# US 2021/0174237 A1 · Hybrid Quantum Computing Network

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0174237 A1
- Document ID: US20210174237A1
- Status: Published application
- Kind: A1
- Application number: 17/110,382
- Filed: 2020-12-03
- Published: 2021-06-10
- Earliest claimed: 2019-12-05
- Assignees: Mellanox Technologies, Ltd.
- Applicants: Mellanox Technologies, Ltd.
- Inventors: Elad Mentovich; Kyle Scheps; Juan Jose Vegas Olmos
- Web page: https://patentgrep.com/patent/US20210174237A1

## CPC

- G06N 10/80: Quantum programming, e.g. interfaces, languages or software-development kits for creating or handling programs capable of running on quantum computers; Platforms for simulating or accessing quantum computers, e.g. cloud-based quantum computing
- G06F 9/4881: {Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues}
- G06F 9/5044: {considering hardware capabilities}
- G06F 13/4022: {using switching circuits, e.g. switching matrix, connection or expansion network (G06F13/4009 takes precedence)}
- G06N 10/60: Quantum algorithms, e.g. based on quantum optimisation, quantum Fourier or Hadamard transforms
- G06F 2209/5017: Task decomposition
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A distributed computing network includes a quantum computation network and a processor. The quantum computation network includes one or more quantum processor units (QPUs) interconnected one with the other using quantum interconnects including each a quantum link and quantum network interface cards (QNICs), where each QPU is further connected to, using the QNIC, a quantum memory. The processor is configured to receive a quantum computation task, and, using a network interface card (NIC) (i) allocate the quantum computation task to the computation network, by activating any of the quantum interconnects between the QPUs according to the quantum computation task, and (ii) solve the quantum computation task using the quantum computation network.

## Claims

### Claim 1

A distributed computing network, the network comprising:
a quantum computation network comprising one or more quantum processor units (QPUs) interconnected one with the other using quantum interconnects comprising each a quantum link and quantum network interface cards (QNICs), where each QPU is further connected to, using the QNIC, a quantum memory; and
a processor, which is configured to:
receive a quantum computation task, and, using a network interface card (NIC), allocate the quantum computation task to the computation network, by activating any of the quantum interconnects between the QPUs according to the quantum computation task; and
solve the quantum computation task using the quantum computation network.

### Claim 2 (depends on 1)

The computing network according to claim 1, further comprising, at least for a given pair of the QPUs, (i) a time counter associated with the quantum interconnect that connects the given pair of the QPUs, and (ii) a classical interconnect connected in parallel with the quantum interconnect, wherein the processor is configured to time-synchronize the given pair of the QPUs with one another using the time counter and the classical interconnect.

### Claim 3 (depends on 1)

The computing network according to claim 1, wherein the quantum computation network has a regular polygon graph topology in which vertices of the topology are the QPUs and edges of the topology are the quantum interconnects.

### Claim 4 (depends on 1)

The computing network according to claim 1, wherein the processor is configured to activate any of the quantum interconnects between the QPUs according to the quantum computation task using an algorithm that maps the quantum computation task to a regular polygon graph.

### Claim 5 (depends on 1)

The computing network according to claim 1, wherein the processor is further configured to receive the solved quantum computation task and, using the solved quantum computation task, solve a full computation task comprising the quantum computation task.

### Claim 6 (depends on 1)

The computing network according to claim 1, wherein the quantum interconnects are configured to support at least one of Remote Direct Memory Access (RDMA) and InfiniBand Trade Association (IBTA) standards.

### Claim 7 (depends on 1)

The computing network according to claim 1, wherein quantum interconnects are configured to communicate qubits.

### Claim 8 (depends on 1)

The computing network according to claim 1, wherein the processor comprises one of a CPU and a GPU.

### Claim 9 (depends on 1)

The computing network according to claim 1, wherein the quantum computation task is separable.

### Claim 10 (depends on 1)

The computing network according to claim 1, wherein the quantum computation task is inseparable.

### Claim 11 (depends on 10)

The computing network according to claim 10, wherein the inseparable quantum computation task is formulated using a quantum-classical path integral (QCPI) model.

### Claim 12

A distributed computing method, comprising:
receiving a quantum computation task, and, using a network interface card (NIC), allocating the quantum computation task to the computation network, by activating any of the quantum interconnects between one or more quantum processor units (QPUs) according to the quantum computation task, wherein the one or more QPUs are interconnected one with the other using quantum interconnects comprising each a quantum link and quantum network interface cards (QNICs), and wherein each QPU is further connected to, using the QNIC, a quantum memory; and
solving the quantum computation task using the quantum computation network.

### Claim 13 (depends on 12)

The computing method according to claim 12, further comprising time-synchronizing at least a given pair of the QPUs with one another using a time counter and a classical interconnect, wherein (i) the time counter is associated with the quantum interconnect that connects the given pair of the QPUs, and (ii) the classical interconnect is connected in parallel with the quantum interconnect.

### Claim 14 (depends on 12)

The computing method according to claim 12, wherein the quantum computation network has a regular polygon graph topology in which vertices of the topology are the QPUs and edges of the topology are the quantum interconnects.

### Claim 15 (depends on 12)

The computing method according to claim 12, and comprising activating any of the quantum interconnects between the QPUs according to the quantum computation task using an algorithm that maps the quantum computation task to a regular polygon graph.

### Claim 16 (depends on 12)

The computing method according to claim 12, and comprising receiving the solved quantum computation task, and using the solved quantum computation task solving a full computation task comprising the quantum computation task.

### Claim 17 (depends on 12)

The computing method according to claim 12, wherein the quantum interconnects are configured to support at least one of Remote Direct Memory Access (RDMA) and InfiniBand Trade Association (IBTA) standards.

### Claim 18 (depends on 12)

The computing method according to claim 12, wherein quantum interconnects are configured to communicate qubits.

### Claim 19 (depends on 12)

The computing method according to claim 12, wherein the processor comprises one of a CPU and a GPU.

### Claim 20 (depends on 12)

The computing method according to claim 12, wherein the quantum computation task is separable.

### Claim 21 (depends on 12)

The computing method according to claim 12, wherein the quantum computation task is inseparable.

### Claim 22 (depends on 21)

The computing method according to claim 21, wherein the inseparable quantum computation task is formulated using a quantum-classical path integral (QCPI) model.
