WO2023189886A1 - 情報処理装置 - Google Patents
情報処理装置 Download PDFInfo
- Publication number
- WO2023189886A1 WO2023189886A1 PCT/JP2023/011095 JP2023011095W WO2023189886A1 WO 2023189886 A1 WO2023189886 A1 WO 2023189886A1 JP 2023011095 W JP2023011095 W JP 2023011095W WO 2023189886 A1 WO2023189886 A1 WO 2023189886A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- quantum
- unit
- information processing
- arithmetic
- processing
- Prior art date
Links
- 230000010365 information processing Effects 0.000 title claims abstract description 93
- 238000012545 processing Methods 0.000 claims abstract description 122
- 238000012546 transfer Methods 0.000 claims abstract description 27
- 238000004364 calculation method Methods 0.000 claims description 76
- 238000012937 correction Methods 0.000 claims description 38
- 239000002096 quantum dot Substances 0.000 claims description 20
- 238000000034 method Methods 0.000 description 41
- 238000010586 diagram Methods 0.000 description 30
- 230000008569 process Effects 0.000 description 27
- 238000004891 communication Methods 0.000 description 10
- 230000007704 transition Effects 0.000 description 10
- 238000001514 detection method Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 4
- 239000003381 stabilizer Substances 0.000 description 4
- 238000003672 processing method Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000006866 deterioration Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000013508 migration Methods 0.000 description 2
- 230000005012 migration Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 150000002500 ions Chemical class 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 230000001151 other effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N10/00—Quantum computing, i.e. information processing based on quantum-mechanical phenomena
- G06N10/70—Quantum error correction, detection or prevention, e.g. surface codes or magic state distillation
Definitions
- the present disclosure relates to an information processing device.
- NISQ Noisy Intermediate Scale Quantum
- This NISQ has the problem that noise is generated in the calculation results and errors are accumulated. Therefore, a system has been proposed for leveling noise by combining NISQ and a classical computer (for example, see Patent Document 1).
- the present disclosure proposes an information processing device that combines NISQ and a quantum operation unit that performs operations while correcting errors.
- the information processing device of the present disclosure includes a first quantum operation unit, a second quantum operation unit, and a control section.
- the first quantum operation unit includes quantum bits and performs quantum operations.
- the second quantum operation unit includes quantum bits and performs quantum operations involving error correction.
- the control unit causes the first quantum calculation unit to perform calculation processing, and controls to transfer the calculation processing to the second quantum calculation unit based on the amount of error in the calculation processing of the first quantum calculation unit. I do.
- FIG. 1 is a diagram illustrating a schematic configuration example of an information processing device according to a first embodiment of the present disclosure.
- 1 is a diagram illustrating a configuration example of a conventional information processing device.
- FIG. 2 is a diagram illustrating an example of conventional information processing.
- FIG. 1 is a diagram illustrating an example configuration of an information processing device according to a first embodiment of the present disclosure.
- FIG. 3 is a diagram illustrating an example of a processing procedure of information processing according to the first embodiment of the present disclosure.
- FIG. 2 is a diagram illustrating an example of a circuit that transfers calculation results according to the first embodiment of the present disclosure. This is an example in which an error correction code is represented by a stabilizer code.
- FIG. 1 is a diagram illustrating a schematic configuration example of an information processing device according to a first embodiment of the present disclosure.
- 1 is a diagram illustrating a configuration example of a conventional information processing device.
- FIG. 2 is a diagram illustrating an example of conventional information processing.
- FIG. 7 is a diagram illustrating an example of a processing procedure of information processing according to a second embodiment of the present disclosure.
- FIG. 7 is a diagram illustrating an example of a circuit that transfers calculation results according to a second embodiment of the present disclosure.
- FIG. 7 is a diagram illustrating a schematic configuration example of an information processing device according to a third embodiment of the present disclosure.
- FIG. 7 is a diagram illustrating an example of a processing procedure of information processing according to a third embodiment of the present disclosure.
- FIG. 7 is a diagram illustrating a configuration example of an information processing device according to a modification of the embodiment of the present disclosure.
- NISQ (1. Background technology)
- FTQC Fault Tolerant Quantum Computer
- ECC error detection and correction code
- T-gate when performing non-Clifford operations, which are key to obtaining quantum superiority, an operation called a T-gate is required. In order for this T-gate to work, it is necessary to generate a special quantum state called a magic state. In order to perform quantum operations with ECC implemented, a large amount of this magical state is required. For example, it is estimated that more than 90% of FTQC's quantum bits are used to generate magical states.
- FIG. 1 is a diagram showing the characteristics of FTQC and NISQ.
- the advantage of FTQC is that by repeating error correction, the quantum state can be preserved for a long time while maintaining the quantum nature. Therefore, FTQC can perform calculations an arbitrary number of times. This allows for complex operations.
- NISQ has the advantage that the CNOT operation (Clifford operation), which involves relatively large noise, can be executed at low cost.
- FTQC FTQC becomes extremely expensive.
