Latest developments in technology

I’ve always been interested in the progress of computers, coming from a science background, and the progress in computing technology has been amazing over the years, but some things are just incredible.

Anyway, I read an article in computer magazine about the latest developments in IBM’s new 2nm chip and it’s absolutely astounding.

As well as delivering dramatic improvements in power consumption and speed, apparently with this new process size they can now fit 333 million transistors onto 1 square mm of silicon, which seems unimaginable to me! :astonished:

As a child I used to take apart and repair transistor radios and other simplistic electronic devices but it’s truly mind blowing just how far things have come.

But can anyone also manufacture 2nm atm?

I know TSMC is on 5nm and will start 3nm next year, Samsung is on 7nm and plan for 3nm this year.

1 Like

They ve made a chip to prove it’s achievable and estimate it’ll be four years to upscale to commercial manufacture. It’s very early days though so that timeframe is just a current best guess.

I have a new Ipad with Apple’s M1 chip, I have no idea what nm that is? The little bugger is really fast.

EDIT: I googled it
The M1 chip uses a 5- nanometer ( nm ) production process,

Yep, Apple reserved 80% of TMSCs 5nm manufacturing.

That´s not good for other other clients, I guess. Also not sure whether it´s a good design with DRAM included

1 Like

TBH I’m amazed that 2nm is possible as I thought they’d pretty much reached the limits of physical possibilities. Every time we hear about limits though they keep finding clever ways of getting around them. It does make you wonder though what the ultimate will eventually be…

Maybe it´s time to looking for new manufacturing processes. I think the actual processes will come to an end

1 Like

There are (again) rumors that Samsungs 5nm manufacturing have big problems with a yield rate <50%

1 Like

That’s always the difficulty. Translating the prototypes and scaling up the manufacturing to mass production.

I suppose they learn more and more with each generation but it’s in the nature of pushing limits that you’re always going beyond what came previously so there are always new difficulties, some of which will be easy to overcome but others may be extremely challenging requiring considerable time to resolve.

I think they are pretty much at the limits of how much more they can shrink the node size. The probable limiting factor being Silicon itself.

For CPU’s. I think that the ‘Large core’ ‘small core’ configuration is the way forward