A team of AMD (Advanced Micro Devices) has filed a patent for teleportation. According to a report from Tom’s Hardware, the patent does not include human teleportation, but include the teleportation of qubits, the quantum computational equivalent of bits in classical computing.
The company’s recent study of quantum teleportation techniques resulted in the new patent, named “Forward-looking teleportation for reliable calculations in a quantum processor with several SIMDs.” It aims to enhance the efficiency of quantum calculations in part by reducing the number of qubits required for each calculation by focusing on this quantum event.
Although Google claimed to have achieved quantum superiority in 2019 and China claims to have the world’s most powerful quantum computer, quantum computing is still a long way to go before becoming a common, scalable calculating method. The fact that quantum states are exceedingly unstable which is one of the most major obstacles to quantum computing. Scalability is particularly difficult because the instability of a system only increases as more qubits are added to a system.
Several figures and diagrams are included in the patent, explaining a quantum architecture that utilizes quantum processing regions, where individual qubits wait in line to process data. The method intends to advance the stability of quantum computing by utilizing qubit teleportation to reduce the number of qubits required for complex computations.
The method allows workloads to be process in a “out-of-order” execution method by teleporting qubits across processing regions. This means that inactive qubits can be executed independently without a prior step being fully completed. This would essentially eliminate the need for qubits to be processed in a sequential order, allowing for independent and concurrent processing.
While we might not be able to teleport from one place to another in the 1960s sci-fi fashion, AMD’s latest patent promises a new quantum computing architecture that is far more stable as well as efficient, taking us one step closer to seeing the quantum computing revolution truly high.
As you might expect, some chip resources (in this example, qubits) are inactive until they’re called upon to perform the next calculation step. Out-of-order execution, on the other hand, examines a particular workload to determine which parts of it rely on prior results and which do not, and then execute each step of the instruction that does not require a previous result, resulting in improving performance through higher parallelism.
AMD’s patent also includes an embedded far-sighted processor in the architecture that analyses the input workload, predicts which steps can be addressed in parallel (and which cannot), and distributes the workload appropriately across qubits utilizing a quantum teleport technique to bring them to the required quantum processing, SIMD-based region.