- a special gate T gate
- ECC processing must be performed at various points in the calculation, there are also disadvantages such as increased processing time and increased power consumption. This is because it is necessary to read out a plurality of quantum bits each time ECC processing is performed. This quantum bit read operation takes several times as long as the gate operation.
- NISQ does not require ECC processing, calculations can be performed at high speed. Furthermore, since one logical qubit is expressed by one physical qubit, NISQ can construct a large-scale system using about 1000 times as many qubits as FTQC.
- the drawbacks of NISQ are that the quantum nature disappears in a finite time, that it can only perform a finite number of operations due to noise generated during quantum operations, and that it is not good at operations that span multiple qubits.
- the present invention was devised for the purpose of using FTQC and NISQ to compensate for each other's shortcomings and achieve the best calculation performance.
- FIG. 2 is a diagram illustrating a schematic configuration example of an information processing device according to the first embodiment of the present disclosure.
- FIG. 1 is a block diagram showing a schematic configuration example of the information processing device 1. As shown in FIG.
- the information processing device 1 includes a control section 10, a first quantum operation unit 20, and a second quantum operation unit 30.
- the first quantum operation unit 20 is equipped with quantum bits and performs quantum operations.
- This first quantum calculation unit 20 is a calculation unit that does not perform error correction during quantum calculation, and corresponds to the above-mentioned NISQ.
- the second quantum operation unit 30 is equipped with quantum bits and performs quantum operations involving error correction. This second quantum operation unit 30 corresponds to the above-mentioned FTQC.
- the first quantum operation unit 20 and the second quantum operation unit 30 in the figure can also be provided with a communication path that provides interaction.
- the dotted arrow in the figure represents this communication path.
- the communication path corresponds to, for example, a quantum communication path via quantum entanglement.
- the control section 10 controls the first quantum operation unit 20 and the second quantum operation unit 30.
- the control unit 10 causes the first quantum calculation unit 20 to perform calculation processing, and also transfers the calculation processing to the second quantum calculation unit 30 based on the amount of error in the calculation processing of the first quantum calculation unit 20. Take control.
- the first quantum computing unit 20 and the second quantum computing unit 30 may be located in different areas within the same chip (die), in different chips, in different packages, or in different locations. It can be placed in a housing, another system, etc. Furthermore, focusing on the physical system to be implemented, superconductivity, ions, cold atoms, light, semiconductors, etc. can be applied to the first quantum operation unit 20 and the second quantum operation unit 30.
- FIG. 3 is a diagram showing an example of the configuration of a conventional information processing device.
- This figure is a block diagram showing a schematic configuration example of an information processing device 50 described as a comparative example.
- the information processing device 50 includes a storage 51 , a classical communication path 52 , an external interface 53 , a classical computer 54 , a logical/physical converter 55 , a quantum operation unit 56 , and a quantum communication path 57 .
- the classical computer 54 is a computer equipped with a CPU (Central Processing Unit), etc., and is a processing system that does not perform quantum operations.
- the external interface 53 is for communicating with the storage 51 and the classical communication channel 52.
- the storage 51 holds programs and data.
- the classical communication channel 52 is a communication channel such as a LAN.
- Logical/physical converter 55 is an interface between classical computer 54 and quantum computing unit 56.
- the quantum operation unit 56 performs quantum operations.
- the classical computer 54 performs processing while communicating with the quantum operation unit 56 via the logical/physical converter 55.
- FIG. 4 is a diagram illustrating an example of conventional information processing. This figure is a diagram illustrating interactions between layers in the information processing apparatus 50 of FIG. 3, focusing on information levels.
- a user 61 inputs commands using a relatively high-level programming language. This instruction is decomposed into low-level elements by a compiler 62 located in the classical computer 54. Various methods can be considered for decomposing this process, depending on various constraints, viewpoints of interest, and evaluation indicators.
- the decomposed processing is reduced to commands associated with physical operations. Based on this command, the control device 63 causes the quantum operation unit 56 to perform a physical operation 64.
- a superconducting quantum computer executes calculations by controlling quantum states called transmons through microwave irradiation.
- the mapping of which physical qubits to allocate logical qubits, the intensity, phase, and irradiation time of the microwaves to be irradiated are determined based on the characteristics and topology of the device used, the processing to be performed, the priority indicators, etc. Optimized.
- the information processing device 1 of the present disclosure unlike the information processing devices shown in FIGS. 3 and 4, repeatedly compiles and executes commands in the quantum operation unit.
- the control unit 10 in FIG. 2 distributes and executes the arithmetic processing to the first quantum arithmetic unit 20 and the second quantum arithmetic unit 30 in FIG. Control is performed to transfer the data to the quantum operation unit 30. Specifically, the control unit 10 compiles the arithmetic processing, generates a command, and causes the first quantum arithmetic unit 20 to execute the command. After that, the control unit 10 interrupts the processing of the first quantum calculation unit 20 depending on the situation of the calculation processing.
- the control unit 10 compiles the remaining arithmetic processing portion again, generates a command, and controls the transfer to the second quantum arithmetic unit 30.
