with frickin lasers!
The path to exascale supercomputing computing is going to require the development of a number of technologies, and one of them is the shift from copper to optical signaling between chips in a complex of computing and memory. And IBM thinks it is going to have an edge in this effort now that it has cooked up photonics components …
with frickin lasers!
Note it's not putting laser light MODEMs (well that's effectively what this functionality is) on a chip.
It's integrating it into the fabrication process flow itself.
Building a laser more or less in a normal Silicon CMOS process, rather than "cheating" by dropping a little GaAs chip on top.
It's good to see a big company create this kind of innovation. Optical pathways actually integrated into the chip!
Seems that IBM should be valued higher than it is. I have read about a lot of impressive technologies they are working at and for the most part accomplishing, but they always seem to be standing in the background. I know they don't generate the dollars a company like Facebook does, but they seem more valuable than Facebook none the less. That might not make sense, but I can't help feel that way. Yes, it's a troll post, just thought I'd share thought.
In 2011 IBM's revenue was $106.91B, facebook's was $3.71B.
IBM doesn't really do stuff that people outside of IT see anymore, it doesn't that they are smaller than consumer tech that people do see.
Yes, I do work for them...
Look at Market cap the true value of a company (at lease how investors view it)
Sadly HP is $28B. I would hate to have to be part of the organization that Mark Hurd set on the course for failure.
Then the board destroyed the company.
You heard it hear first. Carl Icahn will buy a substantial portion of the company and the first thing he will do is fire Ray Lane. Ray is the reason the board is dysfunctional and hires a CEO none of them ever interviewed. Meg is doing her best, but the jury is still out as Ray is still calling the shots.
And adjusted for inflation, IBM's market cap has been higher (in the '70s) than Apple's has ever been.
IBM is simply a blue-chip stock. The irrationally exuberant don't buy it, so it doesn't go to wacky P/E levels. But the market cap is huge and the company continues to be profitable, in no small part because they have a big and very productive R&D organization.
(And had you bought IBM at around $50 in the mid-1990s - about $75, adjusted for inflation - and held it through the subsequent "returns to investors" until today, with the share price just a bit under $200, you'd be sitting pretty.)
If optical memory interrconnects (sorry, from the West Country) will effectively do away with memory latency and speed bottlenecks, what kind of overall improvements will be seen with existing processor clock speeds and architectures? Will the need for 'on die' cache SRAM etc. be negated (freeing die real estate for more cores etc.)?
I went back to re-read the article to make sure, and nowhere is it stated that optical interconnects "do away" with latency and/or bottlenecks.
What the article states is that photons go faster in fibre than electrons do in a wire, therefor optical interconnects are going to minimize latency, not erase them.
And unless quantum entanglement gets even crazier than it is now, you'll always have one absolute physical latency : the distance between RAM and CPU. It cannot go faster than the speed of light in a vacuum, after all.
To all the upstarts that called for Big Blue to be killed as the "dinosaur" of the IT world, this is research, not finding out how many people tweet or facebook their choice of breakfast.
Staying out of the headlines until you have something of value for the world instead of releaseing a "concept" then a few buggy attempts before quietly giving up is the way forward.
This is why Xerox, IBM, CRAY still survive, they dont follow consumer trends.
Funny that: the only person I have ever heard use the term 'dinosaur' in regards to tech was a fellow from Microsoft bashing Big Blue. That was in 2009... as of now IBM's market cap is only lagging Microsoft's by $11.4B (219.75B vs. 231.17B respective). IIRC that might have been the same day (or at least the same week) Microsoft announced their first ever major layoff.
Look at acquisitions too... purchases of fundamental/core technology on the one hand: (IBM) Texas Memory, Netezza, and a whole shopping list of industry-specific cloud based solutions just to name a few... vs. (MSFT) Danger, Skype, and Yammer.
That's not to say Big Blue is perfect or anything, but they are still around and a solid long-term investment for a reason. The same can't be said for a lot of other tech companies.
If you want to hear more people saying dinosaur, just look at the comments here about any article which mentions the BCS. A whole bunch of know-nothing teens and twentysomethings banging on about something of which they have no knowledge. Apparently formalising things like change control make you a dinosaur...
Oh wow, I'm a dinosaur and didn't even know it!