HP's The Machine

[Thor.]

Hey has anyone heard of the machine???, HP seems quite dedicated to send AMD and Intel a message that they are not alone.. Will this be the big step in compute, or computers in general; They have all the inner workings to be a direct competitor to all manufacturers. Is this the change we been looking for.

 

I stumbled across this on the web, and did some research on the subject, the mystery thickens. This caught my attention because there is many breakthrough tech that seem to stall after announcement. The thing different with this computer is they are trying to mass produce it in a huge scale.

 

 

Will it take off?? photonic computing?? the concept of a supercomputer in a cell phone intrigues me, but also scares me :blink:. Thats the kind of power output they are describing.

 

random article found on the net.

 

http://www.veooz.com/news/6HGnnae.html

HP's wiki

http://en.wikipedia.org/wiki/Hewlett-Packard

 

Could this company bring a new age in computers, and leave the rest in the dust. Others will have to adapt pretty quick to keep up.

Edited June 18, 2014 by Thor.

[FMod]

I don't have any plans for buying one currently.

 

 

Could this company bring a new age in computers, and leave the rest in the dust.

HP? No. If you decipher the article, it just talks about how HP will finally make more than a sample of working memristors; samples are like 10 year old.

But whoever it is, don't worry.

 

All the advances - be it linear or breakthrough - have done for the last 70 years is keep upholding Moore's Law (which isn't really a law or really Moore's). No reason to expect anything else from any tech.

 

Big users (military and space) evaluate technologies in terms of readiness level, which is a quantitative assessment on a scale of 1 to 9 of how far away a given tech is from the market. And after it hits 9, there's a period of development and introduction.

[Thor.]

Sad but true

[Vagrant0]

Maybe I failed something in physics class... But how exactly would a binary switch work within a photon based model? Electrons can be stored and discharged while photons tend to be more elusive. Using light just for the larger circuitry in a computer would probably be easier and use less energy, sure, but would also require the more complex components to be more bulky as they would need to be able to transition from electrical impulses to rapid bursts of light, then back again just as quickly as those bursts are happening.

 

Though, in fairness, fiber optics was just becoming a plausible thing when most of my physics books were written, so my information on the subject may be incomplete.

[FMod]

It's just laying buzzwords on thick. What they mean by "photonic" is fiber optics interconnects. All that's new is further minuaturization of optical transceivers.

 

On another note, I'm proud to have in my possession a graded aggregate reinforced cementitous composite hexahedral structural module

a concrete brick

 

[Vagrant0]

It's just laying buzzwords on thick. What they mean by "photonic" is fiber optics interconnects.

Not exactly. It's a similar concept, but would probably be a different physical arrangement. The main idea is to reduce power needs by having fewer areas where data is being transfered by means of electrical signals. This would not be just changing connections between motherboard and hard drive to use fiber, but rather making the complex circuitry of the motherboard, storage drives, and other larger components work off light.

 

The problem though is that unlike electricity which is essentially creating a flow of electrons, with photons you cannot really split paths, but have to interrupt it and create additional pulses at every step. Which would probably make for more bulky components or just force a new way of doing everything to be more linear. Again, unless I'm missing something.

[FMod]

Not exactly. It's a similar concept, but would probably be a different physical arrangement. The main idea is to reduce power needs by having fewer areas where data is being transfered by means of electrical signals. This would not be just changing connections between motherboard and hard drive to use fiber, but rather making the complex circuitry of the motherboard, storage drives, and other larger components work off light.

Aware. And that's still just it - fiber optic interconnects miniaturized. There's no talk yet about replacing the transistor, it would be major on-chip buses, not smaller scale gets speculative.

 

The energy required for electrical<->optical conversion so far is still considerable and generally greater than that lost in electrical transmission.

 

Saying it's no longer an electronic computer isn't exactly honest. You could call it a photonic computer once data state changes - at least some of the state changes - occur while in optical form. This isn't it; what comes in is what comes out, even at its most optimistic processing would remain electronic.

 

 

The problem though is that unlike electricity which is essentially creating a flow of electrons, with photons you cannot really split paths, but have to interrupt it and create additional pulses at every step. Which would probably make for more bulky components or just force a new way of doing everything to be more linear. Again, unless I'm missing something.

"Passive optical network". Splitters can be used fairly reliably.

But there's rarely any splits at a logical level in modern electronic interconnects anyway. Actually, as clock rates shoot past the feasibility of multi-point buses (we're not there yet), signal splitting could even end up an advantage for optical transmission. Or it could prove just as undesirable in miniature optical buses as it is in electronic ones.