- the information processing device 1 can perform arithmetic processing using a combination of the first quantum arithmetic unit 20 that does not perform error correction and the second quantum arithmetic unit 30 that performs error correction.
- the information processing device 1 of the present disclosure will be described in detail using FIG. 5.
- control unit 10 can also allocate the arithmetic processing in anticipation of the state of the arithmetic processing such as the amount of error in the first quantum arithmetic unit 20.
- the control unit 10 generates commands for each of the first quantum operation unit 20 and the second quantum operation unit 30 during compilation. After that, the control unit 10 causes the first quantum operation unit 20 and the second quantum operation unit 30 to execute the generated commands in order. At this time, the control unit 10 performs control to transfer the result of the arithmetic processing of the first quantum arithmetic unit 20 to the second quantum arithmetic unit 30 to execute the remaining arithmetic processing.
- FIG. 5 is a diagram illustrating a configuration example of an information processing device according to the first embodiment of the present disclosure.
- This figure like FIG. 1, is a block diagram showing an example of the configuration of the information processing device 1.
- the control unit 10 includes a classical computer 11 and a logical/physical converter 12.
- the classical computer 11 in the figure controls the above-mentioned compilation and switching between the first quantum operation unit 20 and the second quantum operation unit 30.
- the classical computer 11 causes the first quantum arithmetic unit 20 to execute the compiled command, and interrupts the processing in the first quantum arithmetic unit 20 depending on the status of the arithmetic processing.
- the classical computer 11 compiles the remaining part of the interrupted process again, generates a command, and causes the second quantum operation unit 30 to execute the command, thereby transferring the process.
- the classical computer 11 further performs control to transfer the calculation result of the first quantum calculation unit 20 to the second quantum calculation unit 30. Since the second quantum operation unit 30 performs error correction, the classical computer 11 performs control to add an error correction code to the operation result.
- the classical computer 11 causes the first quantum operation unit 20 and the second quantum operation unit 30 to sequentially execute the generated commands.
- the classical computer 11 performs control to transfer the result of the arithmetic processing of the first quantum arithmetic unit 20 to the second quantum arithmetic unit 30 to execute the remaining arithmetic processing.
- the classical computer 11 transfers processing to the second quantum operation unit 30 depending on the processing status of the first quantum operation unit 20.
- the amount of error in the arithmetic processing of the first quantum arithmetic unit 20 can be adopted as the processing situation of the first quantum arithmetic unit 20 that causes this processing shift.
- the first quantum operation unit 20 accumulates errors as it performs operations. Before this error exceeds an allowable range, the classical computer 11 performs control to transfer the arithmetic processing to the second quantum arithmetic unit 30.
- an estimated value of the accumulated error amount can be used.
- This error amount can be calculated, for example, based on processing time (time elapsed from initialization). Further, the error amount can also be calculated based on, for example, the number of one-qubit gate operations. Further, the error amount can also be calculated based on, for example, a two-qubit gate operation number. Letting the error amount be C op , C op can be expressed by the following equation.
- t represents processing time.
- N s represents the number of 1-qubit gate operations.
- Nd represents the number of 2-qubit operations.
- a, r s and r d represent coefficients. Note that the error also includes deterioration of the quantum state.
- the classical computer 11 includes a register and the like to hold the calculated error amount, and updates the error amount based on information from the first quantum operation unit 20 and the like. If this error amount C op exceeds a predetermined threshold C th , the classical computer 11 interrupts the operation of the first quantum operation unit 20, saves the operation result, and causes the second quantum operation unit 30 to perform the processing. Migrate.
- the threshold value C th is determined based on the accuracy of the solution, the time required to derive the solution, the power consumption of the information processing device 1, and an index for leveling the usage time of the first quantum operation unit 20 and the second quantum operation unit 30. An adjusted C th can be used. Moreover, C th that is a combination of these can also be used. To improve the accuracy of the solution, set C th to a relatively small value. Furthermore, when reducing the power consumption of the information processing device 1, C th is set to a relatively large value.
- FIG. 6 is a diagram illustrating an example of an information processing procedure according to the first embodiment of the present disclosure. The figure is a flowchart showing an example of the processing procedure of the information processing device 1.
- step S100 quantum gate operation, which is calculation processing, is performed in the first quantum calculation unit 20 (step S100).
- the control unit 10 updates the error amount (step S101).
- step S102 determines whether the error amount exceeds a threshold (step S102). As a result, if the error amount does not exceed the threshold (step S102, No), the process returns to step S100, and the control unit 10 causes the first quantum operation unit 20 to continue the quantum gate operation (step S100).
- step S102 if the error amount exceeds the threshold (step S102, Yes), the control unit 10 controls the process to be transferred to the second quantum operation unit 30. Specifically, the control unit 10 maps the processing result of the first quantum operation unit 20 to the error correction code area (step S103).
- step S104 the second quantum calculation unit 30 performs quantum gate operation and error correction processing, which are calculation processing (step S104).
- the control unit 10 determines whether the processing of the second quantum calculation unit 30 has ended (step S105). If the processing has not been completed (step S105, No), the process returns to step S104, and the control unit 10 causes the second quantum operation unit 30 to continue the quantum gate operation and error correction processing (step S104).
- step S105 if the process has ended (step S105, Yes), the control unit 10 reads out the calculation result (step S106), and ends the process.
- the information processing device 1 can perform information processing.
- step S100 the process from step S100 to step S102 is referred to as NISQ mode.
- step S103 the process in step S103 is referred to as a transition mode.
- processing from step S104 to step S106 is referred to as FTQC mode.
- the time obtained by subtracting the time required for the processing in step S103 in FIG. 6 can be applied to the processing time t. This is because errors and deterioration of the quantum state also occur during the mapping in step S103.
- FIG. 7A is a diagram illustrating an example of a circuit for transferring calculation results according to the first embodiment of the present disclosure.
- This figure shows an example of a circuit that transfers processing results from the first quantum operation unit 20 to the second quantum operation unit 30.
- the upper circuit in the figure represents the second quantum operation unit 30, and the lower circuit represents the first quantum operation unit 20.
- One qubit data of the first quantum operation unit 20 is converted to data of five qubits q 0 to q 4 with error codes in the second quantum operation unit 30 .
- the transition is performed by adjusting the amplitude of each basis using a control-unitary gate or a Toffoli gate, according to the correspondence between the logical qubit information and physical qubit information used in the error correction code. This can be done by moving to the second quantum operation unit 30.
- the circuit shown in the figure allows the processing result of the first quantum operation unit 20 to be mapped to the error correction code area. Note that instead of the circuit shown in the same figure, multiple quantum bits in an entangled state are shared in advance, and the quantum state is transferred from the first quantum operation unit 20 to the second quantum operation by quantum communication (quantum teleportation). It can also be transferred to unit 30.
- FIG. 7B shows an example in which the error correction code is represented by a stabilizer code.
- Data with error codes of 5 qubits from q 0 to q 4 can be represented by a stabilizer of 5 qubits in the table of the figure.
- the correspondence relationship formed by the 5-qubit stabilizer in the table of the same figure can be expressed by the mathematical formula of the same figure.
- the information processing device 1 performs arithmetic processing without error correction in the first quantum arithmetic unit 20. Further, the information processing device 1 monitors the amount of error in this arithmetic processing, and before the amount of error exceeds the allowable range, the information processing device 1 shifts to processing in the second quantum arithmetic unit 30 and performs arithmetic processing accompanied by error correction. Thereby, calculation processing can be performed by combining the first quantum calculation unit 20 and the second quantum calculation unit 30.
- the information processing device 1 of the first embodiment described above has shifted from the arithmetic processing of the first quantum arithmetic unit 20 to the arithmetic processing of the second quantum arithmetic unit 30.
- the information processing device 1 according to the second embodiment of the present disclosure transitions from the calculation processing of the second quantum calculation unit 30 to the calculation processing of the first quantum calculation unit 20. This embodiment differs from the embodiment of .
- the information processing device 1 shifts from the calculation processing of the second quantum calculation unit 30 to the calculation processing of the first quantum calculation unit 20. For this reason, the control unit 10 of the information processing device 1 distributes calculation processing to the second quantum calculation unit 30 and the first quantum calculation unit 20. Then, when compiling the arithmetic processing, the control unit 10 embeds a command for outputting a flag indicating the end of the arithmetic processing in the second quantum arithmetic unit 30 and performs the compilation. Note that the configuration of the information processing device 1 is the same as that of the information processing device 1 in FIG. 5, so a description thereof will be omitted.
- FIG. 8 is a diagram illustrating an example of an information processing procedure according to the second embodiment of the present disclosure. Similar to FIG. 6, this figure is a flowchart showing an example of the processing procedure of the information processing device 1. The processing procedure differs from the processing procedure in FIG. 6 in that processing is performed in the order of FTQC mode and NISQ mode.
- step S110 quantum gate operation and error correction processing, which are calculation processing, are performed in the second quantum calculation unit 30 (step S110).
- step S111 the control unit 10 determines whether to shift to NISQ mode. This can be done by detecting a flag from the second quantum computing unit 30. If the NISQ mode is not to be entered (step S111, No), the process returns to step S110, and the control unit 10 causes the second quantum operation unit 30 to continue quantum gate operation and error correction processing (step S110).
- step S111 when shifting to the NISQ mode (step S111, Yes), control is performed to shift to the processing of the first quantum operation unit 20. Specifically, the control unit 10 maps the processing result of the second quantum calculation unit 30 to the NISQ domain (step S112).
- step S113 the first quantum operation unit 20 performs a quantum gate operation, which is an operation process (step S113).
- the control unit 10 determines whether the error amount exceeds a threshold (step S114). As a result, if the error amount does not exceed the threshold (step S114, No), the process moves to step S115.
- step S115 the control unit 10 determines whether the processing of the first quantum operation unit 20 has ended (step S115). If the processing has not ended (step S115, No), the process returns to step S113, and the control unit 10 causes the first quantum operation unit 20 to continue the quantum gate operation (step S113).
- step S115 Yes
- the control unit 10 moves to the process in step S116.
- step S114 if the error amount exceeds the threshold (step S114, Yes), the control unit 10 moves to the process of step S116.
- step S116 the control unit 10 reads out the calculation result (step S116), and ends the process.
- the information processing device 1 can perform information processing. Note that the process of step S114 in the figure can also be omitted.
- step S110 to step S112 corresponds to the FTQC mode. Furthermore, the process in step S112 corresponds to the transition mode. Further, the processing from step S113 to step S116 corresponds to the NISQ mode.
- FIG. 9 is a diagram illustrating an example of a circuit for transferring calculation results according to the second embodiment of the present disclosure.
- This figure shows an example of a circuit that transfers processing results from the second quantum operation unit 30 to the first quantum operation unit 20. Similar to FIG. 7A, the upper circuit in the figure represents the second quantum operation unit 30, and the lower circuit represents the first quantum operation unit 20.
- the logic state of the second quantum operation unit 30 is represented by a linear sum of states made up of multiple qubits.
- the calculation result can be converted to data of the first quantum calculation unit 20 by performing a multi Toffoli gate calculation after performing appropriate base conversion.
- the processing result of the second quantum operation unit 30 including the error correction code can be mapped to the area of the first quantum operation unit 20.
- the information processing device 1 first selects the FTQC mode and shifts to the NISQ mode (processing procedure in FIG. 8), and first selects the NISQ mode and shifts to the FTQC mode (processing procedure in FIG. 6). ) can also be selected and executed. Further, the information processing device 1 can also alternately operate the NISQ mode and the FTQC mode multiple times.
- the configuration of the information processing device 1 other than this is the same as the configuration of the information processing device 1 in the first embodiment of the present disclosure, so a description thereof will be omitted.
- the information processing device 1 transfers the calculation process from the second quantum calculation unit 30 to the first quantum calculation unit 20.
- the second quantum operation unit 30 performs error correction processing.
- the information processing device 1 according to the third embodiment of the present disclosure differs from the above-described first embodiment in that the coding distance for error correction is adjusted.
- FIG. 10 is a diagram illustrating a schematic configuration example of an information processing device according to a third embodiment of the present disclosure.
- This figure like FIG. 5, is a block diagram showing a configuration example of the information processing device 1.
- the control unit 10 in the figure differs from the information processing apparatus 1 in FIG. 5 in that it further includes an error rate detection unit 13 and performs control to alternately execute the NISQ mode and the FTQC mode a plurality of times.
- the error rate detection unit 13 detects and holds the error rate in the calculation process of the second quantum calculation unit 30.
- the classical computer 11 in the figure further performs control to adjust the coding distance based on the error rate held in the error rate detection section 13.
- the code distance can be adjusted in this plurality of error correction encodings. For example, if the frequency of error correction is lower than expected and the error rate of the system is small, a code with a smaller coding distance can be selected. This makes it possible to reduce the number of quantum bits and the calculations required for error correction encoding. Furthermore, if the frequency of error correction is higher than expected, the coding distance can be increased to improve error correction capability.
- FIG. 11 is a diagram illustrating an example of an information processing procedure according to the third embodiment of the present disclosure. Similar to FIG. 6, this figure is a flowchart showing an example of the processing procedure of the information processing device 1.
- control unit 10 executes the NISQ mode (step S151).
- control unit 10 executes the transition mode (step S152).
- the control unit 10 performs error correction encoding of the encoding distance a.
- the control unit 10 executes the FTQC mode (step S153).
- the control unit 10 alternately executes the NISQ mode and the FTQC mode. At this time, the control unit 10 adjusts the coding distance based on the detection result of the error rate detection unit 13, and generates the coding distance b.
- control unit 10 executes the NISQ mode (step S154) and executes the transition mode (step S155). In this transition mode, the control unit 10 performs error correction encoding of the encoding distance b. Next, the control unit 10 executes the FTQC mode (step S156). Through the above processing, the information processing device 1 can perform information processing.
- the configuration of the information processing device 1 other than this is the same as the configuration of the information processing device 1 in the first embodiment of the present disclosure, so a description thereof will be omitted.
- the information processing device 1 performs the NISQ mode and the FTQC mode multiple times and adjusts the encoding distance during error correction encoding. Thereby, the error correction encoding process can be optimized.
- FIG. 12 is a diagram illustrating a configuration example of an information processing device according to a modification of the embodiment of the present disclosure.
- the figure is a block diagram showing an example of the configuration of the information processing device.
- the information processing device 4 in the figure includes n information processing devices (information processing devices 1a to 1b), n classical computers (classical computers 2a to 2b), and a virtualization layer 3.
- This virtualization layer 3 virtualizes a plurality of information processing devices to which the first quantum operation unit 20 and the second quantum operation unit 30 are allocated.
- the present technology can also have the following configuration.
- control unit performs control to transfer the arithmetic processing to the second quantum arithmetic unit based on the error amount according to the number of one-qubit gate operations in the arithmetic processing.
- control unit performs control to transfer the arithmetic processing to the second quantum arithmetic unit based on the error amount according to the number of two-qubit gate operations in the arithmetic processing.
- control unit further performs control to transfer the arithmetic processing to the first quantum arithmetic unit after selecting the second quantum arithmetic unit and causing the arithmetic processing to be performed.
- the information processing device according to any one of. (6) The information according to any one of (1) to (5) above, wherein the control unit further performs control to transfer the arithmetic processing transferred to the second quantum arithmetic unit to the first quantum arithmetic unit again. Processing equipment. (7) The information processing device according to any one of (1) to (6), wherein the control unit adjusts the coding distance for the error correction in the second quantum operation unit.
Landscapes
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Mathematical Analysis (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Condensed Matter Physics & Semiconductors (AREA)
- Computational Mathematics (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Error Detection And Correction (AREA)
Abstract
Description
1.背景技術
2.第1の実施形態
3.第2の実施形態
4.第3の実施形態
5.変形例
上述のNISQは、量子操作の度に生じるノイズの影響により、量子性が保たれる時間が有限長である。このため、NISQは、有限の回数の量子操作しか行うことができないという問題がある。このNISQに対して、誤りを訂正しながら量子演算を行う誤り訂正量子計算機(FTQC:Fault Tolerant Quantum Computer)は、任意の量子演算を任意の回数実行可能である。このFTQCを実現するには、多数の物理量子ビットを用いて誤り検出訂正符号(ECC:Error detection and Correction Code)を実装する必要がある。例えば、ECC無しのエラー率が0.1%の場合、1論理量子ビットの構成には1000個の物理量子ビットが必要とされる。このECCを実装した場合であっても、クリフォード演算であれば効率的に処理ができる。しかし、量子優位性を得るための要となる非クリフォード演算を行う場合には、Tゲートと呼ばれる操作が必要となる。このTゲートを作用させるには魔法状態(magic state)と呼ばれる特殊な量子状態を生成する必要がある。ECCを実装した状態において量子演算を実行するには、この魔法状態が大量に必要となる。例えば、FTQCの量子ビットの90%以上は、魔法状態の生成に費やされると試算されている。
[情報処理装置の構成]
図2は、本開示の第1の実施形態に係る情報処理装置の概略の構成例を示す図である。同図は、情報処理装置1の概略の構成例を表すブロック図である。情報処理装置1は、制御部10と、第1の量子演算ユニット20と、第2の量子演算ユニット30とを備える。
Cop=f(t,Ns,Nd)=a×t+rs×Ns+rd×Nd
ここで、tは処理時間を表す。Nsは1量子ビットゲート演算数を表す。Ndは2量子ビット演算数を表す。a、rs及びrdは係数を表す。なお、誤りには、量子状態の劣化も含まれる。
図6は、本開示の第1の実施形態に係る情報処理の処理手順の一例を示す図である。同図は、情報処理装置1の処理手順の一例を表す流れ図である。
図7Aは、本開示の第1の実施形態に係る演算結果を移行する回路の一例を示す図である。同図は、第1の量子演算ユニット20から第2の量子演算ユニット30に処理結果を移行する回路の例を表したものである。同図の上側の回路が第2の量子演算ユニット30の部分を表し、下側の回路が第1の量子演算ユニット20の部分を表す。第1の量子演算ユニット20の1量子ビットのデータが第2の量子演算ユニット30のq0乃至q4の5量子ビットの誤り符号付きのデータに変換される。移行は、誤り訂正符号にて使用する論理量子ビット情報と物理量子ビット情報の対応関係に従い、制御ユニタリゲート(Control-unitary gate)やトフォリゲート(Toffoli gate)を用いて各基底の振幅を第2の量子演算ユニット30に移すことにより行うことができる。
上述の第1の実施形態の情報処理装置1は、第1の量子演算ユニット20の演算処理から第2の量子演算ユニット30の演算処理に移行していた。これに対し、本開示の第2の実施形態の情報処理装置1は、第2の量子演算ユニット30の演算処理から第1の量子演算ユニット20の演算処理に移行する点で、上述の第1の実施形態と異なる。
図8は、本開示の第2の実施形態に係る情報処理の処理手順の一例を示す図である。同図は、図6と同様に、情報処理装置1の処理手順の一例を表す流れ図である。FTQCモード及びNISQモード及の順に処理を行う点で、図6の処理手順と異なる。
図9は、本開示の第2の実施形態に係る演算結果を移行する回路の一例を示す図である。同図は、第2の量子演算ユニット30から第1の量子演算ユニット20に処理結果を移行する回路の例を表したものである。図7Aと同様に、同図の上側の回路が第2の量子演算ユニット30の部分を表し、下側の回路が第1の量子演算ユニット20の部分を表す。第2の量子演算ユニット30は、多量子ビットで構成される状態の線形和によって論理状態が示される。演算結果の移行は、適当な基底変換を行った後にマルチトフォリゲート(multi Toffoli gate)演算を行うことにより、第1の量子演算ユニット20のデータに変換することができる。
上述の第1の実施形態の情報処理装置1は、第2の量子演算ユニット30において誤り訂正処理を行っていた。これに対し、本開示の第3の実施形態の情報処理装置1は、誤り訂正のための符号化距離を調整する点で、上述の第1の実施形態と異なる。
図10は、本開示の第3の実施形態に係る情報処理装置の概略の構成例を示す図である。同図は、図5と同様に、情報処理装置1の構成例を表すブロック図である。同図の制御部10は誤り率検出部13を更に備え、NISQモード及びFTQCモードを交互に複数回実行する制御を行う点で、図5の情報処理装置1と異なる。
図11は、本開示の第3の実施形態に係る情報処理の処理手順の一例を示す図である。同図は、図6と同様に、情報処理装置1の処理手順の一例を表す流れ図である。
[情報処理装置の構成]
図12は、本開示の実施形態の変形例に係る情報処理装置の構成例を示す図である。同図は、情報処理装置の構成例を表すブロック図である。同図の情報処理装置4は、n個の情報処理装置(情報処理装置1a乃至1b)と、n個の古典計算機(古典計算機2a乃至2b)と、仮想化層3とを備える。この仮想化層3は、第1の量子演算ユニット20及び第2の量子演算ユニット30の割り振りを行う複数の情報処理装置を仮想化するものである。
(1)
量子ビットを備え、量子演算を行う第1の量子演算ユニットと、
量子ビットを備え、誤り訂正を伴う量子演算を行う第2の量子演算ユニットと、
前記第1の量子演算ユニットに演算処理を行わせるとともに、前記第1の量子演算ユニットの演算処理における誤り量に基づいて前記演算処理を前記第2の量子演算ユニットに移行させる制御を行う制御部と
を有する情報処理装置。
(2)
前記制御部は、前記演算処理の処理時間に応じた前記誤り量に基づいて前記演算処理を前記第2の量子演算ユニットに移行させる制御を行う前記(1)に記載の情報処理装置。
(3)
前記制御部は、前記演算処理における1量子ビットゲート演算数に応じた前記誤り量に基づいて前記演算処理を前記第2の量子演算ユニットに移行させる制御を行う前記(1)に記載の情報処理装置。
(4)
前記制御部は、前記演算処理における2量子ビットゲート演算数に応じた前記誤り量に基づいて前記演算処理を前記第2の量子演算ユニットに移行させる制御を行う前記(1)に記載の情報処理装置。
(5)
前記制御部は、前記第2の量子演算ユニットを選択して前記演算処理を行わせた後に前記演算処理を前記第1の量子演算ユニットに移行させる制御を更に行う前記(1)から(4)の何れかに記載の情報処理装置。
(6)
前記制御部は、前記第2の量子演算ユニットに移行させた前記演算処理を前記第1の量子演算ユニットに再度移行させる制御を更に行う前記(1)から(5)の何れかに記載の情報処理装置。
(7)
前記制御部は、前記第2の量子演算ユニットにおける前記誤り訂正のための符号化距離を調整する前記(1)から(6)の何れかに記載の情報処理装置。
2a、2b、11 古典計算機
10 制御部
13 誤り率検出部
20 第1の量子演算ユニット
30 第2の量子演算ユニット
Claims (7)
- 量子ビットを備え、量子演算を行う第1の量子演算ユニットと、
量子ビットを備え、誤り訂正を伴う量子演算を行う第2の量子演算ユニットと、
前記第1の量子演算ユニットに演算処理を行わせるとともに、前記第1の量子演算ユニットの演算処理における誤り量に基づいて前記演算処理を前記第2の量子演算ユニットに移行させる制御を行う制御部と
を有する情報処理装置。 - 前記制御部は、前記演算処理の処理時間に応じた前記誤り量に基づいて前記演算処理を前記第2の量子演算ユニットに移行させる制御を行う請求項1に記載の情報処理装置。
- 前記制御部は、前記演算処理における1量子ビットゲート演算数に応じた前記誤り量に基づいて前記演算処理を前記第2の量子演算ユニットに移行させる制御を行う請求項1に記載の情報処理装置。
- 前記制御部は、前記演算処理における2量子ビットゲート演算数に応じた前記誤り量に基づいて前記演算処理を前記第2の量子演算ユニットに移行させる制御を行う請求項1に記載の情報処理装置。
- 前記制御部は、前記第2の量子演算ユニットを選択して前記演算処理を行わせた後に前記演算処理を前記第1の量子演算ユニットに移行させる制御を更に行う請求項1に記載の情報処理装置。
- 前記制御部は、前記第2の量子演算ユニットに移行させた前記演算処理を前記第1の量子演算ユニットに再度移行させる制御を更に行う請求項1に記載の情報処理装置。
- 前記制御部は、前記第2の量子演算ユニットにおける前記誤り訂正のための符号化距離を調整する請求項1に記載の情報処理装置。
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2024511926A JPWO2023189886A1 (ja) | 2022-03-31 | 2023-03-22 | |
EP23779900.2A EP4502875A1 (en) | 2022-03-31 | 2023-03-22 | Information processing device |
CN202380029313.7A CN119032365A (zh) | 2022-03-31 | 2023-03-22 | 信息处理装置 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2022-061245 | 2022-03-31 | ||
JP2022061245 | 2022-03-31 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2023189886A1 true WO2023189886A1 (ja) | 2023-10-05 |
Family
ID=88201298
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2023/011095 WO2023189886A1 (ja) | 2022-03-31 | 2023-03-22 | 情報処理装置 |
Country Status (4)
Country | Link |
---|---|
EP (1) | EP4502875A1 (ja) |
JP (1) | JPWO2023189886A1 (ja) |
CN (1) | CN119032365A (ja) |
WO (1) | WO2023189886A1 (ja) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2009260882A (ja) * | 2008-04-21 | 2009-11-05 | Sumitomo Electric Ind Ltd | 復号化装置及び光通信システムの宅内装置 |
CN111510157A (zh) * | 2020-04-15 | 2020-08-07 | 腾讯科技(深圳)有限公司 | 基于神经网络的量子纠错解码方法、装置及芯片 |
US20210272002A1 (en) | 2019-09-27 | 2021-09-02 | Zapata Computing, Inc. | Computer Systems and Methods for Computing the Ground State of a Fermi-Hubbard Hamiltonian |
-
2023
- 2023-03-22 CN CN202380029313.7A patent/CN119032365A/zh active Pending
- 2023-03-22 EP EP23779900.2A patent/EP4502875A1/en active Pending
- 2023-03-22 WO PCT/JP2023/011095 patent/WO2023189886A1/ja active Application Filing
