The limitations of contemporary supercomputers, as well as the ramifications for academics and institutions worldwide, are drawing attention in the scientific community. For example, researchers may use current technology to perform more complicated simulations, such as those that focus on chemistry and the reactive properties of each element. However, when the intricacy of these interactions increases, they become far more challenging for current supercomputers to manage. Due to the limited processing capability of these devices, finishing these sorts of computations is nearly impossible, which is forcing scientists to choose between speed and precision while doing these studies.
To provide some context for the breadth of these experiments, let's start with the example of modeling a hydrogen atom. With just one proton and only one electron in hydrogen, a researcher could easily do the chemistry by hand or depend on a computer to complete the calculations. However, depending on the number of atoms and whether or not the electrons are entangled, this procedure becomes more difficult. To write out every conceivable result for an element such as thulium, which contains a staggering 69 electrons that are all twisted together, would take upwards of 20 trillion years. Obviously, this is an inordinate amount of time, and standard techniques must be abandoned.
Quantum computers, however, open the door to a whole new world of possibilities.
source: Quantum Computing: Current Progress and Future Directions | EDUCAUSE
Eduard Babulak I think part of the problem is the lack of understanding of the nature of the electron and the nature of molecular bonds:
Preprint The Hydrogen Bond (June 2022)
It might be the case that once we have a realistic model to work with then the computer models become simpler. Even the heavier elements with many electrons can be considered with the electrons occupying shells so that only the outer shell electrons are involved in molecular bonding.
I wouldn’t recommend waiting for quantum computers. They lack accuracy. For example ask yourself what is the probability that a quantum computer will give the answer 7 when adding 3 and 4.? The probability is not 1 which rules out their use for many applications.
Richard
Gaussian can handle most small molecules, quantum chemists do take
up large chunks of computer time. Always the same problem, you can handle
either small systems, or very large ones. The middle ones are hard.
Supercomputers dont last a long time, they are replaced fairly fast by even
faster ones.
The quantum ones are still hypothetical, but could work in parallel computation. They would need taxing physical conditions as with low temperature, prone to mistakes.
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling