IMB have revealed that they are working on a 7nm chip in conjunction with GlobalFoundries, Samsung and SUNY(State University of New York). The development are set to push Intel in the Process war seeing as this is one of the first times recently that Intel have not been well ahead of they competitors.
The 7nm Chip employs a type of silicone called silicon-germanium (SiGe), this type of silicon allows the transistors to be sub 10nm and still allow of electron mobility, this was previously a problem because regular silicon would not allow enough current through to make an effect. They are still using the standard FinFET transistor which is widely used in the commercial industry at the moment. They are now using UV lasers to etch the channels in the silicon, they are using an Argin Flouride laser which has a wave length of 193nm.
The chip is said to have improved performance by 50% and also brought down power consumption by 50%. Although the technology is still in its infancy it is set to look very exciting, the chip is set to released for commercial use in 2017/2018 although the use of EUV lasers to etch the chips is not yet practical. Also with Intel running into problems with 10nm production at the moment we may see some delays.
This has the potential to spice things up the processor market for the next few years and may even spur Intel to get of the arses even more and charge for the metaphorical hills of graphine. IBM may now take the place of AMD as the competition to Intel.
I loved IBM's Power PC series they smoked intel/amd back in the day. Seems like IBM is going full corporate support and leaving consumer all together this will probably get used in their server equipment and not in stuff we will ever use.
"The 7nm Chip employs a new type of silicone called silicon-germanium (SiGe)"
I don't think this is new SiGE is over 20 years old just the size is new.
I would like to see ibm re-enter the consumer cpu market with a socket cpu with as much punch as late 90's early 2000's relative to its competitors. It will never happen though.
It has been used in chunky diodes for a long time now, but as above it is new to the size it is being used in now.
Or more specifically silicone and germanium have both been used for diodes, I am not sure on the specifics of if they have been used together like this.
wiki says 1989 on SiGe but i agree this must be a new process. im interested and looking into it. I just watched a doc on silicon chips and it was very interesting. They touched on this tech but it was in the theoretical stages.
How about no. IBM is a development company now, not a manufacturer. Notice in the article it says they are licensing the design only to GlabalFoundries. This basically means that after this round of R&D they won't make more chips for themselves and will instead be collecting the royalties on the design from GloFo.
Of course this leaves Intel out of the secret circle, unless they make a deal with GloFo... AMD on the other hand is going to get a nice good look at what this is all about, and we may actually see them jump down to 7nm within the next 3 years. (Assuming their jump to 14nm is fruitful.)
They always throw around these small numbers. I think this video gives a really good idea of exactly what kind of scale were talking about. (its also a good explanation of transistors)
AMD is going to be manufacturing these according to yahoo.
"The company will license its technology to a variety of manufacturers, which will then mass-produce the chips with a variety of companies, including Broadcom, Qualcomm and AMD, to name a few."
After watching this vid it's hard to imagine them going much farther with the current manufacturing processes, the good bin parts at 7nm vs the rejects has to be a high enough number to even make it worth manufacturing at that die size which is difficult to wrap your head around, I'm really surprised that quantum mechanics hasn't reared it's ugly head at much bigger die sizes like 22nm, to build on the 7nm die and be successful is really quite a engineering feat in my little mind...lol
From what little I know about quantum mechanics, we have a long way to go before the the die sizes reach those levels. Considering DNA is 2.5nm wide and that is made up of proteins, which are chains of atoms. We'd have to get much much smaller to the subatomic to get to that funky world.
I imagine that DNA is a slightly different process from transistors. Maybe the switch of materials is one of the reasons they got 7nm, i would think that it will get harder as they continue to try and reduce the size.
I didn't know much about transistors, i didn't realize that were essentially have switches that are only dozens of atoms across.
Yeah...the article I just finished reading basically says that with the new manufacturing process of 7nm ( Gate All Around (GAA) FETs and vertical tunneling FETs (TFETs) make 5nm and 3nm a possibility.
For some reason I feel like this 3nm isn't as far off as we'd normally think. As with all technological advances it's on an exponential curve up. Landed on the moon 66 years after the Wright brothers first flew. The advances in just the last 5 years alone is pretty remarkable. I'm excited for all this new fun stuff. Hopefully I can upload my consciousness into the "cloud" soon... This whole eating and taking care of a body is getting really old....
LOL.... put yourself in my shoes I'm 59!, and while I wasn't around for the first flight I do remember the moon landing when I was in grade school.
But your right success in 7nm will build into success in 5nm then 3nm, but I'm more interested on the consumer side, sure it will offer more power at lower wattage but the CPU's we have now are pretty over powered for daily use for the average consumer.