Support NeoGAF

[FightyF]

http://money.cnn.com/2005/02/01/technology/hp_transistor.reut/index.htm?cnn=yes

Quite interesting! Of course, I'm not getting giddy until I see some actual devices use this technology. But it's the kind of thing that could revolutionize the World as we know it.

 

[demon]

They can create a technology that changes the world of computers forever, but they can't create a fucking printer that doesn't jam on you one day and uncontrollably print out random shit the next.

 

[Artificial Meat]

Hopefully, in ten years, they have this stuff in everything we would use then. The current technology used for VLSI's and stuff is aging fast, and I really don't want to see the silicon industry going the way of the steel industry.

 

[Scrow]

what're the actual advantages this provides over exisiting transistors?

 

[Jonnyram]

what're the actual advantages this provides over exisiting transistors?

They don't "leak" the state. That's the quantum problem, apparently.

 

[Scrow]

They don't "leak" the state. That's the quantum problem, apparently.

bear with me.... what's the advantage in not leaking the state?

 

[Jonnyram]

Leaking the state, as in losing memory, so they have to be refreshed, which loses time. This doesn't require that.

 

[fart]

i haven't read the paper, but HP is hardly the only one working on this stuff. regardless, this quantum level stuff is much farther away than on-chip optical, nanoscale, etc.

 

[marsomega]

Leaking the state, as in losing memory, so they have to be refreshed, which loses time. This doesn't require that.

No...


As far as memory having to be "refreshed", that deals with capacitor based cells. DRAM cells have to be refreshed because the energy is stored in a capacitor (which represents the bit) discharges. The capacitor is like a "leaky" bucket holding energy that goes empty over time, in other words it leaks energy. Thus Memory (DRAM) has to be "refreshed" to regenerate the bit stored in the capacitor.

This talks about replacing the transistor not the capacitors in DRAM cells which are the cause for the leak since Capacitors discharge over time. SRAM which comprises of "static latches" as cells don't have to be refreshed.


This new tech does everything a transistor can however the edge this has over the transistor is that it can shrink down to a few nanometers and remain operable. Least, this is HP's claim as well as relatively simple and cheap (but how cheap?) to produce in mass quantity.

 

[Scrow]

well now i have no fucking idea what's going on

once more... what're the actual advantages this new technology provides over exisiting transistors?

 

[marsomega]

well now i have no fucking idea what's going on

once more... what're the actual advantages this new technology provides over exisiting transistors?

http://www.hp.com/hpinfo/newsroom/press/2005/050201a.html

QSR works on nanoscale electronic devices that will first supplement, and someday perhaps replace, silicon technology, which is expected to reach its physical limits in about a decade. In addition to exploring the fundamental scientific principles of computing at the molecular level, QSR is also looking at architectural issues and determining how such tiny devices - thousands of which could fit across the diameter of a human hair - could be fabricated economically and in mass quantities.

This basically says they can be made very tiny and produced cheaper in mass quantities. It is implied that it will be cheaper then transistors but not that cheap.

The experimentally demonstrated latch consists of a single wire acting as a signal line, crossed by two control lines with an electrically switchable molecular-scale junction where they intersect. By applying a sequence of voltage impulses to the control lines and using switches oriented in opposite polarities, the latch can perform the NOT operation, which, along with AND and OR, is one of three basic operations that make up the primary logic of a circuit and are essential for general computing. In addition, it can restore a logic level in a circuit to its ideal voltage value, which allows a designer to chain many simple gates together to perform computations.

Standard semiconductor circuits require three-terminal transistors to perform the NOT operation and restore signals. However, it is generally believed that transistors will not be able to shrink down to the size of a few nanometers and remain operable.

Basically saying it can do what a transistor can however QSR's can do it too much cheaper and much smaller. Here is where IMO it is implied that they are cheaper then transistors but not that cheap.

DRAM cells comprise of transitor/capacitor combos and because of the nature of the capacitor (discharges over time), which stores the bit, must be refreshed. SRAM cell on the other hand doesn't have to be refreshed because it comprises of transistors solely. Why don't we just use the faster SRAM? Because all those transistors are expensive, the more the you have, the higher the cost. We are talking 6 transistors per cell in SRAM versus 1 transitor and 1 capacitor in a DRAM cell. Of course, with the size of ram, 6 transistors per cell in 64 MB of ram will escolate the price tremendously. Here you see why would anyone deal with refreshing memory versus not having to worry about refreshing it at all.

But this talks about a combination of QSR's and capacitors. If they are that cheap, why not just replace the whole cell with a bunch of QSR's?

"We have previously demonstrated that we could make a working memory with molecular-scale junctions and logic devices that could perform simple logic operations such as AND and OR," Stewart said. "With the crossbar latch, we now have the final component theoretically needed for performing the multiple processing steps required for useful computing at the nanoscale."

This just says that QSR's will bring us closer to much smaller computing devices. They make this claim because apparently transistors can't scale as small and be operable as QSR's.

Of course this is all according to HP's claim.

 

[Scrow]

cheers

 

[marsomega]

cheers

Hopefully I didn't confuse you. I wasnt done with my post (unorganized). I meant to click on the preview button.

Should be fine now.

 

[Forsete]

So.. Is it kind of like this?

http://science.box.sk/newsread.php?newsid=4775

 

[GG-Duo]

I'm happy that people are working hard to keep Moore's Law intact.

wait, processors will still come from the same waffer technology right?

 

[sprsk]

When is this due out?