antiterror13
Brigadier
Totally agreed, as it is a National Security issue context, the same reasons what the US govt use
Chinese semiconductor industry
- Thread starter Hendrik_2000
- Start date
- Status
Totally agreed, as it is a National Security issue context, the same reasons what the US govt use
That is the point. If it is going to take that long, 2-3 years, then it is not worth it.
If China has fabs inside China manufacturing 10nm or 7nm or 5nm, then China can go buy 28nm chips anywhere in the world.
China cannot achieve self sufficiency at 14nm and below for the time being.
28nm it can.
Look at how US plans to rip out old Huawei equipments.
China will "rip out "existing US semiconductor equipments and replace with domestic even if they are old.
Just an interesting fact about the benefits of new nodes.
The difference between the current fastest single thread CPU - 7nm AMD Ryzen 9 5900X and between the 2010 vintage 45nm Intel Core i7-880 is fivefold , means ten development works years, 41 times more transistor managed so little advancement .
Most likely best part of the difference coming from the new architecture, not from the manufacturing technology.
Just reminder , the speed difference between the 180nm and 90nm CPUs was close to 6, in five years time ,with magnitude less development fund.
If the speed difference should be 540* bigger between the 45nm and the 7nm if the scaling similar like between the 180nm and 90nm.
C'mon, I managed to show 2 times performance improvement on manufacturing lines, using existing machining centres in two years time ,without changing the product.
The difference between the current fastest single thread CPU - 7nm AMD Ryzen 9 5900X and between the 2010 vintage 45nm Intel Core i7-880 is fivefold , means ten development works years, 41 times more transistor managed so little advancement .
Most likely best part of the difference coming from the new architecture, not from the manufacturing technology.
Just reminder , the speed difference between the 180nm and 90nm CPUs was close to 6, in five years time ,with magnitude less development fund.
If the speed difference should be 540* bigger between the 45nm and the 7nm if the scaling similar like between the 180nm and 90nm.
C'mon, I managed to show 2 times performance improvement on manufacturing lines, using existing machining centres in two years time ,without changing the product.
I think a more fruitful way of looking at these goals is seeing if there are any gaps where China's tech industry is stalled or severely hampered. Even if true self-sufficiency is long-term, there have to be enough bridging strategies to get there: i.e., smuggling US equipment, getting alternative suppliers from abroad, buying off-the-shelf chips instead of self-developing them, maintaining stockpiles etc. It would be absolutely unacceptable if a Chinese tech company had to sit and wait for two years twiddling its thumbs.
That's what the contours of the conversation should be about, not hyper-focusing on narrow goals. Having said that, it's well accepted that SMEE will be supplying its 28nm ArF immersion lithography machine next year (it's also been reported that SMIC has at least one prototype for verification); what basis do you have for claiming that the 28nm goal will take 2-3 years? In that kind of timeframe a domestic EUV solution should be emerging.
Well, neither can the Americans.
That is why this whole fight that the Americans started with China over chips, is a little odd.
Intel is basically dead, Microsoft is moving to ARM long term for its servers and cloudWell, neither can the Americans.
That is why this whole fight that the Americans started with China, is a little odd.
I build a new PC with 4790k back in late 2014, and recently had upgraded to 9900k more than half a decade later.... in most games I've benchmarked I barely even get 5fps in improvement and its so marginal you cannot notice without a benchmark tool...
I remember back in the golden age of computing that I would need to upgrade every other year and then every upgrade was at least 2x in improvement (200%+)
In a few years even GPU will slow down... I bought the RTX3090 and the thing takes up all three slots of my machine... and looks like a large monster and to be honest only yielded like 10% improvement in frames in most modern AAA games compared to the much smaller and much cheaper RTX3080 that I upgraded from....
We have reached point of deminishing returns in all regards
Nvidia sees writing on the walls this is why they rebrand themselves AI company and purchased ARM
- Status