- 2023-03-22 JP JP2024511926A patent/JPWO2023189886A1/ja active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2009260882A (ja) * | 2008-04-21 | 2009-11-05 | Sumitomo Electric Ind Ltd | 復号化装置及び光通信システムの宅内装置 |
US20210272002A1 (en) | 2019-09-27 | 2021-09-02 | Zapata Computing, Inc. | Computer Systems and Methods for Computing the Ground State of a Fermi-Hubbard Hamiltonian |
CN111510157A (zh) * | 2020-04-15 | 2020-08-07 | 腾讯科技(深圳)有限公司 | 基于神经网络的量子纠错解码方法、装置及芯片 |
Also Published As
Publication number | Publication date |
---|---|
CN119032365A (zh) | 2024-11-26 |
EP4502875A1 (en) | 2025-02-05 |
JPWO2023189886A1 (ja) | 2023-10-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111417965B (zh) | 软件限定的量子计算机 | |
CN109858628B (zh) | 编译量子电路的方法、装置、设备和计算机可读存储介质 | |
Balensiefer et al. | An evaluation framework and instruction set architecture for ion-trap based quantum micro-architectures | |
Tannu et al. | Taming the instruction bandwidth of quantum computers via hardware-managed error correction | |
US20220199888A1 (en) | Automatic Real-Time Calibration of Qubit Chip | |
Alam et al. | An efficient circuit compilation flow for quantum approximate optimization algorithm | |
EP3803716A1 (en) | Generating quantum computing circuits by distributing approximation errors in a quantum algorithm | |
Almudever et al. | Realizing quantum algorithms on real quantum computing devices | |
Lin et al. | Scalable optimal layout synthesis for NISQ quantum processors | |
Peham et al. | On optimal subarchitectures for quantum circuit mapping | |
WO2022206842A1 (zh) | 量子线路编译方法、装置、编译框架及量子操作系统 | |
Metodi et al. | Scheduling physical operations in a quantum information processor | |
Holmes et al. | Resource optimized quantum architectures for surface code implementations of magic-state distillation | |
Paler et al. | Really small shoe boxes: On realistic quantum resource estimation | |
US11954563B2 (en) | Apparatus and method for error reduction in distributed quantum systems via fusing-and-decomposing gates | |
WO2023189886A1 (ja) | 情報処理装置 | |
Niemann et al. | Combining SWAPs and remote Toffoli gates in the mapping to IBM QX architectures | |
CN114970866B (zh) | 一种量子计算任务计算方法、装置及可读存储介质 | |
US20230016817A1 (en) | Apparatus and method for coherent error mitigation using clifford gate injection | |
Hwang et al. | Integrated analysis of performance and resources in large-scale quantum computing | |
Abdessaied et al. | Reversible circuit rewriting with simulated annealing | |
Yuan et al. | The T-complexity costs of error correction for control flow in quantum computation | |
Zhang et al. | MECH: Multi-entry communication highway for superconducting quantum chiplets | |
CN113269325B (zh) | 基于指令重排的量子程序执行方法及装置 | |
CN114912615A (zh) | 一种量子计算机操作系统及量子计算机 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 23779900 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2024511926 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202380029313.7 Country of ref document: CN |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2023779900 Country of ref document: EP |
|
ENP | Entry into the national phase |
Ref document number: 2023779900 Country of ref document: EP Effective date: 20241031 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |