Share anything you find interesting broadly related to science and engineering that others might not know about. (Computer science, mathematics, physics, chemistry, biology, material science, geology, metrology, meteorology …)
Write the name of the topic on the top of the comment and at least one link below with a short description. Credit to @metamachine@FaunCB and a few others who gave me the idea with this and this thread. Feel free to add interesting links, videos and such.
The von Neumann-Landauer principle fundamentally limits the possible density of processors and logic gates because of the fundamental entropy change in non-reversible bit operations. Reversible logic does not need power on a fundamental level.
Please only post new interesting topics, research fields, concepts, … here. You can discuss and chat about these topics in the discussion thread (the intention is to keep this thread easily browsable for people who are only interested in the topics):
Here is another interesting topic. Something that I have linked here before:
Ultimate physical limits to computation
This paper discusses the extreme limits ( from a physics standpoint ) to how fast and how dense computers can become. (There are a couple of assumptions made, this does not apply in the same way to reversible logic that I mentioned in the op.)
Spoiler (funny quote):
“Nonetheless, although we do not know the exact number of bits that can be registered by a one kilogram computer confined to a volume of a liter, we do know the exact number of bits that can be registered by a one kilogram computer that has been compressed to form a black hole. … The amount of information that can be stored by the 1 kilogram computer in the black-hole limit is 3.826x10^16 bits. Note that this number is independent of the physics of the standard model, and relies only on the physics of quantum Fields on curved spacetime. The black-hole computer can perform 2.7129x10^50 operations per second, the same as the 1 liter computer.”
@Cobra92fs I have not seen that book before, thanks for the link!
To increase the critical mass, here are two further topics. I would really like to see some posts from the computer science guys in this forum or even better, interesting stuff all you everyday programmers stumbled upon.
A Fully-Integrated Flexible Photonic Platform for
Chip-to-Chip Optical Interconnects
This is an older paper, however there are some amazingly streight forward ideas in there and a lot of interesting keywords to google.
My personal recommendation, beeing a Astrophysics nerd:
The Laser Interferometer Gravitational-wave Observatory
surface level talk about IncludeOS (service orinted unikernel):
I know very little about operating systems and the technology around it, but I found some of the points made very intriguing. I’d like to know more from those people who know more about the pros and cons of Unikernels over here.
Post any interesting sciency stuff? Count me in. Just published a few weeks ago, out of CIT.
“Synthetic protein-level circuits could enable engineering of powerful new cellular behaviors.”
The paper is pretty dense (as are pretty much all papers from Science) and there’s 30 pages of supplemental material, but this is an idea I’ve had kicking around for a long time. There’s some good and some less good in the article though. Basically, they were able to program cell death based on semisynthetic logic circuits. One cool thing is that they used multiple logic gates, with each dependant on the result of the prior, which as far as I can tell is a first.
This is a big deal because it could potentially lead to protein complexes or circuitry that can make intelligent, autonomous decisions about drug delivery, or adopt conformations or fluorescence intelligently to relay information to scientists and doctors about the state of a cell. For instance, we can kill cancer cells fairly easily. The problem is killing only the cancer cells. There’s often not one clear biological indicator that something is a cancer cell - it’s usually determined by a series of information. We could theoretically design proteins that could examine multiple aspects of a cell, do some basic math, and if the cell scores high enough, kill it. That would be an incredible change in cancer therapy.
The downside to this is that it’s semisynthetic, like most of the novel protein design today, it isn’t truly novel. We know a lot about how proteins fold but still don’t have a clear model that works with a high enough success rate to think about designing proteins from scratch. Until we get that, the most we can do is adapt existing proteins and make mutations to those, which has a whole host of limitations.
Ultimately this is a computing problem. Once we crack the folding code (most likely to be done by brute force since it’s tough to get a clear picture of what’s happening on an atomic level when proteins are folding) it’ll be possible to design de novo proteins that are super-capable and super-efficient for our purposes.
The Antikythera Mechanism
Most of you will probably know about this amazing piece of ancient Greek machinery. I am following this video series for a while now and I have learned a lot in the process. Also production quality is through the roof on this one:
also, here is a series on
Ancient Tool technology
It does not mean that this is exactly how machine work was done but it shows some of the viable options.
And if you didn’t know about the Anthikythera mechanism I would suggest to start here:
Electrons are more cooperative (or at least we understand how to guide their movement better right now).
Leakage of electrons is a big issue in transistors and many decades of refinement has gone into address containment of and efficiency of movement of (lowering resistance) electrons in semiconductors.
Then there is manufacture… you’d need a scalable/cheap per unit process to lay down complex switching networks that scaled to billions of gates… I saw a fair amount of work 20 years ago go into “photonic switching networks”, but they were not trying to handle a lot of “logic”, but rather forestall the conversion to electrons as long as possible in a router to allow light to bend/redirect/shift mechanically to accomplish “most” of the routing logic.
Now try to scale that to registers, ALUs, caches, etc… materials have to change their game…
“The telescope at this point has long been operated by scientists working remotely, with only a small support staff onsite. In 1998, the physicists that manned the Parkes Radio Telescope began to notice random interference in the normal signals from space, interference that seemed to be—possibly– relayed from deep space. Interference described as fleeting bursts of radio signals. Though not a predictable regular occurance, these anomalies persisted. Finally, convinced they had made an actual discovery of a new form of radio wave, they named their discovery after a mythical chimera half stag and half bird: the Peryton. It was high up at the time of scientific mysteries.”
“And then—finally, they with a little bit of earthbound research—found exactly where the mysterious perytons were coming from.
And like every good scientist does, they wrote their findings up for inclusion in a scientific journal.”
“Perytons are millisecond-duration transients of terrestrial origin, whose frequency-swept emission mimics the dispersion of an astrophysical pulse that has propagated through tenuous cold plasma. In fact, their similarity to FRB 010724 had previously cast a shadow over the interpretation of fast radio bursts, which otherwise appear to be of extragalactic origin. Until now, the physical origin of the dispersion-mimicking perytons had remained a mystery. We have identified strong out-of-band emission at 2.3(2.5 GHz) associated with several peryton events. Subsequent tests revealed that a peryton can be generated at 1.4 GHz when a microwave oven door is opened prematurely and the telescope is at an appropriate relative angle. Radio emission escaping from microwave ovens during the magnetron shut-down phase neatly explain all of the observed properties of the peryton signals.”
It was the microwaves at the Parkes radio telescope facility. Every time one of the few staff members there grabbed their hot pocket before the timer went off.
The peryton was actually, all these 17 years, was actually a microwave."
ACF is an adhesive electrical interconnect that electrically connects opposing surfaces but does not conduct electricity parallel to said surfaces. This technology is commonly used in LCD manufacturing to connect electronic components to contacts on the display. To me this seems incredibly useful for hobby electronics projects to connect very small scale pads to each other.
I just now learned about this, I had now idea this existed and it is such a simple concept. Credit to the youtube channel Strange Parts for mentioning and explaining this simple but very useful technology that seems to be widely used in the electronics manufacturing industry.
Interposers traditionally have been various substrates that include metal traces for additional signal routing space for CPUs, FPGAs or other highly integrated logic circuits. HBM2 and GPU dies for example are connected via a silicon interposer. Active interposers, meaning silicon interposers with logic on the die, can improve signal routing capability and add other functionality to save space on the compute dies.