Calibration method among multiple channels of photon calculation chip

Abstract

The invention discloses a method for calibrating multiple channels of a photon calculation chip, which comprises the following steps of: inputting same calibration optical signals to a calculation module channel by channel, and acquiring a calibration output value of each channel; selecting a certain channel as a reference channel, and taking a calibration output value of the reference channel as a reference calibration output value; the other channels are used as channels to be calibrated; and the input optical signal of the calculation module of each to-be-calibrated channel is adjusted according to the external control electric signal, so that the calibration output value of each to-be-calibrated channel of the calculation module is the reference calibration output value under the condition that the light source module inputs the calibration optical signal. Original errors caused by manufacturing process errors and optical fiber coupling are adjusted through input optical signals, so that final output structures are consistent. According to the invention, the calibration module is arranged on hardware to adjust the optical signal, and a calibration coefficient does not need to be added into the calculation module, so that the burden of the calculation module is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of optical chips, and in particular relates to a calibration method between multiple channels of a photonic computing chip. Background Art

[0002] In the process of AI calculation, electronic chips play an important role in data transmission and calculation. The whole process of artificial neural algorithm contains a large number of matrix multiplication operations, and the CPU based on traditional computer architecture is very laborious to process these operations, and the calculation efficiency is low. Therefore, academia and industry have turned their attention to new hardware structures specifically used for artificial neural networks and deep learning, such as GPU, ASIC and FPGA. However, no matter what kind of technical architecture, it is designed and manufactured using traditional microelectronics technology. The improvement of AI chip performance is inseparable from the improvement of microelectronics integration. However, at the beginning of the 21st century, it was difficult for microelectronics technology to develop according to the prediction of Moore's Law, and the difficulty of improving chip integration continued to increase.

[0003] Compared with electrons, photons have many unique properties: photons have no rest mass, no interaction between photons, and almost no interference; different wavelengths of light can be used for multi-channel simultaneous communication, and at a modulation frequency of tens of hertz, photons can still obtain stable modulation and information transmission, while electrical signals face the problem of radiation loss at high frequencies; in addition, optical signals are not affected by electromagnetic fields and have strong confidentiality. Even more unique is that by using some specific optical structures, photons can perform some mathematical operations with zero energy consumption. Therefore, the use of photons can achieve ultra-high-speed, low-energy or even zero-energy computing, thereby breaking through the performance and cost bottlenecks of traditional microelectronic chips.

[0004] As the semiconductor industry gradually enters the post-Moore era, the development of integrated circuits continues to evolve in different directions. On the one hand, the development of new semiconductor materials, especially carbon nanotubes and two-dimensional semiconductor materials, continues the essence of Moore's Law and continues to reduce the size of devices or chips, that is, "More Moore"; on the other hand, for specific application fields, the development of new architectures and heterogeneous integrated chips, such as neuromorphic chips, optoelectronic chips, quantum chips, etc., achieves "More than Moore".

[0005] Among them, photonic computing chips based on silicon-based optoelectronic technology use materials and processes compatible with integrated circuits to integrate micron and nanoscale photonics, electronics and optoelectronic devices on the same silicon substrate to achieve functional integration and complementary advantages of microelectronic devices and optoelectronic devices, and obtain optoelectronic chips with superior performance, which is an effective way to solve the performance bottleneck and information congestion faced by traditional integrated circuits. Thanks to the mature application of optical fiber communication, photons, as information carriers, have more multiplexing dimensions than electrons, such as amplitude, phase, wavelength, mode, etc., and thus have larger bandwidth, faster speed and lower energy consumption. Early silicon-based optoelectronic chips were designed to replace copper interconnect technology and solve the bottleneck of interconnection communication between the processor core and memory of microelectronic chips. The microprocessor and storage unit are realized by microelectronic devices, while the photonic devices mainly complete signal transmission and information transmission. With the increasing maturity of silicon photonic technology and the huge advantages of optical communication, people's attention to silicon photonic computing chips has gradually shifted from information transmission to information processing, including analog computing, quantum computing, brain-like computing and other cutting-edge application fields.

[0006] Due to the process errors in the production of optical chips and the coupling errors of the optical fiber array, it is impossible for each calculation channel to be completely consistent. The inconsistency between channels will affect the calculation accuracy. Summary of the invention

[0007] The purpose of the present invention is to provide a calibration method between multiple channels of a photonic computing chip, which partially solves or alleviates the above-mentioned deficiencies in the prior art and can perform consistency calibration between computing channels, thereby ensuring the computing accuracy of the photonic computing chip.

[0008] In order to solve the above-mentioned technical problems, the present invention specifically adopts the following technical solutions: a calibration method between multiple channels of a photonic computing chip, comprising: Input the same calibration optical signal into the calculation module channel by channel to obtain the calibration output value of each channel; Select a channel as the reference channel, and use the calibration output value of the reference channel as the reference calibration output value; the remaining channels are used as channels to be calibrated; The input optical signal of the calculation module of each channel to be calibrated is adjusted, so that when the light source module inputs the calibration optical signal, the calibration output value of each channel to be calibrated of the calculation module is the reference calibration output value.

[0009] As an improvement, a calibration module is used to adjust the input optical signal of the channel to be calibrated; the calibration module includes a plurality of submodules, and the submodules correspond to the modulators in the modulator array one by one; the calibration module is arranged between the light source module and the modulator array, or between the modulator array and the calculation module; In the case where the calibration module is arranged between the light source module and the modulator array, the output optical signal of the light source module is the input optical signal of the calibration module; the calibration module adjusts the output optical signal of the light source module according to the external control electrical signal, thereby forming the input optical signal of the modulator array; When the calibration module is arranged between the modulator array and the computing module, the output optical signal of the modulator array is the input optical signal of the calibration module; the calibration module adjusts the output optical signal of the modulator array according to the external control electrical signal, thereby forming the input optical signal of the computing module.

[0010] As an improvement, the step of adjusting the input optical signal of each channel to be calibrated includes: Input the reference calibration output value as an independent variable into a matching function pre-fitted in the channel to be calibrated to obtain a corresponding external control electrical signal; or, search for an external control electrical signal corresponding to the reference calibration output value in a matching table pre-constructed in the channel to be calibrated; The output light signal from the light source module is adjusted according to the external control electrical signal by the calibration module.

[0011] As an improvement, the fitting step of the matching function includes: Recording several external control electrical signals and corresponding output values ​​of the channel to be calibrated under the calibration light signal; The external control electrical signal and the corresponding output value are regarded as discrete points in a two-dimensional coordinate system; Fit discrete points into a curve; The expression function of the curve is used as the matching function of the channel to be calibrated.

[0012] As an improvement, the method of fitting discrete points into a curve is one of least squares fitting, polynomial fitting, spline interpolation fitting, and non-parametric fitting.

[0013] As an improvement, the step of constructing the matching table includes: According to the set step length, all external control electrical signals and corresponding output values ​​of the channel to be calibrated under the calibration light signal are obtained; A matching table of the channel to be calibrated is constructed according to the acquired external control electrical signal and the corresponding output value.

[0014] As an improvement, the steps for obtaining the set step size include: Initializing the step length, and adjusting the external control electrical signal according to the initialization step length; Compare the output values ​​of the optical signal modulator under adjacent external control electrical signals; if the absolute value of the difference is less than the difference threshold range, add the current step size to the preset incremental step size as the step size for the next adjustment; if the absolute value of the difference is within the difference threshold range, the step size remains unchanged; if the absolute value of the difference is greater than the difference threshold range, subtract the incremental step size from the current step size as the step size for the next adjustment.

[0015] As an improvement, the external control electrical signal is adjusted upward from the lower limit of the normal working voltage range of the external control electrical signal, or the external control electrical signal is adjusted downward from the upper limit of the normal working voltage range of the external control electrical signal.

[0016] As an improvement, the external control electrical signal is voltage or current.

[0017] As an improvement, the channel with the smallest calibration output value is used as the reference channel.

[0018] Beneficial effects: The photonic computing chip provided by the present invention is provided with a calibration module; the calibration module includes a plurality of sub-modules, and the sub-modules correspond one-to-one to the optical signal modulators in the modulator array; the calibration module can adjust the input optical signal of the computing module on each channel to be calibrated, so that when the light source module inputs the calibration optical signal, the calibration output value of each channel to be calibrated of the computing module is the benchmark calibration output value.

[0019] The present invention changes the input of the corresponding optical signal path in the calculation module by adjusting the output optical signal of the light source module or the output optical signal of the modulator. The inconsistency of the optical signal path originally caused by the manufacturing process error and the optical fiber coupling precision error is adjusted so that the final output structure is consistent. The present invention changes the optical signal by setting a calibration module in hardware, and there is no need to add a calibration coefficient in the calculation module, thereby reducing the burden of the calculation module.

[0020] In addition, the present invention also provides a calibration method for a photonic chip based on the above-mentioned photonic computing chip. During calibration, the light source module inputs the same calibration light signal to the computing module channel by channel to obtain the calibration output value of each channel. The channel with the smallest calibration output value is used as the reference channel, and the calibration output value of the reference channel is used as the reference calibration output value; the remaining channels are used as channels to be calibrated; the calibration module adjusts the input light signal of the computing module of each channel to be calibrated according to the external control electrical signal, so that when the light source module inputs the calibration light signal, the calibration output value of each channel to be calibrated of the computing module is the reference calibration output value.

[0021] After the calibration is completed, the external control electrical signal of the calibration module submodule corresponding to each channel is solidified when the output reference calibration output value is output, so that the optical signal is pre-adjusted each time it is used in the future. No soft calibration is required in the calculation module.

[0022] The present invention also provides a method for obtaining the specific value of the external control electrical signal, one of which is a function fitting method and the other is a table lookup method. Through the above two methods, as long as the reference calibration output value is obtained, the required external control electrical signal can be inferred, which is simple and fast. When the optical chip is initialized, the calculation channel can be calibrated quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, each element or part is not necessarily drawn according to the actual scale. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings without paying creative labor.

[0024] Figure 1 It is a structural schematic diagram of the present invention; Figure 2 A schematic diagram of another structure of the present invention; Figure 3 Flow chart of the calibration method of the present invention. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0026] Herein, suffixes such as "module", "component" or "unit" used to represent elements are only used to facilitate the description of the present invention, and have no specific meanings by themselves. Therefore, "module", "component" or "unit" can be used mixedly.

[0027] In this document, the terms "upper", "lower", "inner", "outer", "front", "back", "one end", "the other end" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance.

[0028] In this document, unless otherwise clearly specified and limited, the terms "installed", "provided with", "connected", etc. should be understood in a broad sense. For example, "connected" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, a direct connection, or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0029] Herein "and / or" includes any and all combinations of one or more of the associated listed items.

[0030] Herein, "plurality" means two or more than two, ie, it includes two, three, four, five, etc.

[0031] Embodiment 1: Prior art Chinese patent application CN201310044171.X discloses a method for calibrating the consistency of laser paths of an optical vector-matrix multiplier, the optical vector-matrix multiplier comprising: a microprocessing module and a plurality of laser paths controlled by the microprocessing module; the method is executed by the microprocessing module, comprising: obtaining a correction coefficient k of each laser path with reference to a reference laser path before actual operation; [n] and b [n] ; and in the actual operation process, the input vector V of the current laser path is corrected by the correction coefficient k [n] and b [n] Corrected to V′=k [n] V+b [n] .

[0032] In the above-mentioned prior art, the correction coefficient of the reference laser path is first obtained, and then the input vector of the current laser path is corrected using the correction coefficient. This is a soft calibration, and the correction coefficient is used as a parameter in the calculation, which will bring additional calculation to the calculation unit, increase the power consumption of the calculation unit, and increase the system overhead.

[0033] To solve this problem, Figure 1 , Figure 2As shown, the present invention provides a photonic computing chip with a multi-channel calibration module, which realizes the calibration between the channels in the photonic computing chip from the hardware level.

[0034] The computing power of a photonic computing chip is proportional to the size of its computing matrix. The greater the computing power, the larger the matrix size. As the matrix size increases, the number of channels also increases. For example, a 64X64 computing matrix includes 64 optical signal channels and 64 electrical signal channels. Similarly, a 128X128 computing matrix includes 128 optical signal channels and 128 electrical signal channels. Due to differences in technology, the optical signal channels in a photonic computing chip cannot be completely consistent, which affects the computing accuracy.

[0035] The problem to be solved in the present invention is to calibrate the consistency of several optical signal channels in the calculation matrix (the calibration method of the electrical signal channel will not be described in detail in the present invention).

[0036] Specifically, the photonic computing chip provided by the present invention specifically includes: A light source module for providing a light source, a modulator array composed of a plurality of modulators for modulating the light signal output by the light source module, and a computing module for performing optical calculations; further comprising a calibration module, wherein the calibration module comprises a plurality of submodules, and the submodules correspond one-to-one to the light signal modulators in the modulator array; the calibration module can adjust the input light signal of the computing module.

[0037] The present invention changes the input of the corresponding optical signal path in the calculation module by adjusting the output optical signal of the light source module or the output optical signal of the modulator. The inconsistency of the optical signal path originally caused by the manufacturing process error and the optical fiber coupling precision error is adjusted so that the final output structure is consistent. The present invention changes the optical signal by setting a calibration module in hardware, and there is no need to add a calibration coefficient in the calculation module, thereby reducing the burden of the calculation module.

[0038] In this embodiment, the calibration module may be arranged between the light source module and the modulator array, or between the modulator array and the calculation module.

[0039] In the case where the calibration module is arranged between the light source module and the modulator array, the output optical signal of the light source module is the input optical signal of the calibration module; the calibration module adjusts the output optical signal of the light source module according to the external control electrical signal, thereby forming the input optical signal of the modulator array, such as Figure 1 shown.

[0040] In the case where the calibration module is arranged between the modulator array and the computing module, the output optical signal of the modulator array is the input optical signal of the calibration module; the calibration module adjusts the output optical signal of the modulator array according to the external control electrical signal, thereby forming the input optical signal of the computing module, such as Figure 2 shown.

[0041] Embodiment 2: The present invention also provides an optoelectronic hybrid computing system, including the photonic computing chip with a multi-channel calibration module provided in Embodiment 1; more specifically, the optoelectronic hybrid computing system also includes a control module, and the controller is connected to a DAC module; the DAC module converts the digital signal emitted by the control module into an analog signal and sends it to the calibration module and the modulator array respectively, and in some embodiments, it also needs to be sent to the computing module.

[0042] Some computing modules use the optical signals sent by the modulator array and the electrical signals sent by the DAC module to implement matrix multiplication. In this case, the DAC module needs to send electrical signals to the computing module. In other computing modules, due to the different designs of the optical chips themselves, the calculation weights can also be adjusted through optical signals, thus eliminating the need for additional DAC modules to send electrical signals.

[0043] In the photonic computing chip, the function of the DAC (Digital-to-Analog Converter) module is to convert digital signals into analog signals. In the present invention, the DAC module converts the control signal (digital signal) sent by the control module into analog signals and sends them to the calibration module, the modulator array and the computing module respectively, thereby controlling the above modules.

[0044] like Figure 1 As shown, in some embodiments, the DAC module includes a first DAC submodule connected to the calibration module, a second DAC submodule connected to the modulator array, and a third DAC submodule connected to the calculation module. The calibration module, the modulator array, and the calculation module correspond to a DAC submodule respectively, so as to facilitate receiving the control signal sent by the controller. Of course, as mentioned above, some calculation modules do not need to send electrical signals to the calculation module, so there is no need for a third DAC module, such as Figure 2 shown.

[0045] In addition, the photonic computing chip also includes a detection module and an ADC module connected to the output end of the computing module in sequence; the detection module converts the optical signal output by the computing module into an electrical signal; the output end of the ADC (Analog-to-Digital Converter) module is connected to the control module, thereby converting the output of the computing module from an analog signal to a digital signal and sending it to the controller.

[0046] In addition, the detection module includes a PD (Phtodetector) module and a TIA (Transimpedance amplifier) ​​module.

[0047] The function of the PD module is to convert the optical signal into an electrical signal, while the function of the TIA module is to amplify the weak current signal from the PD module into a voltage signal, so as to facilitate the subsequent ADC module to perform signal processing.

[0048] Embodiment 3: Figure 3 As shown, based on the above-mentioned photon computing chip with a multi-channel calibration module, the present invention also provides a multi-channel calibration method for a photon computing chip, and the specific steps thereof include: During S1 calibration, the light source module inputs the same calibration light signal to the calculation module channel by channel to obtain the calibration output value of each channel.

[0049] In this embodiment, calibration is performed when the photonic computing chip leaves the factory or in other scenarios where reinitialization is required.

[0050] First, the same calibration light signal is input into each light signal channel in the calculation module through the light source module, and the calibration output value inputted for each channel is obtained. It can be foreseen that since each channel has differences, the calibration output value outputted by each channel also has differences.

[0051] S2 selects a certain channel as a reference channel, and uses the calibration output value of the reference channel as a reference calibration output value; the remaining channels are used as channels to be calibrated.

[0052] In this step, one channel needs to be selected as the reference channel, and the remaining channels are used as channels to be calibrated, so that the output value of the channel to be calibrated is consistent with the output value of the reference channel under the same input.

[0053] In some embodiments, the channel with the smallest calibration output value is used as the reference channel. The reason is that in a photonic computing chip, the enhancement of the optical signal may be constrained by the characteristics of the device itself. For example, the intensity of the optical signal output by the light source module to a certain channel is fixed and cannot be increased. Therefore, the channel with the smallest calibration output value is selected as the reference channel. In order to achieve the same output, the other channels to be calibrated only need to reduce the intensity of the optical signal during transmission.

[0054] The S3 calibration module adjusts the input optical signal of the calculation module of each channel to be calibrated according to the external control electrical signal, so that when the light source module inputs the calibration optical signal, the calibration output value of each channel to be calibrated of the calculation module is the reference calibration output value.

[0055] In this step, the external control electric signal is actually the electric signal after the control signal sent by the control module is converted by the DAC module. The external control electric signal can be either a voltage signal or a current signal, which is not limited in the present invention.

[0056] In this embodiment, a calibration module is used to adjust the input optical signal of the channel to be calibrated; according to the description of the first embodiment, the calibration module includes a plurality of submodules, and the submodules correspond to the modulators in the modulator array one by one; the calibration module is arranged between the light source module and the modulator array, or between the modulator array and the calculation module. Therefore, in this embodiment, the calibration module adjusts the input optical signal of each channel to be calibrated according to the external control electrical signal, which also includes two situations: First, when the calibration module is arranged between the light source module and the modulator array, the output light signal of the light source module is the input light signal of the calibration module; the calibration module adjusts the output light signal of the light source module according to the external control electrical signal, thereby forming the input light signal of the modulator array.

[0057] Second, when the calibration module is arranged between the modulator array and the computing module, the output optical signal of the modulator array is the input optical signal of the calibration module; the calibration module adjusts the output optical signal of the modulator array according to the external control electrical signal, thereby forming the input optical signal of the computing module.

[0058] Under the control of the external control electrical signal, the calibration module adjusts the optical signal emitted by the light source module or the output optical signal of the modulator array through the submodule, thereby changing the output of the light source module or the modulator on the corresponding channel, and finally changing the input optical signal of the calculation module of each channel to be calibrated. It can be foreseen that the adjustment of the input optical signal of the channel to be calibrated is adjusted by a certain proportion, the purpose of which is to make the calibration output values ​​of all channels to be calibrated consistent with the reference calibration output value when the calibration optical signal is input to the light source module.

[0059] In this embodiment, the step of the calibration module adjusting the input optical signal of each channel to be calibrated according to the external control electrical signal includes: S31 inputs the reference calibration output value as an independent variable into a matching function pre-fitted for the channel to be calibrated to obtain a corresponding external control electrical signal; or, searches for an external control electrical signal corresponding to the reference calibration output value from a matching table pre-constructed for the channel to be calibrated.

[0060] This embodiment provides two methods for obtaining specific external control signals, one is a function fitting method, and the other is a table lookup method.

[0061] The fitting step of the matching function in the function fitting method includes: S311 records a number of external control electrical signals and corresponding output values ​​of the channel to be calibrated under the calibration optical signal.

[0062] For each channel to be calibrated, different external control electrical signals are used to intervene when the same calibration optical signal is input, and each external control electrical signal and the corresponding output value are recorded.

[0063] For example, when inputting the same calibration optical signal P ref Under the premise of , set the external control electrical signal I1 to obtain the output value S1 of the channel; set the external control electrical signal I2 to obtain the output value S2 of the channel; set the external control electrical signal I3 to obtain the output value S3 of the channel... and so on.

[0064] S312 takes the external control electrical signal and the corresponding output value as discrete points in a two-dimensional coordinate system.

[0065] For example, the values ​​obtained in the previous step are combined into discrete points in a two-dimensional coordinate system, that is, P1=(I1, S1), P2=(I2, S2), P3=(I3, S3)... It can be understood that in a two-dimensional coordinate system, the external control electrical signal value is the x-coordinate of the point, and the corresponding output value is the y-coordinate.

[0066] S313 fits the discrete points into a curve.

[0067] In this embodiment, the method for fitting discrete points into a curve is one of least squares fitting, polynomial fitting, spline interpolation fitting, and non-parametric fitting.

[0068] The least squares method is a commonly used data fitting method, especially suitable for fitting discrete points in a two-dimensional coordinate system into a curve. Its basic idea is to find the best fitting curve for the data by minimizing the sum of squares of errors. Its general steps include: 1. Determine the model function, 2. Establish the error function, 3. Solve the coefficients, 4. Calculate the fitting curve, and 5. Evaluate the fitting effect.

[0069] Since this method is a prior art, it will not be described in detail in the present invention. Of course, other methods can also be selected to fit the curve, which will not be described in detail in the present invention.

[0070] S314 uses the expression function of the curve as the matching function of the channel to be calibrated.

[0071] The expression of the curve fitted in the previous step is the matching function. It can be foreseen that when obtaining the external control electrical signal, it is only necessary to input the reference calibration output value as an independent variable into the matching function to obtain the corresponding external control electrical signal.

[0072] Each channel to be calibrated needs to be fitted with a matching function.

[0073] In the table lookup method, the steps of constructing the matching table include: S321 obtains all external control electrical signals and corresponding output values ​​of the channel to be calibrated under the calibration optical signal according to the set step size.

[0074] Similar to the function fitting method, for each channel to be calibrated, different external control electrical signals are used to intervene when the same calibration optical signal is input, and each external control electrical signal and the corresponding output value are recorded.

[0075] For example, when inputting the same calibration optical signal P ref Under the premise of , set the external control electrical signal I1 to obtain the output value S1 of the channel; set the external control electrical signal I2 to obtain the output value S2 of the channel; set the external control electrical signal I3 to obtain the output value S3 of the channel... and so on.

[0076] In the table lookup method, the adjacent external control electrical signals, such as I1 and I2, I2 and I3, are set step lengths. The steps for obtaining the set step lengths include: S3211 initializes the step length and adjusts the external control electrical signal according to the initialized step length.

[0077] In this embodiment, the external control electrical signal may be a voltage signal or a current signal. In the initial state, the external control electrical signal is adjusted upward from the lower limit of the normal working voltage or current range of the external control electrical signal, or the external control electrical signal is adjusted downward from the upper limit of the normal working voltage or current range of the external control electrical signal.

[0078] When adjusting for the first time, adjust according to the preset initialization step length. The initialization step length can be adaptively adjusted based on the technician's experience.

[0079] Assume that the normal range of the external control electrical signal is [I s ,I e ], the initialization step is I0. Then first record the lower limit of the range I s Then record the output value of I s Output value at +I0.

[0080] S3212 compares the output values ​​of the optical signal modulator under adjacent external control electrical signals; if the absolute value of the difference is less than the difference threshold range, the current step size plus the preset incremental step size is used as the step size for the next adjustment; if the absolute value of the difference is within the difference threshold range, the step size remains unchanged; if the absolute value of the difference is greater than the difference threshold range, the current step size minus the incremental step size is used as the step size for the next adjustment.

[0081] In this embodiment, the step size is not fixed, but will be adaptively adjusted. Compare the outputs under two adjacent external control voltages, for example, compare the outputs S1 and S2 under I1 and I2. If the absolute value of the difference between the two is less than the difference threshold range, it is considered that the difference between S1 and S2 is not only within the tolerance range of the system, but also too dense. In this embodiment, the incremental step size is preset. In the above case, the current complement step size plus the incremental step size is used as the step size of the next step.

[0082] If the absolute value of the difference between the two is within the difference threshold range, it is considered that the difference between S1 and S2 is tolerable by the system. During calibration, even if there is no S1, S2 as output is within the allowable deviation range of the system. In the above case, the current step size remains unchanged.

[0083] Of course, if the absolute value of the difference between the two exceeds the difference threshold range, it is considered that the two are irreplaceable and interpolation is required. Therefore, the step size should be shortened to obtain a more densely distributed output value. In the above case, the current complement step size minus the incremental step size is used as the step size of the next step.

[0084] The above rules are used in analogy until the normal working range of the entire external control electrical signal is traversed. S322 constructs a matching table for the channel to be calibrated according to the acquired external control electrical signal and the corresponding output value.

[0085] After traversing the normal working range of the entire external control electrical signal, a number of external control electrical signals and corresponding outputs are obtained, and the two are compiled into a matching table for the channel to search.

[0086] It is understandable that a matching table needs to be pre-built for each channel.

[0087] S32 adjusts the output light signal from the light source module according to the external control electrical signal through the calibration module.

[0088] It can be understood that the calibration module adjusts the calibration light signal from the light source module in a certain submodule according to the external control electrical signal, so that the input light signal of the corresponding modulator changes, thereby changing the output light signal of the modulator. The output light signal of the adjuster is used as the input light signal of the channel in the calculation module, so that the output of the channel is the reference calibration output value.

[0089] After the calibration is completed, the external control electrical signal of the calibration module submodule corresponding to each channel is solidified when the output reference calibration output value is output, so that the optical signal is pre-adjusted each time it is used in the future. No soft calibration is required in the calculation module.

[0090] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0091] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a computer terminal (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0092] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A calibration method between multiple channels of a photonic computing chip, characterized in that include: Input the same calibration optical signal into the calculation module channel by channel to obtain the calibration output value of each channel; Select a channel as the reference channel, and use the calibration output value of the reference channel as the reference calibration output value; The remaining channels are used as channels to be calibrated; The input optical signal of the calculation module of each channel to be calibrated is adjusted, so that when the light source module inputs the calibration optical signal, the calibration output value of each channel to be calibrated of the calculation module is the reference calibration output value.

2. The method for calibrating multiple channels of a photonic computing chip according to claim 1, characterized in that: The calibration module is used to adjust the input optical signal of the channel to be calibrated; the calibration module includes a plurality of submodules, and the submodules correspond to the modulators in the modulator array one by one; the calibration module is arranged between the light source module and the modulator array, or between the modulator array and the calculation module; In the case where the calibration module is arranged between the light source module and the modulator array, the output optical signal of the light source module is the input optical signal of the calibration module; the calibration module adjusts the output optical signal of the light source module according to the external control electrical signal, thereby forming the input optical signal of the modulator array; In the case where the calibration module is arranged between the modulator array and the calculation module, the output optical signal of the modulator array is the input optical signal of the calibration module; The calibration module adjusts the output optical signal of the modulator array according to the external control electrical signal, thereby forming the input optical signal of the calculation module.

3. A calibration method between multiple channels of a photonic computing chip according to claim 2, characterized in that The step of adjusting the input optical signal of each channel to be calibrated comprises: Input the reference calibration output value as an independent variable into a matching function pre-fitted in the channel to be calibrated to obtain a corresponding external control electrical signal; or, search for an external control electrical signal corresponding to the reference calibration output value in a matching table pre-constructed in the channel to be calibrated; The output light signal from the light source module is adjusted according to the external control electrical signal by the calibration module.

4. The method for calibrating multiple channels of a photonic computing chip according to claim 3, characterized in that The fitting step of the matching function comprises: Recording several external control electrical signals and corresponding output values ​​of the channel to be calibrated under the calibration light signal; The external control electrical signal and the corresponding output value are regarded as discrete points in a two-dimensional coordinate system; Fit discrete points into a curve; The expression function of the curve is used as the matching function of the channel to be calibrated.

5. The method for calibrating multiple channels of a photonic computing chip according to claim 4, characterized in that: The method of fitting discrete points into a curve is one of least squares fitting, polynomial fitting, spline interpolation fitting, and non-parametric fitting.

6. The method for calibrating multiple channels of a photonic computing chip according to claim 4, characterized in that The steps of constructing the matching table include: According to the set step length, all external control electrical signals and corresponding output values ​​of the channel to be calibrated under the calibration light signal are obtained; A matching table of the channel to be calibrated is constructed according to the acquired external control electrical signal and the corresponding output value.

7. A method for calibrating multiple channels of a photonic computing chip according to claim 6, characterized in that The steps to obtain the set step size include: Initializing the step length, and adjusting the external control electrical signal according to the initialization step length; Compare the output values ​​of the optical signal modulator under adjacent external control electrical signals; if the absolute value of the difference is less than the difference threshold range, add the current step size to the preset incremental step size as the step size for the next adjustment; if the absolute value of the difference is within the difference threshold range, the step size remains unchanged; if the absolute value of the difference is greater than the difference threshold range, subtract the incremental step size from the current step size as the step size for the next adjustment.

8. The method for calibrating multiple channels of a photonic computing chip according to claim 7, characterized in that: The external control electrical signal is adjusted upward from the lower limit of the normal operating voltage or current range of the external control electrical signal, or the external control electrical signal is adjusted downward from the upper limit of the normal operating voltage or current range of the external control electrical signal.

9. The method for calibrating multiple channels of a photonic computing chip according to claim 2, characterized in that: The external control electrical signal is voltage or current.

10. The method for calibrating multiple channels of a photonic computing chip according to claim 1, characterized in that: The channel with the smallest calibration output value is used as the reference channel.

Images