You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
IBM says they have produced the first commercially viable photonic chip on 90nm. I wonder how far others are behind or this is even accurate for I know Intel among others is working on the same thing. I feel this is the first step to a photonic computer and even this step should vastly increase the power of systems by allowing separate processors to work together at a vastly increased speed. I have been told just the increase in communications speed in data centers and systems will be a game changer. Will this be the ultimate game changer? Any opinions or information on this area would be appreciated.
Arthur do you have a link you can share on this. There have been many electronic / photonic ICs made and I'd like to understand what makes this unique. Is this using the Global Foundries (e.g. Ex IBM monolithic process)? What is the application etc etc?
In short answer to your question, yes Integrated Photonics stands to play a larger role as we move past the 100 Gigabit per second space and into 200 Gps and then 400 Gbps arena.
Its already the predominant work horse for long haul communications but as speeds and distances in the mega data centers increases you will start to see integrated photonics move closer and closer to the processor clusters on the boards.
It it sounds like IBM may be stepping into this realm now integrating both digital logic with the photonics needed for hyper-speed I/O.
Mitch, I just have the link I posted, but have listened to lectures at UCB in which Intel was involved in pure photonic computing. Look at my forum post on Opatlysys.
Ok, that makes more sense. The May 2015 article is also good and gives a clear picture of where photonics is going in the data centers. Global Foundries announced last year that they were going to provide foundry services for this type of integrated photonics and as you know, they leverage IBM's technologies so I'm hoping to see more of these types of ICs coming from the fabless community this year. Photonics activity in North America has picked up considerably, especially with the advent of the AIM Institute that is now entering it's second year of operations. Their version 1.0 PDK is due to be released this month and they recently announced an advanced test and packaging facility in conjunction with ON Semi in Rochester, NY for integrated photonics. Should be an interesting year.
Dan, the reason I gave that link is because the more current link was from the current MIT Technology Review which is behind a pay wall. I have also listened to lectures on this at UCB online in the last few months. gI make my money reading between the lines and many companies are not talking. Careful homework and conjecture will give one a fairly good idea. I make my living doing this, so this is not idle chatter.
Absolutely Arthur. You have rightfully noted an important trend. I was just thrown off by the timeline of the IBM claim as I knew that work had gone on previously and rightfully so as shown by the article you referenced. I'm currently doing a lot of work in this space at the moment and I can tell you that momentum has picked up significantly in the last year or so. Keep up the good and inciteful comments!
Best regards,
Mitch Heins
Mitch, if you could keep me posted. I have assembled a very high end team in this area to advise me on investments. I'm now investing in second tier companies that supply key technology, for I don't know who the final winner of the actual optical computer or communications equipment will be. The high speed data transfer alone is valuable. This technology will also be critical for lab on a chip and sensor technology.
