ARM CEO Predicts Processor the size of a blood cell.

RobertKaucher

ARM CTO predicts chips the size of blood cells - smartphones, processors, consumer electronics, Components, Arm Holdings - Hardware - Techworld

This is coming from a guy in the know. So it makes the vision we were discussing in this thread (http://www.techexams.net/forums/off-topic/71472-5-10-years-future.html) all the more possible.

Here are some similar items dealing with materials that will allow flexible LCD displays for tablets of the future.

http://www.techexams.net/forums/off-topic/71452-nokia-kinetic.html
http://www.nature.com/nnano/journal/vaop/ncurrent/full/nnano.2011.184.html
Nanotube Springs Stretch Skin-Like Sensor - YouTube

Everyone

A red blood cell is about 6 to 8 micrometers in diameter and 2 micrometers thick. 1 micrometer = 1000 nanometers. Chips built on a 28nm process are just now coming out, and we're projected to have chips built on an 11nm process by 2015.

CPUs are still at 32nm, it takes a 512mm squared size chip to pack in 2.6 Billion transistors (10 core Xeon). 512 square mm = 262,144,000,000 square micrometers. That's about 0.0099 transistors per square micrometer. The surface area of a red blood cell would be about 201 square micrometers, which is 1,304,199,005 times smaller than today's top of the line CPU on today's 32nm process. So a chip the size of a red blood cell would only be able to hold almost 2 transistors today. By 2015 we may be able to fit 5.7.

I'm not sure what's smaller than a nanometer, but until we get into semiconductor manufacturing processes that are smaller than a nanometer, I don't see chips the size of a red blood cell being very useful. Intel's first CPU, the 4004 had 2,300 transistors.


Oh and Flexible LCD isn't really the right term... it's FOLED, which is not really an LCD. Flexible Organic Light Emitting Diode. Like I said in the other thread, those are already possible, it's just a matter of time before the manufacturing process is tuned enough to bring the price down enough to get them into general use.

RobertKaucher

Everyone wrote: »

A red blood cell is about 6 to 8 micrometers in diameter and 2 micrometers thick. 1 micrometer = 1000 nanometers. Chips built on a 28nm process are just now coming out, and we're projected to have chips built on an 11nm process by 2015.

CPUs are still at 32nm, it takes a 512mm squared size chip to pack in 2.6 Billion transistors (10 core Xeon). 512 square mm = 262,144,000,000 square micrometers. That's about 0.0099 transistors per square micrometer. The surface area of a red blood cell would be about 201 square micrometers, which is 1,304,199,005 times smaller than today's top of the line CPU on today's 32nm process. So a chip the size of a red blood cell would only be able to hold almost 2 transistors today. By 2015 we may be able to fit 5.7.

I'm not sure what's smaller than a nanometer, but until we get into semiconductor manufacturing processes that are smaller than a nanometer, I don't see chips the size of a red blood cell being very useful. Intel's first CPU, the 4004 had 2,300 transistors.


Oh and Flexible LCD isn't really the right term... it's FOLED, which is not really an LCD. Flexible Organic Light Emitting Diode. Like I said in the other thread, those are already possible, it's just a matter of time before the manufacturing process is tuned enough to bring the price down enough to get them into general use.

I know. I just hate FOLED because it looks like it should be a propper acronym but it isn't. I also hate OLEDB and try as hard as I can to not say it. I'll change that to flexible LED, then to be more accurate.

But as far as processor size I think that there are techologies in the pipes right now that might make this possible within that timeframe. The main one I can think of is ferroelectrics; which could make the power requirements, and therefore heat output, of processors reduce by several orders of magnitude. I believe that once something like that comes into common use we will break the 4 GHZ mark and I also think miniturization could occur faster. But who knows.

Now that being said, when someone like Ray Kurzweil says something that X is going to happen I might roll my eyes and read on, but when the CEO of a company like ARM says something like this I pay a little more attention. But of course it could be marketing.

Everyone

I actually think that the "3D" process is what will make it happen faster. Intel developed a 22nm 3D transistor process earlier this year. Being able to pack transistors on top of each other in addition to just next to each other should be a pretty significant increase in processing power, while shrinking chip size.

