Transmission and calculation integrated intelligent optical computing chip architecture and system

Abstract

The invention relates to the technical field of optical computing, in particular to a transmission and computing integrated intelligent optical computing chip architecture and system. The architecture comprises a transmitting module used for receiving an optical input signal corresponding to a target task, loading the optical input signal to an initial carrier, and obtaining and transmitting a to-be-transmitted signal set; the transmission and calculation integrated intelligent optical calculation chip architecture controls the to-be-transmitted signal set to execute optical calculation corresponding to the target task in the propagation process to obtain a to-be-received signal set; and the receiving module is used for receiving the to-be-received signal set and decoding the to-be-received signal set to obtain an optical output signal, and the optical output signal is in direct proportion to the optical input signal. According to the scheme, optical calculation of intelligent tasks can be carried out while signal transmission is realized, and optical communication and calculation with ultrahigh energy efficiency are realized.

Description

Technical Field

[0001] The present disclosure relates to the field of optical computing technology, and in particular to an integrated transmission and computing intelligent optical computing chip architecture and system. Background Art

[0002] With the rapid development of artificial intelligence and scientific computing, the complexity and scale of computing needs are also increasing. However, the existing electronic computing technology is limited by Moore's Law, and its performance is gradually approaching saturation, making it difficult to effectively cope with the increasingly stringent requirements of large-scale complex algorithms on computing power and power consumption. Light has natural advantages such as high throughput and low latency in the propagation process. Optical computing technology that uses photons instead of electrons as computing carriers is seen as the key to breaking the existing computing bottleneck. Summary of the invention

[0003] The present disclosure aims to solve one of the technical problems in the related art at least to some extent.

[0004] To this end, the first purpose of the present disclosure is to propose an integrated transmission and computing intelligent optical computing chip architecture to perform optical computing of intelligent tasks while realizing signal transmission, thereby achieving ultra-high energy-efficiency optical communication and computing.

[0005] The second objective of the present disclosure is to propose an integrated transmission and computing intelligent optical computing chip system.

[0006] To achieve the above objectives, the first embodiment of the present disclosure proposes an integrated transmission and computing intelligent optical computing chip architecture, including: A transmitting module, used to receive an optical input signal corresponding to a target task, and load the optical input signal onto an initial carrier, to obtain and transmit a set of signals to be transmitted; The transmission-computation integrated intelligent optical computing chip architecture controls the set of signals to be transmitted to perform optical calculations corresponding to the target task during the propagation process to obtain a set of signals to be received; The receiving module is used to receive the set of signals to be received, and decode the set of signals to be received to obtain an optical output signal, wherein the optical output signal is proportional to the optical input signal.

[0007] Optionally, the set of signals to be transmitted transmitted by the transmitting module is infinitely propagated through the air to the receiving module.

[0008] Optionally, when the transmission-computation integrated intelligent optical computing chip architecture controls the set of signals to be transmitted to perform the optical computing corresponding to the target task during the propagation process, it is specifically used to: The signal set to be transmitted is controlled to perform optical calculation corresponding to the target task based on the optical propagation matrix corresponding to the target task during the propagation process.

[0009] Optionally, before controlling the set of signals to be transmitted to perform optical calculations corresponding to the target task based on the optical propagation matrix corresponding to the target task during the propagation process, the transmission-computation integrated intelligent optical computing chip architecture is further used to: The initial optical propagation matrix is ​​trained according to the target task to obtain an optical propagation matrix corresponding to the target task.

[0010] Optionally, the transmission-computation integrated intelligent optical computing chip architecture is used to train the initial optical propagation matrix according to the target task, specifically for: According to the target task, initial matrix parameters in the initial optical propagation matrix are trained, wherein the initial matrix parameters include at least one of a phase parameter and an amplitude parameter.

[0011] Optionally, the transmitting module is used to load the optical input signal onto an initial carrier, and when obtaining and transmitting a set of signals to be transmitted, is specifically used to: Loading the optical input signal onto an initial carrier to obtain a loaded carrier; Modulating the loaded carrier to obtain a set of signals to be transmitted; The modulation transmitter is controlled to transmit the set of signals to be transmitted.

[0012] Optionally, when the transmitting module is used to load the optical input signal onto an initial carrier, it is specifically used to: The optical input signal is loaded onto the initial carrier by means of amplitude modulation.

[0013] Optionally, when the transmitting module is used to modulate the loaded carrier, it is specifically used to: The electro-optic modulator and / or the thermal modulator is controlled to modulate the loaded carrier.

[0014] Optionally, when the transmitting module is used to modulate the loaded carrier, it is specifically used to: The loaded carrier is modulated by modulating the phase and / or the amplitude.

[0015] To achieve the above-mentioned purpose, the second aspect of the present disclosure proposes an integrated intelligent optical computing chip system, including: the integrated intelligent optical computing chip architecture shown in any one of the first aspects mentioned above.

[0016] In summary, the integrated transmission and computing intelligent optical computing chip architecture and system provided by the present invention can perform optical computing of intelligent tasks while realizing signal transmission by controlling the set of signals to be transmitted to perform optical computing corresponding to the target task during the propagation process, thereby achieving ultra-high energy-efficient optical communication and computing, realizing the computing architecture of disordered end-side intelligent processors, and breaking through the energy efficiency contradiction of existing high-power processors.

[0017] Additional aspects and advantages of the present disclosure will be given in part in the following description and in part will be obvious from the following description or learned through practice of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above and / or additional aspects and advantages of the present disclosure will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which: Figure 1 A background diagram of an integrated transmission and computing intelligent optical computing chip architecture provided by an embodiment of the present disclosure; Figure 2 A schematic diagram of the structure of an integrated transmission and computing intelligent optical computing chip architecture provided by an embodiment of the present disclosure; Figure 3 A schematic diagram of the operation of an integrated transmission and computing intelligent optical computing chip architecture provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0019] Embodiments of the present disclosure are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present disclosure, and should not be construed as limiting the present disclosure.

[0020] A century has passed since the birth of modern computers, and computing has always been closely linked to processors. With the rapid development of artificial intelligence technology and the widespread popularity of smart devices, a new era is quietly approaching, in which everything will be connected and enhanced by the power of artificial intelligence. The remarkable progress of artificial intelligence depends largely on the continuous innovation of high-performance processors. In order to reduce energy consumption and improve computing power, people have made a lot of efforts. Today, a series of new processors designed specifically for artificial intelligence computing have emerged, including graphics processing units (GPUs), neural network processing units (NPUs), and memory computing chips. However, we still have an urgent need for solid-state dedicated processors, especially in the context of Moore's Law gradually becoming saturated, and building energy-efficient processors has become increasingly difficult. In addition, the surge in the number of smart devices has also brought new challenges, because equipping each device with a high-power processor not only takes up too much space, but also seems impractical in terms of cost. Light has natural physical properties such as high throughput, high speed and high energy efficiency during propagation. Optical computing technology that uses photons instead of electrons as computing carriers is seen as the key to breaking the existing computing bottleneck.

[0021] Using optical technologies, scientists have proposed a series of advanced transmission methods and devices to achieve sub-terahertz bit-rate optical modulation and detection. By utilizing the ability of spatial light propagation to achieve high-channel parallel data transmission, optical transmission is expected to become a high-speed and energy-efficient universal computing method, thereby reducing the time delay and energy overhead associated with dedicated processors. However, existing spatial optical computing systems are large-scale optical systems on optical platforms, which limits many edge-side artificial intelligence computing applications.

[0022] The present disclosure is described in detail below with reference to specific embodiments.

[0023] Figure 1 This is a background diagram of an integrated intelligent optical computing chip architecture provided by the present disclosure. Figure 1 As shown in Figure 1, in a pervasive interconnected system, agents continuously communicate with each other to perform collective tasks. This wireless connectivity spans a variety of scales, including centimeter-scale computer networks for distributed computing, meter-scale robot networks for task collaboration, and kilometer-scale collaboration between drones and ground vehicles. Typically, data is transmitted and processed locally on dedicated processors. However, the separation between data transmission and processor-based data computation results in energy consumption and time delays.

[0024] To address this challenge, the present disclosure introduces a processor-free wireless AI computing method, called Transputing, which utilizes light propagation during wireless transmission as a computing medium, making it possible to perform optical computing during transmission to execute AI tasks.

[0025] For example, Figure 2 This is a schematic diagram of the structure of an integrated intelligent optical computing chip architecture provided by the embodiment of the present disclosure. Figure 2 As shown, the architecture of the integrated intelligent optical computing chip includes: A transmitting module is used to receive an optical input signal corresponding to a target task, load the optical input signal onto an initial carrier, obtain and transmit a set of signals to be transmitted; The transmission and calculation integrated intelligent optical computing chip architecture controls the set of signals to be transmitted to perform optical calculations corresponding to the target task during the propagation process to obtain the set of signals to be received; The receiving module is used to receive a set of signals to be received and decode the set of signals to be received to obtain an optical output signal, wherein the optical output signal is proportional to the optical input signal.

