than it... looks bad. Or you hold back Fury X2 and this year you release low end Polaris, medium Polaris maybe 20% faster than Fury at half the size and then Fury X2, another 30% faster than medium Polaris.
It's also quite possible that Fury has a fair amount of hidden DP performance and the Fury X2 gets released as a firepro card with epic DP performance as AMD continue to make fairly significant market share gain in the professional market. Then this is the expensive card we see allocated as part of the new range in terms of number part.
Last edited:
Its said that if you buy a Polaris you get laid a lot.
thats how sexy it is to own.
Roy talks some VR and how far ahead they are vs never taped out Pascal.
http://www.gamecrate.com/interview-amds-roy-taylor-dawn-virtual-reality-age/12842
The future is brighter with AMD
thats how sexy it is to own.
Roy talks some VR and how far ahead they are vs never taped out Pascal.
http://www.gamecrate.com/interview-amds-roy-taylor-dawn-virtual-reality-age/12842
The future is brighter with AMD
Its said that if you buy a Polaris you get laid a lot.
thats how sexy it is to own.
Roy talks some VR and how far ahead they are vs never taped out Pascal.
http://www.gamecrate.com/interview-amds-roy-taylor-dawn-virtual-reality-age/12842
The future is brighter with AMD
Soldato
- Joined
- 19 Feb 2011
- Posts
- 5,849
yeah except its Roy talking so grains of salt needed to be used, he often talks out of his rear.
Hope theres some elements of truth with regards to good price.
Hope theres some elements of truth with regards to good price.
Paper launch, but yes, Samsung 14nm LPP is half a year ahead of tsmc 16nm FF+.
Let's hope AMD are able to maximise whatever window of opportunity TSMC are able to give them
They could use a shot in the arm right now. Hopefully AMD don't snatch defeat from the jaws of victory, by mucking something up.
Paper launch
Which is irrelevant because Nvidia weren't trying to launch Pascal 6 months ago when they would have experienced a delay. Since TSMC is in full volume production their delay is no consequence to Nvidia in the slightest.
Let's hope AMD are able to maximise whatever window of opportunity TSMC are able to give them
They could use a shot in the arm right now. Hopefully AMD don't snatch defeat from the jaws of victory, by mucking something up.
Yeh hopefully they'll be able to use this delay to their advantage
i watched the video with roy while eating some sugarpuffs and didnt spit any out
so yeh
i think AMD talking about the right things
the software is super important
and a 970 but much cheaper would get crazy sales
but not going to lie i hope they bring something a lot faster too
so yeh
i think AMD talking about the right things
the software is super important
and a 970 but much cheaper would get crazy sales
but not going to lie i hope they bring something a lot faster too
I still see people talking Samsung... there's no proof whatsoever AMD will use them for Polaris. GF, which they have confirmed they ARE using, is pretty much neck and neck with TSMC, although TSMC are far more competent.
Funny how Roy said ahead to market btw, they haven't hit the market yet.
Funny how Roy said ahead to market btw, they haven't hit the market yet.
Last edited:
Soldato
- Joined
- 30 Nov 2011
- Posts
- 11,499
If they are low power parts coming first then it could well be Samsung LPP (which GF have licenced so they could be referring to the technology rather than the physical foundry). But they'll be low power, so we aren't talking Fury replacements, but that would explain the comparisons to 950 type GPU's.
But it would also mean basically having two completely different designs on two different processes for their high/mid range parts and their low end low power parts, and taking a big chunk of work away from TSMC isn't exactly going to endear them as customers.
Last edited:
http://www.techtimes.com/articles/118904/20151224/amd-partners-with-samsung-globalfoundries-to-build-new-14nm-gpu-report.htm
"Because Samsung Electronics and GF have same IP for 14-nano processing, chips that are designed by AMD will all be produced at both factories," a person familiar with the matter tells Electronic Times. "If products are produced from both factories, AMD won't have to worry about a problem regarding lack of supplies."
http://www.techpowerup.com/218578/samsung-to-fab-amd-zen-and-arctic-islands-on-its-14-nm-finfet-node.html
"It has been confirmed that Samsung will be AMD's foundry partner for its next generation GPUs......It gets better - not only will Samsung manufacture AMD's next-gen GPUs, but also its upcoming "Zen" family of CPUs, at least a portion of it. AMD is looking to distribute manufacturing loads between two foundries, Samsung and GlobalFoundries, perhaps to ensure that foundry-level teething trouble doesn't throw its product launch cycle off the rails. "
It could be tho that samsung will make ZEN, and GloFo will make the GPUs
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
Samsung are going to be making early AMD parts I would imagine and moving to GloFo as and when capacity becomes available.
There is a reason that TSMC/Samsung started production and neither AMD or Nvidia have parts out, the industry and AMD/Nvidia's position in it has changed entirely.
AT 55nm those two used up most of the early capacity of a process for two reasons, yields on bigger chips were better back then and they absolutely require making profit on every chip. Newer processes have worsening yields on bigger chips and as such the point a process becomes profitable comes much later on for bigger chips than it used to. Mobile chips dominate new process capacity for the opposite reasons, they are smaller and don't need to make a profit so can start instantly. AMD and Nvidia sell chips for profit, that is their core business, selling chips at a loss does nothing for them. Apple/Samsung make chips that cost maybe $20, their biggest customer is themselves and having a superior chip means selling more of their own devices that cost $200 to make and sell for $600. Their profit comes from devices, if that $20 chip makes $10 profit or a $20 loss doesn't matter as long as it enables a $400 profit from a phone sale they otherwise wouldn't get.
AMD/Nvidia now have to wait for capacity on a new process to exceed the requirements of those who can pay more for the wafers because profit on the silicon itself isn't their core business.
D.P doesn't seem to get this as he thinks TSMC being later doesn't matter, except by his version of events AMD/Nvidia should be shipping today, there is a reason they aren't and that is production capacity.
Production capacity doesn't increase incrementally over time, it happens in giant leaps but takes time. All fabs only start with limited capacity as you need only a small number of lines working making mostly broken chips till you get the 'recipe' right and they start working well. They you start full scale production but it's still very low capacity, you can't flick a switch and have more lines. A line has multiple machines all costing 10's of millions each which take time to build, ship, install and then tune all the machines to get equally good yields. So you might start day one with 5k wafers a month and 5 months later it's still 5k wafers, but a month after that 10 more production lines are ready and you jump to 50k wafers a month.
Apple will pay double the cost for a wafer start at TSMC over Nvidia because every wafer means another maybe 400-500 devices they can sell at a profit, double the cost of the wafer has a small hit on overall profit for them. Any company whose core business is making profit per chip will lose out to a customer who is making commodity chips where profit on the chips themselves doesn't matter. So if TSMC are 6 months later to production starting, they are almost certainly 6 months behind in getting to enough capacity for someone other than Apple/other mobile chip makers to be making chips. There is also a reason so many people shifted from TSMC to Samsung, they have GloFo as a second source and they will get up to capacity earlier.
The industry has completely changed in this sense, AMD/Nvidia used to drive the new processes and be amongst the first chips out, they pretty much never will again.... unless Apple buy their own fabs in the future which is possible if not probable.
There is a reason that TSMC/Samsung started production and neither AMD or Nvidia have parts out, the industry and AMD/Nvidia's position in it has changed entirely.
AT 55nm those two used up most of the early capacity of a process for two reasons, yields on bigger chips were better back then and they absolutely require making profit on every chip. Newer processes have worsening yields on bigger chips and as such the point a process becomes profitable comes much later on for bigger chips than it used to. Mobile chips dominate new process capacity for the opposite reasons, they are smaller and don't need to make a profit so can start instantly. AMD and Nvidia sell chips for profit, that is their core business, selling chips at a loss does nothing for them. Apple/Samsung make chips that cost maybe $20, their biggest customer is themselves and having a superior chip means selling more of their own devices that cost $200 to make and sell for $600. Their profit comes from devices, if that $20 chip makes $10 profit or a $20 loss doesn't matter as long as it enables a $400 profit from a phone sale they otherwise wouldn't get.
AMD/Nvidia now have to wait for capacity on a new process to exceed the requirements of those who can pay more for the wafers because profit on the silicon itself isn't their core business.
D.P doesn't seem to get this as he thinks TSMC being later doesn't matter, except by his version of events AMD/Nvidia should be shipping today, there is a reason they aren't and that is production capacity.
Production capacity doesn't increase incrementally over time, it happens in giant leaps but takes time. All fabs only start with limited capacity as you need only a small number of lines working making mostly broken chips till you get the 'recipe' right and they start working well. They you start full scale production but it's still very low capacity, you can't flick a switch and have more lines. A line has multiple machines all costing 10's of millions each which take time to build, ship, install and then tune all the machines to get equally good yields. So you might start day one with 5k wafers a month and 5 months later it's still 5k wafers, but a month after that 10 more production lines are ready and you jump to 50k wafers a month.
Apple will pay double the cost for a wafer start at TSMC over Nvidia because every wafer means another maybe 400-500 devices they can sell at a profit, double the cost of the wafer has a small hit on overall profit for them. Any company whose core business is making profit per chip will lose out to a customer who is making commodity chips where profit on the chips themselves doesn't matter. So if TSMC are 6 months later to production starting, they are almost certainly 6 months behind in getting to enough capacity for someone other than Apple/other mobile chip makers to be making chips. There is also a reason so many people shifted from TSMC to Samsung, they have GloFo as a second source and they will get up to capacity earlier.
The industry has completely changed in this sense, AMD/Nvidia used to drive the new processes and be amongst the first chips out, they pretty much never will again.... unless Apple buy their own fabs in the future which is possible if not probable.
Last edited: