A Powerful Co-Processor with Speed Like No Other
MemComputing dramatically reduces computation time and costs by delivering solutions orders of magnitude faster than any prior technology.*
MemComputing unifies the memory and processing components of a traditional computer to dramatically accelerate computation.
By translating a computer science problem into a physics problem, we revamped traditional computing to mimic the network in the human brain, producing a computational collective state. As a result, our multidirectional circuitry is able to scale in linear time and manipulate large data sets at lightning speed.
MemComputing solves optimization problems at never-before-seen speeds so you can tackle previously impossible problems.
The Science Under the Hood
Combinatorial Optimization Problems
MemComputing uses a proprietary feature called “Self-Organizing Logic Gates” (SOLGs) to rapidly search for optimization problem solutions. Unlike traditional combinatorial methods, SOLGs are assembled in active, dissipative “Self Organizing Logic Circuits” with an equilibria associated with either the optimal minima of the optimization problem or approximations to the optimum far beyond the heuristic limits of current methods.
MemComputing can also solve hard binary optimization problems natively by formulating them as “Boolean Maximum Satisfiability” (MaxSAT) problems or “Quadratic Unconstrained Binary Optimization” (QUBO) problems.
When implemented in hardware MemComputing operates at room temperature, without the extreme environmental requirements of other advanced computational technologies. And it’s primed to scale, since all components that create MemComputing hardware are readily available.
Available as Software
Since MemComputing uses standard electronic circuitry, we can reproduce the technology in software. Our first product, Falcon, is available on the cloud as a Software as a Service (SaaS) and through custom Software Development Kits (SDK).
”Improved Performance of Optimization Calculation by 4 or More Orders of Magnitude for a ‘simple’ computation”
MEMCOMPUTING BEATS FASTEST AVAILABLE SOLUTIONS FOR MAXSAT PROBLEM
Time increases exponentially as the number of variables increases, capping the total amount of variables and creating impossibly long computation times.
Time increases linearly as the number of variables increases, enabling efficient scaling of the number of variables while maintaining feasible computation times.
When every millisecond and every mile counts, MemComputing calculates optimal routes orders of magnitude faster than other systems.
We can help:
Dramatically reduce time of routing computation
Increase cost-effectiveness of production
Improve accuracy of solutions
Home Package Delivery
Chip Routing Design
DYNAMIC MOTION PLANNING
Beyond static routes, MemComputing provides the most efficient trajectories in environments with moving obstacles at near real time.