Schneier on Security

Clive Robinson • January 25, 2024 10:49 AM

@ Bruce,

Do you know about the connection between running Quantum Computers the making nuclear bombs and Security Theatre?

Basically the Qbits are in a liquid gas Helium-3 which is lighter than air and even though helium is one of the most abundant elements in the universe,

“It’s all out there way above our gravity well and of a density to low to measure conventionally.”

Remember the child hood fun of a ballon? Well few realise that it was the start of an interstellar journy for the atoms of gas inside.

They ballon bursts or leaks the atoms out, they rise up to the top of our atmosphere where “space weather” scours them free and accelerates then out of Earth’s influence into the space between the planets in our solar system.

The solar irradience of our sun pushes most of them ever outwards into interstellar space and thus outwards to who knows where.

Romantic as that might sound to a young childs dreams and wonders, it’s actually a bit of a nightmare.

Those atoms are a very finite and mostly a non renuable resource and we are going through them at an alarming rate. As I’ve indicated before they are now considered a critical resource and not as once regarded an anoying byproduct of the natural gas drilling operations.

https://www.science.org/doi/10.1126/science.326_778

Note the connection to the DHS and “Security Theatre”.

The other source is a byproduct of making nuclear bombs and in one case nuclear power,

https://laurentisenergy.com/story/laurentiss-production-of-helium-3-supports-high-tech-research-security-and-health-care/

So Quantum Computing may not happen because,

“The price of gas is to high and it’s become ‘unobtanium’.”

Clive Robinson • January 25, 2024 2:42 PM

@

Re : Untill heat part us all.

From the aricle on Carver Mead’s comments,

“In [his] view, one of the thorniest challenges for the chip industry is power dissipation. For decades now, the focus has been on faster and faster chips, but the heat issue can’t be ignored.”

If you look back through this blog you will see I’ve likewise been worried about “heat death” of chips.

The funny side if you are amused by people succeeding on target shooting their feet is in a number of cases it’s a problem that need not have happened.

It’s often portrayed by journalists as a transistor count, transistor density or clock speed or similar issue. Actually they are secondary effects to the real issue that’s not as easy to explain to journalists and hence them to others.

The problem is how long a transistor is in a linear or resistive portion of it’s charecteristic because that is when the heating effect due to “I squared R” causes issues. One way to reduce the losses is to examine how the transistor moves through that linear region. When off resistance is very high but the current is just about zero so the power is likewise close to zero. When the transistor is hard on it’s resistance is very low but the design is such that the voltage is very low so the V2/R power is very small. The design is often shown as a totempole where on transistor is on and the other off. Thus both I and V are minimised in either the on or the off conditions. Howrver switching the totempole from one state to another not just takes it through the linear stage, for speed there must be considerable movment of charge, not just through the totempole transistors but any storage element connected to them. Thus the likes of input and output capacitance make a difference by prolonging the transition time as well as increasing the current or voltage. Whilst there is little that can be done about the capacitance, resistance and current it’s important to remember that the supply voltage can be reduced. At one time logic chips had 24v supplies that came down to 5v and now as low as 1.8 volts below which life starts getting problematical for the likes of a totempole arrangement. But consider that 5 squared is 25 and 1.8 squared is 3.24 just bringing the supply voltage down brings the heat down to 3.24/25 or a little under 13% of what it was all other things being equal. Which unfortunately they are not, because like charge heat can be stored. There are videos around of people holding in their bare hand one end of iron wire the other end goes into a gas burner. The end in the flame gets hot enough to glow bright orange, yet the other end of the wire can still be held. The wire is taken out of the gas burner but still continues to glow for a few seconds. Usually though the heat finally gets to the end of the wire being held and the person has to let go. The fact the other end may still be glowing tells you about the storage effect, the fact that the wire could be held in a bare hand tells you about the slow speed of heat transmission in a nominal conductor. These storage and transmission issues are both worse in semiconductors…

Without going into more details there is less and less that can be done in a general way to chips and heat. However you can still get it down in the individual transistors by not switching them as often.

One of the failing with all of Intel’s “Go Faster Strip” circuits is that they have to run at the processor internal clock speed even though they are mostly not doing anything particularly usefull. As the old joke about the country gent leaning on the gate and after being asked for directions by newlyweds says on reflection “I wouldn’t start from here if I were you”… The fact is if you were to design a CPU Intel’s IAx86 architectute is one you most definately would not consider unless you want to double up as the element in a toaster (just one reason why ARM has such advantages).

But Heat Death is just a side effect of unwanted heat from computation, although not talked about Quantum Computers do generate quite a bit of heat so keeping the Qbits down to near absolute zero is actually quite an engineering issue and one of the reasons the chandeliers are designed the way they are. Not just a design with technical beauty but actually beautiful to look at as an object in it’s own right. It’s why the IBM office in London just down the street from Waterloo had one in it’s window, as a display last time I was down that way (if it’s still in the window I’d recomend a look in if you are passing).

Clive Robinson • January 31, 2024 4:37 PM

@ piglet, ALL,

Re : Encryption is not the only jewel.

“[T]he only problem that a workable quantum computer would actually be able to “solve” (which isn’t easily solvable with conventional computing) is breaking encryption.”

Not totally true but there are not that many exceptions to make “Quantum Computing”(QC) anything close to main stream.

The other problem QC has that is not much talked about is that it is actually very slow. In effect “snail on Mogadon (Nitrazepam)” slow. Which is why other “hunting” or “annealing” algorithms on modern hardware can compeate (with the movment to FPGA hardware outperform at a lower price point is possible).

In “theory” encryption is a “special case” set of algorithms where you can not “hunt in” via successively closer approximations (actually not entirely true as we are finding in more modern times).

What is “big and scary” is how asymetric crypto algorithms “have changed the world”. If you came up with an inexpesive and easy to run fasy algorithm that killed asymetric crypto, “Internet Commerce” and much of the US and other Western Nations economies would be “gone by the end of the week” or very shortly there after as would all online banking etc, and ammusingly so would all crypto-coins and their infrastructure.

Third world countries would not notice, nor would most second world. After all you don’t need asymetric crypto to herd herbivors or cultivate crops, or build buildings, roads or other infrastructure. Or anything that was “normal” prior to the mid 1990’s…

Would it hurt to kill the “big tech bros cash cows?” probably not, though they might scream “Snot Fair, Snot Fair” like a six year old in the school playground…