The benefits of being able to build computing components that make use of 3 dimensions rather than 2 have been around for a few years now with Storage devices. It'll be cool to see how they work out with processors.

With 3D transistors, a chip the size of a red blood cell (roughly, lets say 8 micrometers cubed) in 2015 may be able to hold about as many transistors as 386 CPU did in 1985. Pretty impressive considering that was a 104 square mm chip.

RobertKaucher

Everyone wrote: »

I actually think that the "3D" process is what will make it happen faster. Intel developed a 22nm 3D transistor process earlier this year. Being able to pack transistors on top of each other in addition to just next to each other should be a pretty significant increase in processing power, while shrinking chip size.

The benefits of being able to build computing components that make use of 3 dimensions rather than 2 have been around for a few years now with Storage devices. It'll be cool to see how they work out with processors.

With 3D transistors, a chip the size of a red blood cell (roughly, lets say 8 micrometers cubed) in 2015 may be able to hold about as many transistors as 386 CPU did in 1985. Pretty impressive considering that was a 104 square mm chip.

Yea, the 3d architecture would help in this, but I think at that size we are still going to have issues with heat - but maybe not. For consumer electronics I really think heat and powers consumption are the two big things that need to be addressed.

Turgon

RobertKaucher wrote: »

ARM CTO predicts chips the size of blood cells - smartphones, processors, consumer electronics, Components, Arm Holdings - Hardware - Techworld

This is coming from a guy in the know. So it makes the vision we were discussing in this thread (http://www.techexams.net/forums/off-topic/71472-5-10-years-future.html) all the more possible.

Here are some similar items dealing with materials that will allow flexible LCD displays for tablets of the future.

http://www.techexams.net/forums/off-topic/71452-nokia-kinetic.html
http://www.nature.com/nnano/journal/vaop/ncurrent/full/nnano.2011.184.html
Nanotube Springs Stretch Skin-Like Sensor - YouTube

ARM are a British company based in Cambridge. They know what they are doing and have done for years..

ARM Holdings - Wikipedia, the free encyclopedia

mattlee09

RobertKaucher wrote: »

Yea, the 3d architecture would help in this, but I think at that size we are still going to have issues with heat - but maybe not. For consumer electronics I really think heat and powers consumption are the two big things that need to be addressed.

Just saying this off the cuff,

I wonder if, along with power consumption, we need to work on the batteries themselves too.

I guess in time this will all be realized. I mean, this industry is still in it's infancy, right?

RobertKaucher

mattlee09 wrote: »

Just saying this off the cuff,

I wonder if, along with power consumption, we need to work on the batteries themselves too.

I guess in time this will all be realized. I mean, this industry is still in it's infancy, right?

Yes, you are certainly correct in that. But one thing we can keep in mind with mobile devices as teh consumtion levels come down and battery efficiency goes up we will also see an increase in bandwidth allowing for the OS itself (for many mobile devices) to be run in the cloud. So your 25", wafer thin LED tablet needs almost NOTHING because it's not actually ruunning any OS other than what is required to boot and pick the cloud OS you want to load. You might have a VM for your entertainment, one for your work, etc. and all of these could be loaded in multiple "desktops" on your tablet. Sure these mobile processors do need to get more powerful and efficient - but not like desktops. It's just not required.

So you are really right. The batteries need to get thinner and more efficient as well.

Thin film rechargeable lithium battery - Wikipedia, the free encyclopedia

Everyone

The smaller the nm process for creating a chip is, the less power it consumes, even though the transistor count is greatly increased. We already have chips using only a fraction of a watt of energy. Lower wattage = less heat.

The thermodynamic challenges of 3D processors is probably pretty interesting. It seems like the transistors at the center would dissipate heat into the transistors around them, making cooling the entire processor evenly quite difficult.

ARM's own Cortex A5 processes are only 0.68mm squared, delivering 1GHz while using only 1v and consuming as little as 80mW of energy. That processor is 2 years old now, and powers a lot of mobile devices out there.

There's a lot more too it than that of course.

Everyone

Here you go, this is sort of inline with this... Researches at Cornell created transistors made out of cotton...

Cornell Chronicle: Transistors made from natural cotton fibers

Your t-shirt may also be your computer in the future. Dry-clean only!