Monday, December 5th 2022

AMD Still Believes in Moore's Law, Unlike NVIDIA

Back in September, NVIDIA's Jensen Huang said that Moore's Law is dead, but it seems like AMD disagrees with NVIDIA, at least for now. According to an interview with AMD's CTO, Mark Papermaster, AMD still believes that Moore's Law will be alive for another six to eight years. However, AMD no longer believes that transistor density can be doubled every 18 to 24 months, while remaining in the same cost envelope. "I can see exciting new transistor technology for the next - as far as you can really plot these things out - about six to eight years, and it's very, very clear to me the advances that we're going to make to keep improving the transistor technology, but they're more expensive," Papermaster said.

AMD believes we'll see a change in how chips are being designed and put together, with chiplets being the future of semiconductors. Papermaster calls this "a Moore's Law equivalent, meaning that you continue to really double that capability every 18 to 24 months" although it's not exactly Moore's Law in the traditional sense. AMD also appears to be betting heavily on FPGA technology in some of its market segments, for something the company calls adaptive computing. As to how things will play out, time will tell, but with both AMD and Intel going down the chiplet route, albeit in slightly different ways, we should continue to see new innovations from both companies, with or without Moore's Law.
Source: The Register
Add your own comment

42 Comments on AMD Still Believes in Moore's Law, Unlike NVIDIA

#26
Jism
RavenlordMoore's Law is biggest joke in a whole IT. I never liked this fake "law" which isn't law at all. 100% dependent on how tech companies are doing.
So tell me whats beyond 1nm then?
Posted on Reply
#27
TheoneandonlyMrK
JismSo tell me whats beyond 1nm then?
Angstrom level part's obviously.
Posted on Reply
#28
cyneater
Pretty sure Nvidias version of moores law is we can charge twice as much for this generation and then double the price every 12 months
Posted on Reply
#29
1d10t
You can always go horizontal ( i.e modular ) or vertical ( i.e stacking ), performance might regressed, but at least there's a way of jumping brick walls.
Posted on Reply
#30
Ravenlord
JismSo tell me whats beyond 1nm then?
It has nothing to do with this fake law. Still 100% dependent on how tech companies are doing - ask them.
Posted on Reply
#31
Assimilator
RavenlordIt has nothing to do with this fake law. Still 100% dependent on how tech companies are doing - ask them.
Keep on ignoring the laws of physics.
Posted on Reply
#32
Ravenlord
AssimilatorKeep on ignoring the laws of physics.
What does Moor's law have to do with the laws of physics? Please enlight me.
Posted on Reply
#33
Assimilator
RavenlordWhat does Moor's law have to do with the laws of physics? Please enlight me.
Moore's law is the observation that the number of transistors in a dense integrated circuit (IC) doubles about every two years.
Once we hit the wall of how precisely we can build tools to physically etch silicon, we are no longer able to increase transistor density. And we are currently almost at that wall.

Even if by some miracle we are able to get down to etching at angstrom-level (1Å = 0.1nm), silicon atoms are 2Å wide and it is quite literally impossible to etch anything smaller than an atom.
Posted on Reply
#34
Ravenlord
AssimilatorOnce we hit the wall of how precisely we can build tools to physically etch silicon, we are no longer able to increase transistor density. And we are currently almost at that wall.

Even if by some miracle we are able to get down to etching at angstrom-level (1Å = 0.1nm), silicon atoms are 2Å wide and it is quite literally impossible to etch anything smaller than an atom.
This answer has nothing about question which was asked. You are talking about known facts and it's not related to question.
Posted on Reply
#35
AnotherReader
AssimilatorOnce we hit the wall of how precisely we can build tools to physically etch silicon, we are no longer able to increase transistor density. And we are currently almost at that wall.

Even if by some miracle we are able to get down to etching at angstrom-level (1Å = 0.1nm), silicon atoms are 2Å wide and it is quite literally impossible to etch anything smaller than an atom.
A process with features that small would be dominated by quantum effects, and won't yield a working classic ASIC as we know it. So you're right; the wall will come before that, but we still have some time to go. As the following table from RealWorldTech shows, process names are far from the feature sizes that they used to correspond to.


Table 1 – Intel process node comparison
Posted on Reply
#36
ARF
AnotherReaderwe still have some time to go
I think they are beyond already since the price increases began to follow the performance uplift which is unsustainable and will lead to widespread semiconductors companies bankruptcies.
Posted on Reply
#37
Bwaze
Nah, it's completely normal, world has tons of millionaires that are gamers, not to mention people that earn money with productivity on their gaming cards...

2020, RTX 3080 - $700
2022, RTX 4080 - $1200 <- WE ARE HERE
2024, RTX 5080 - $2040
2026, RTX 6080 - $3468
2028, RTX 7080 - $5896
2030, RTX 8080 - $10022
2032, RTX 9080 - $17038
2034, GTX 1080 - $28965

Posted on Reply
#38
R-T-B
BwazeNah, it's completely normal, world has tons of millionaires that are gamers, not to mention people that earn money with productivity on their gaming cards...

2020, RTX 3080 - $700
2022, RTX 4080 - $1200 <- WE ARE HERE
2024, RTX 5080 - $2040
2026, RTX 6080 - $3468
2028, RTX 7080 - $5896
2030, RTX 8080 - $10022
2032, RTX 9080 - $17038
2034, GTX 1080 - $28965

Sir this is not Quadroville.
Posted on Reply
#39
AnotherReader
ARFI think they are beyond already since the price increases began to follow the performance uplift which is unsustainable and will lead to widespread semiconductors companies bankruptcies.
Companies will stay on older nodes or only migrate to new nodes when they are mature. You can see this in AMD's transition to N5: they followed Apple by 2 years whereas for previous nodes, they were among the first customers.
Posted on Reply
#40
ARF
Define mature. We don't need high yields but cheap, affordable processes. We need back-to-normal pricing:

599 for top
499 for cut-down
399 for mid-high-end
249 for mid-range
119 for low-end and entry

You don't get these prices with 20,000$ N4 or N5 wafers on super mature, high-yield process.
Posted on Reply
#41
Kyan
sam_86314Makes me wonder what will be used in the future. GaN?
www.electronicsforu.com/news/new-semiconductor-material-for-computer-chips
I think have read about that material some month ago but there is other candidate too. It's like car actually. We are going to the limit of what we were used to use for making chip/power car and now we have to find alternative ressources or a new way on how to continue with the old material.
Posted on Reply
#42
N3utro
All big CEOs of techs companies are singing together the "it will cost more" song thinking it will ready us to pay more for the same performance/price ratio. But what will really happen is when people dont buy it because it's not worth it, they'll lower their prices.
Posted on Reply
Add your own comment
Nov 21st, 2024 09:55 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts