Quantum computing for the defense industry

33

Rainier Quantum Processor for D-Wave One Computer

Scientific organizations from different countries are working on the creation of the so-called. quantum computers. Devices of a special architecture should show increased performance and simplify the solution of a number of tasks. It is quite natural that the military and defense industries are already interested in such technologies.

Implementation process


The first to appear on the market were quantum computers from the Canadian company D-Wave Systems. Since 2007, it has introduced a range of processors based on a different number of qubits with different capabilities. In this case, we are not talking about a full-fledged universal computer, but about a specialized system for solving specific problems. However, in this case, superiority over the systems of "classical" architecture was shown.



In 2011, a 128-qubit D-Wave One computer was introduced, capable of only discrete optimization. Soon there was a contract worth more than $ 10 million with Lockheed Martin for the supply of one machine and its subsequent maintenance. In its communications, the client organization indicated that the computer will be used to solve the most complex problems and problems in the field of software.

In 2013, Lockheed-Martin ordered a newer D-Wave Two computer with a 512-qubit processor. The next computer, Type Two, was sold to a group of organizations led by NASA for a joint project. The third contract with Lockheed Martin was signed in 2015 and provided for the supply of the D-Wave 2X product, which has 1152 qubits. Other customers include NASA and Los Alamos National Laboratory. At the beginning of 2017, sales of the D-Wave 2000Q computer (2048 qubits) started - they again became interested in NASA and related organizations. The Advantage system (5640 qubits) is entering the market this year.

In June, it became known that the USC-Lockheed Martin quantum computing center based on the USC Viterbi technical school will receive an Advantage computer in the near future. It is expected that the receipt of a more powerful machine will expand the capabilities of the center to conduct research and create practical systems. In addition, the new computer will be included in the Leap quantum cloud complex.


D-Wave 2000Q computer

Despite the criticism and limited capabilities, quantum computers from D-Wave Systems have been the subject of several contracts and have been used in various organizations for almost 10 years. The main customer for such equipment was Lockheed Martin, one of the largest organizations in the defense industry, operating in all major areas. Also, scientific and research organizations, incl. employed in the applied field.

The future from DARPA


In March of this year, the DARPA agency launched its quantum computer project. Within the framework of the ONISQ (Optimization with Noisy Intermediate-Scale Quantum) program, it is planned to work out general issues of creating new computers, and then create ready-made samples. Seven "teams" are involved in the work, in which scientific and design organizations participate.

The first stage of ONISQ will last a year and a half; within its framework, it is required to develop techniques and algorithms for solving combinatorial optimization problems. Then the second stage will begin, the purpose of which will be to improve the created products and programs. The results of the program can be applied both in the military and in the civilian sphere.

DARPA takes into account that creating a universal quantum computer is an extremely difficult task, and therefore, for now, they set more modest goals. In particular, it is allowed to create simulators of a quantum system based on a "regular" computer or to develop a hybrid architecture with a limited number of qubits.

Russian perspective


In other countries, incl. in Russia, the development of quantum computing is still lagging behind. Simulators and prototypes with a small number of qubits are available and are in use, but commercial sales and mass adoption are still a long way off. However, the necessary measures are being taken, and the desired results will appear in the near future.


So, in 2018, the Russian Foundation for Advanced Study launched the "Optical Systems of Quantum Computing" project. As part of this work in 2018-2021. it was planned to create demonstrators of computers with 50 qubits based on neutral atoms and integrated optical circuits. The main executor of the project is Moscow State University, and a number of other organizations are also involved in the work.

The work has not yet been completed, but there are already plans to introduce new technologies. Quantum computing is one of the promising areas of the Digital Economy state program. Fundamentally new computers with increased performance will be used in various areas of industry and economics. In the future, it is proposed to create means of quantum encryption with increased stability.

The current project is of interest to various government and commercial organizations. Thus, the state corporation "Rosatom" and some large state and commercial organizations are interested in obtaining quantum computers. The possibility of their introduction in the defense industry is being discussed - but specific enterprises have not yet been named. Most likely, such issues will be resolved in the future, after the appearance of ready-made prototypes.

Applied tasks


The main advantage of quantum computers over traditional systems is their increased performance. Thanks to this, a quantum machine can be used for faster calculations or for performing specific work in which the use of other means is impractical.

Lockheed Martin has been involved with quantum computing for several years. Detailed information about the progress of such works, their real tasks and the successes achieved were not disclosed. At the same time, general data on the goals and prospects of the direction have been published, which allows us to draw various conclusions.


The company's official resources mention the use of computers from D-Wave Systems as a software verification tool. When writing software, certain errors are possible, the search and correction of which takes up a lot of the developer's time and resources. A high-speed quantum computer can test a program in the shortest possible time and identify any existing problems. In some cases, quantum computing is able to cope with tasks that are almost impossible for systems with other architectures.

Code review can be used in various fields. As an example, the development of software for aviation technology. The quantum computer will speed up the process of testing and improving programs, and a ready and safe aircraft will be released for testing. High-speed computers are also suitable for calculations in the space industry. Within a few milliseconds, D-Wave Two is able to calculate many trajectories of the ship and select the optimal one.

Performance issues


Lockheed Martin has only quantum systems with limited capabilities - computers from D-Wave solve only a narrow range of problems. In the future, the emergence of universal systems with wide areas of application is expected, and this will make it possible to more fully use the possible speed.

Fast processing of large amounts of data or performing complex calculations is required in various areas of the defense industry. The introduction of quantum computers will simplify software development, speed up the design of various structures, and also reduce the number of field tests required. High performance can be useful when creating systems with artificial intelligence for various purposes - this area is also of interest to military and defense enterprises.

In general, quantum computing, as well as general purpose or specialized computers, are of great interest and can find applications in various fields. The introduction of such systems, as expected, began with the defense sector and, most likely, it will be the leader in the development of newer technologies and designs. It is expected that the fundamentally new means of computing will seriously affect the industry and the military, but how quickly such results will appear and what changes will lead to it is not yet clear.
33 comments
Information
Dear reader, to leave comments on the publication, you must sign in.
  1. +11
    13 August 2020 06: 29
    Who knows...
    Sometime in the 90s, my dorm neighbor and I read in the newspaper about the creation of "Pentium 2". And the neighbor is a computer scientist and he alone had 486th.
    He ridiculed the article and authoritatively stated that it was impossible to jump above 486. My son is now wondering how we could count on anything at all on the first PCs!
    1. +4
      13 August 2020 08: 37
      When the first 30 * 30 "hard drives" were released, many did not understand how to fill such a huge amount of memory, 60 megabytes. How people developed counted planes, missiles, abacus and slide rule. in some cases, computerization tends to complicate rather than facilitate labor. Previously, accounting and settlements were carried out on accounts, and everything was fine. Now on computers, program 1-C, have you seen the instructions for the user? These are 3 thick books, and updates are constant. The computer counts in milliseconds, before these people will enter data into the computer for a couple of days, and then for a week they will read what the system has issued. The person becomes a very weak link .. I hope the machines will not say goodbye.
      1. +6
        13 August 2020 10: 15
        How people developed counted planes, missiles, abacus and slide rule.

        With numbers and graphs, I was not surprised. But about seven years ago "Matkad" (the diploma with the youngest was "composed"), when in symbolic the integral decides for you or does the Laplace transform, for which neither brains nor life is enough - that's when I said "Wow! This creature really thinks"
        1. +5
          13 August 2020 10: 54
          Quote: dauria
          "Wow! This creature really thinks."

          Everything flows, everything develops. So far, the problem with quantum computers is software and error correction. I tried to understand the principle and rested (well, stereotypical thinking), I realized that qubits are both "0" and "1" in other words, calculates one qubit at once! 3 responses when adding bits. those. answers "00" "01" and "11". But the question of applying stereotypical thinking does not allow to comprehend. But what I realized is that this logic is the best fit for machine learning, and this is already the door to artificial intelligence. In all areas where it is required to calculate almost infinite options for solutions, they are irreplaceable. With enough power, it will calculate all possible options for anything with an acceptable speed. There will be a breakthrough in chemistry (especially organic). In general, the prospects are more than serious.
          PS: Simulators are already available to those who are interested. For acquaintance https://habr.com/ru/post/418505/
          1. +3
            13 August 2020 11: 24
            But the question of applying stereotypical thinking does not allow to comprehend.

            I also said to myself - "nafig, interesting, but sorry for the time." And I also remembered the analog computers on the op-amp. The flight of the Kh-23 guided missile was perfectly modeled by a simple scheme. Something similar - you won't see intermediate results, but get the overall result. Very fast and accurate enough for the eyes.
            Rough, or maybe wrong - how to weigh a trifle. You can count the pennies, or you can use the scales. recourse
            1. +2
              13 August 2020 11: 44
              Quote: dauria
              and get the overall result. Very fast and accurate enough for the eyes.

              That's what is interesting (well, this is so to the topic). If, for example, take any computer (for example) game, then on quantum computing it will give all possible endings, you just have to choose the option you need. Well, in real life, chemists (organic in particular) today can only work with theory, modeling reactions at the atomic level using computer technology is carried out only with hydrogen ... And how many problems are there in atomic, quantum physics? Modeling a rocket flight in an acceptable range can be calculated on a calculator (and it will go in a column), but they launched rockets into space and managed to dock them. But to simulate the result of combat use in real time at least 2 aircraft, subject to the build-up of forces and assets on both sides ... request It seems to me that with the development of quantum computing (hardware will be created normal, I have no doubt), the military race will be in this area.
              Yours!hi
            2. BAI
              +1
              13 August 2020 12: 45
              analog computers

              At the dawn of the development of computers, it was believed that the future belonged to analog machines. Even in the first half of the 80s, they were still studied in institutes.
              1. +1
                13 August 2020 18: 14
                And even later, new models appeared.


                The model appeared in the late 80s.
              2. 0
                13 August 2020 22: 59
                Quote: BAI
                At the dawn of the development of computers, it was believed that the future belonged to analog machines. Even in the first half of the 80s, they were still studied in institutes.

                It must be because, along the way, there was also an opinion about the constant and continuous intellectual development of the human brain, because to work with an analog computer, it is required to represent the entire process of calculation, and not just do the stupid monkey work of entering data and reading the result. Well, who knew that they would be so wrong ...
      2. +2
        13 August 2020 10: 46
        Now on computers, program 1-C, have you seen the instructions for the user? These are 3 thick books, and updates are constant. The computer counts in milliseconds, before these people will enter data for a couple of days
        And before you enter data in 1c, you have to write with your pens another 1c on another platform, because these 3 thick books cannot take into account all the nuances of owls. accounting department.
      3. AUL
        -1
        13 August 2020 18: 45
        Quote: Free Wind
        When the first 30 * 30 "hard drives" were released, many did not understand how to fill such a huge amount of memory, 60 megabytes

        Generally 40 MB. And there was enough for all programs, data, and even a bunch of toys like "Princesses"!
    2. +3
      13 August 2020 09: 05
      Once upon a time ... there was such a computer - ZX "Spectrum". On tape. As a child, I played toys on it.
      We thought it wouldn't be better.

      I will say - the slogan. If Russia does not lose sight of its own in the development of quantum computers, this can nullify our chronic lag in the development of computers.
      1. +3
        18 August 2020 08: 26
        And there is. Why invest billions in processors with a conventional 7 nnm, when you can invest billions in quantum systems?
  2. +5
    13 August 2020 09: 25
    The amount of data that has to be stored and processed is growing exponentially. And there is no end to it.
    In Israel, two large firms are building their own data centers at once: Microsoft and Lockheed. And he is going to build Amazon.
    1. +1
      13 August 2020 11: 36
      You are somewhat wrong! The general information system does not have new incoming data. Therefore, internal flows change, namely, the density of these processes is transformed.
    2. 0
      13 August 2020 11: 44
      Data centers require high energy availability. Therefore, critical levels will come very quickly, if you do not try to solve problems in the technology of transferring a unit of information. The code as an electronic magnetic impulse as part of a natural series of numbers can be rotated so to speak, and not to look for a new place in space for it, and therefore the code is written into a numerical structure, which is the data array that is then determined by the capacity of this data, and not by speed and energy consumption by other transmission parameters.
    3. +1
      14 August 2020 06: 45
      ... firms are building their data centers:

      This data center is just like a factory in our plant, large rectangular buildings, electrical substations, galleries with cables, but unlike ore processing, the data center grinds the figure and their profits are a thousand times greater.
      1. 0
        14 August 2020 10: 17
        There is not even an electrical substation. Special power plants are being built for them,
        refrigeration units with cooling towers, concrete reinforcements against air strikes.
        Like military installations. Date is expensive and can be critical to
        defense of the country.
  3. 0
    13 August 2020 11: 28
    Superbig data and their combinatorial optimization, and even the construction of variable combinations of data at all levels of fractality cannot be carried out on inaccurate definitions of quantum computations. Well, on the computational principles of modern mathematics, and even more so. Therefore, it is necessary to change the set of approaches to solving these problems. Starting from using the function of constant value of a number, which means using only integer values ​​and building algorithms specifically for the relationship of a number, and not abstract processuality. And binding the work of arithmetic operations to the work of a multipolar processor. Therefore, it is very easy to tie a number to just an electronic magnetic pulse. It is impossible to implement this on binary logic of both simple computers and quantum computers. But the main thing from which you need to get away is from tying the performance of the machine to the transmission rate of a unit of information. Therefore, the performance of the machine increases unimaginably many times if not to increase and fight for the data transfer rate, but to work with arrays of local and algorithmically related numeric codes. That is, to transfer already optimized data, then the energy aspects of data transfer are also solved. It is worth repeating once again that the solutions are not in mechanical technologies, but in ideological methods of working with big data. AI is primarily an analysis technology, not working with statistical data.
  4. +1
    13 August 2020 11: 32
    delivery of a D-Wave 2X product with 1152 qubits. His other customers were NASA and Los Alamos National Laboratory.

    Mathematical modeling of processes, such as nuclear explosions, requires efficient machines.
    1. +1
      13 August 2020 11: 39
      Not always ... As an example, you yourself, that is, your brain ... laughing
      Well, there is also such a topic - Quantum annealing. Read at your leisure.
      1. 0
        13 August 2020 12: 50
        Quote: Shahno
        theme-Quantum annealing. Read at your leisure


        Already. Thank.
  5. +2
    13 August 2020 11: 49
    53 versus 5000. Whose qubits are thicker?

    https://mcs.mail.ru/blog/5000-kubit-kvantovoe-prevochodstvo

    From a marketing and PR point of view, the announcement of a 5000-qubit system, and even with such an odious name, looks like a major event in the quantum computing industry. Especially against the backdrop of the latest news from IBM about the opening of the IBM Quantum Computing Center in New York based on five 20-qubit and one 53-qubit system.

    "Five thousand qubits" sounds much more powerful than the recent landmark event of Google's "quantum supremacy" using a 53-qubit superconducting system.

    Despite numerous thousands of qubits, D-Wave systems are significantly inferior to "full-fledged" quantum computers with tens of qubits from other manufacturers. The problem is that the limitation of the capabilities of quantum D-Wave systems operating on the principle of quantum annealing


    So "50 full-fledged" domestic qubits is not so far from Google's 53-qubit system and it is not a fact that D-Wave Advantage is worse.

    but ... the big difference is that about 60 applications have already been developed for D-Wave and they work.
    1. +2
      13 August 2020 11: 56
      As the teacher said ... kills the qubit of strife.
      We have a colloquium tomorrow on optical degradation in ising machine ... Generalization of these models is generally much higher than that of quantum ... yes, and I'm not talking about keeping a stable state in quantum qubits.
      Ps. That is, to what I am all this. Each qubit for its own class of problems. It can't be better or worse. This is still a piece. What is not good.
  6. 0
    13 August 2020 11: 52
    Within the framework of the global development of mankind, there are obviously three stages of evolution. This is a mankind without technologies for the conscious use of analysis technologies. This stage is a stage of technologies of low potential analysis of physical processes based on computational mathematics. And a new stage is the analysis of capacious and highly potential physical processes. And of course, this is, first of all, an analysis technology based on a function of a constant value of a number, because it is the fpsc that allows you to optimize ultra-high data and solve this issue absolutely accurately in every solution and solve this issue with optimal energy security.
  7. BAI
    0
    13 August 2020 12: 49
    Everything related to computers (both hardware and programming) is developing so rapidly that certain technologies become obsolete before they find industrial application. It is possible that quantum computing will be replaced by something else before large-scale production begins.
  8. +1
    13 August 2020 13: 03
    Once Griboyedov said - grief from the mind to people ... The current slide rule does not know, as well as the purpose of the protractor with the abacus ... They turned off the electricity and the battery in the beech dried up, no, they will not determine the square of the circle ... . area of ​​a circle on paper with a pencil ......
  9. +1
    13 August 2020 16: 37
    Quote: Free Wind
    Previously, accounting and settlements were carried out on accounts, and everything was fine. Now on computers, program 1-C, have you seen the instructions for the user? These are 3 thick books

    Excuse me, but accounting in 1C is far from the most basic function for which it is used. Along with the growth of technical capabilities, the tasks also expand. Increasing the efficiency, depth and variability of the analysis of large amounts of data is the main task of the systems for managing the financial and economic activities of an enterprise. So this comparison is incorrect. And by the way, 1C is far from the best system here. The fact that it is crammed where it does not hit, well, this is the old trouble of progress - it is not the best technical / engineering solution that captures the market, but a commercially successful one - remember, for example, the capture of the PC market by IBM or software by Microsoft ... here is 1C from the same operas.
  10. -1
    14 August 2020 00: 17
    For the curious:
    Lockheed Martin Computing Center video
    https://youtu.be/Fls523cBD7E
  11. 0
    15 August 2020 10: 43
    It is very interesting that competent scientists do not see the prospects for the development of quantum technologies. At least in the direction which is now determined. Everyone is waiting for something and does not know what. And it’s quite funny to watch how, presenting non-standard ideas, but solutions, everyone gets stupid. The masses are suppressed by the binary logic of reasoning. Even the elementary game technologies of spatial chess in a cube are shocking to people. Nobody can break through the level of underdevelopment of human civilization. Only she herself knows the algorithms of timeliness and dynamics of her evolution.
  12. 0
    22 August 2020 20: 14
    As far as I know, experts in quantum computing themselves admit that so far they can give practical results only in very narrow encryption problems (albeit very important ones).

    At the same time, a number of cybernetics claim that in the near future, more exhaust can be given by hardware neuroprocessors - matrices of artificial neurons.
    As modern processors are very well suited for imperative programming, so neuroprocessors are ideal for running adaptive, genetic algorithms on them,
    machine learning problems - when an algorithm is "born" in the process of solving a problem.
    It is neuroprocessors that are capable of making a revolution in the coming years, because they are very well suited for "inaccurate calculations" - these are just all behavioral tasks: making decisions in conditions of insufficient and / or conflicting input data (and this is all command and control command systems), control of all unmanned vehicles; optimization (for example, search for optimal solutions in design, TRIZ); extraction of noise-like signals from noise: this is pattern recognition, optical observations (the so-called tasks like "reading a newspaper through a layer of milk"), radar, acoustics and sonar, applications in flaw detection, medicine (search for tumors), geology and geophysics, etc.

    The development of such hardware It is imperative that you keep up with the widespread introduction of "language-oriented programming" methods. In the case of neuro-computing, the traditional concept of "source code" loses its meaning - behavior is generated during the operation of the computing system. Here the so-called. "open specifications", which describe not HOW the program should work, but WHAT it should do (see "Why is Shalyto right about open specifications? - YouTube", https://www.youtube.com/watch?v=qWKAqLazoRI) - It is necessary to give specialists in the implementation of such systems in subject areas ("programmers-trainers") convenient tools for formulating problems in a given subject area for input into neuroprocessor systems (as now even a child can "teach" a CPU to solve problems in Basic, Python).
  13. 0
    22 September 2020 12: 06
    Here, this direction did not exist in Soviet times.
    And where is the RF in this sector?
    Do you need to explain?
    But I still want to scold the Soviet government.
    Is not it?
    Mediocrity and thieves.
  14. 0
    1 October 2020 19: 27
    Why do we need defense!
    Russia has long been occupied. We are a resource appendage of the West. Don't make me laugh