[0026] According to some embodiments, the target task refers to an intelligent task that needs to be performed by the integrated transmission and calculation intelligent optical computing chip architecture. The target task includes but is not limited to various intelligent computing tasks such as classification and matrix calculation.

[0027] It is easy to understand that this architecture controls the set of signals to be transmitted to perform optical calculations corresponding to the target task during the propagation process. Therefore, the transmission can be used to complete the calculation, and the optical calculations of intelligent tasks can be performed while realizing signal transmission, thereby achieving ultra-high energy-efficient optical communication and computing. It can realize the computing architecture of disordered end-side intelligent processors and break through the energy efficiency contradiction of existing high-power processors.

[0028] Optionally, the transmitting module is used to load the optical input signal onto the initial carrier, and when obtaining and transmitting the set of signals to be transmitted, is specifically used to: Loading the optical input signal onto the initial carrier to obtain a loaded carrier; Modulating the loaded carrier to obtain a signal set to be transmitted; The modulated transmitter is controlled to transmit a set of signals to be transmitted.

[0029] According to some embodiments, when the transmitting module is used to load the optical input signal onto the initial carrier, the optical input signal may be loaded onto the initial carrier by modulating the amplitude.

[0030] In some embodiments, when the transmitting module is used to modulate the loaded carrier, the electro-optic modulator and / or the thermal modulator may be controlled to modulate the loaded carrier.

[0031] In some embodiments, the loaded carrier may be modulated by modulating the phase and / or the amplitude.

[0032] When the loaded carrier is modulated by modulating the phase and / or the amplitude, an application specific integrated circuit (ASIC) interface may be used to perform phase modulation on the loaded carrier. For example, the ASIC interface may be driven by a voltage source after being led out through an electrode to achieve phase modulation of the loaded carrier, or the phase modulation of the loaded carrier may be achieved through a phase shifter array.

[0033] Optionally, the set of signals to be transmitted transmitted by the transmitting module is infinitely propagated through the air to the receiving module.

[0034] It should be noted that in the process of the signal set to be transmitted emitted by the transmitting module being infinitely propagated through the air to the receiving module, based on random dispersion disturbance, the signal set to be transmitted can form different transmission modes for each wavelength of the signal to be transmitted during the transmission process. Therefore, it is possible to encode the signal to be transmitted by configuring an independent transmission channel for each wavelength.

[0035] In some embodiments, random dispersion perturbation refers to dispersing the phase of single-wavelength light through the length of the waveguide, the wavelength dimension is unraveled, and the transmission modes of single-wavelength light of similar different wavelengths are no longer similar, that is, the results after diffraction of single-wavelength light of different wavelengths are very different.

[0036] Optionally, when the transmission-computation integrated intelligent optical computing chip architecture controls the set of signals to be transmitted to perform optical computing corresponding to the target task during the propagation process, it is specifically used to: The signal set to be transmitted is controlled to perform optical calculations corresponding to the target task based on the optical propagation matrix corresponding to the target task during the propagation process.

[0037] According to some embodiments, the optical propagation matrix refers to a matrix used when implementing optical diffraction calculations during the propagation process of a set of signals to be transmitted.

[0038] In some embodiments, when the set of signals to be transmitted performs optical calculations corresponding to the target task during the propagation process, it can be modeled as:

[0039] in, P is the set of signals to be transmitted, is the set of signals to be received, U is the optical propagation matrix.

[0040] According to some embodiments, when controlling the transmission of a signal set, before performing optical calculations corresponding to a target task based on an optical propagation matrix corresponding to the target task, the transmission-computation integrated intelligent optical computing chip architecture can train the initial optical propagation matrix according to the target task to obtain the optical propagation matrix corresponding to the target task.

[0041] In some embodiments, when the integrated intelligent optical computing chip architecture trains the initial optical propagation matrix according to the target task, the initial matrix parameters in the initial optical propagation matrix can be trained according to the target task, wherein the initial matrix parameters include at least one of phase parameters and amplitude parameters.

[0042] Take a scenario as an example. Figure 3 The following is a schematic diagram of the working of an integrated intelligent optical computing chip architecture provided by the embodiment of the present disclosure. Figure 3 As shown, first, the optical input signal A ( t ) is loaded onto the initial carrier by amplitude modulation to obtain the loaded carrier M ( t ); Then, the electro-optic modulator M ( t ) signal is modulated into t Rapidly changing set of signals to be transmitted ,in, n is a positive integer; then, the transmitter is modulated Broadcast transmission of signals, During the propagation process, the optical calculation corresponding to the target task is performed to obtain the set of signals to be received. ; Finally, the receiving module receives , and sum them to get the light output signal I ( t ).

[0043] It should be noted that, in theory, if the transmitting module is not Perform any modulation, the receiving module receives will be the same as the input A ( t ) is a relatively strict proportional relationship, so a series of communication theories and methods can be used to decode the received light and restore the original signal. A ( t ). However, if the loading modeling is calculated according to the optical propagation matrix, that is, After modulation, will keep changing, making demodulation impossible. The sum is I ( t), then demodulation can be performed again, the specific reasons are as follows:

[0044]

[0045] in, λ is the wavelength; α is λ Wavelength I ( t )and A ( t ) between the two; in, Each signal is a series of points, which can be described as a vector. Each vector consists of multiple units. i The rth unit in the vector corresponding to the signal is expressed as That is to say, when the receiving module decodes the set of signals to be received, it can decode by summing.

[0046] It is easy to understand that I ( t )and A ( t ) in a proportional form, the communication link can be completed, the signal is demodulated to achieve information transmission, and at the same time distribution, completing the optical diffraction calculation.

[0047] In summary, the architecture provided in this embodiment realizes the simultaneous implementation of optical computing and optical communication without interfering with each other by jointly modeling the transmission channel and the computing channel, so that the architecture can be applied to various existing communication systems to obtain massive additional computing resources.

[0048] In order to implement the above embodiments, the present disclosure also proposes an integrated transmission and computing intelligent optical computing chip system, including: the integrated transmission and computing intelligent optical computing chip architecture provided by the above embodiments.

[0049] It should be noted that the integrated intelligent optical computing chip system, which includes 1,000 integrated intelligent optical computing chip architectures, can obtain an additional 10Tera-MACS of computing power from a 10GHz transmission channel. Due to the use of wireless transmission channels, the additional computing energy overhead of Transputing can be only the driving power of the phase shifter array. Among them, when using low-power microwatt phase shifters, through resonance enhanced modulation and over-coupling configuration, a 15-micron-long active modulation area only requires 16.69 nanowatts of power to achieve 0.067 radians of modulation. Through coupling mode analysis, the phase shifter reduces the power consumption of the phase shifter to 0.067 radians per π phase change through efficient resonant phase modulation. With microwatt-class phase shifters and full reconfigurability, the total computational power can be estimated as NMP, and the energy efficiency of Transputing can be defined as ( ), where NMF refers to the number of operations per second (OPS) of computing power, and F represents the operating frequency of data modulation and detection frequency, which can reach at least tens of GHz.

[0050] In summary, the system provided in this embodiment can achieve an ultra-high energy efficiency of 11.8 Peta OPS / W by reducing the computing burden of a dedicated processor, which exceeds the energy efficiency of a traditional computing system by 3-5 orders of magnitude.

[0051] The collection, storage, use, processing, transmission, provision and disclosure of user personal information involved in this disclosure shall comply with the relevant laws and regulations and shall not violate public order and good morals.

[0052] It should be noted that personal information from users should be collected for legitimate and reasonable purposes and should not be shared or sold outside of these legitimate uses. In addition, such collection / sharing should be carried out after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign the agreement / authorization including authorization of relevant user information before the user uses the function. In addition, any necessary steps should be taken to protect and safeguard access to such personal information data and ensure that others who have access to personal information data comply with its privacy policy and procedures.

[0053] The present disclosure anticipates providing implementation schemes for users to selectively block the use or access of personal information data. That is, the present disclosure anticipates providing hardware and / or software to prevent or block access to such personal information data. Once the personal information data is no longer needed, risks can be minimized by limiting data collection and deleting the data. In addition, when applicable, such personal information is de-identified to protect the privacy of the user.

[0054] The acquisition, transmission, storage, use, and processing of data in the technical solution disclosed in this disclosure are in compliance with the relevant provisions of national laws and regulations.

[0055] It should be noted that in the embodiments of the present disclosure, certain software, components, models and other existing solutions in the industry may be mentioned, which should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0056] In the description of the aforementioned embodiments, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they contradict each other.

[0057] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present disclosure, "plurality" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0058] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present disclosure includes additional implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present disclosure belong.

[0059] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or otherwise processing in a suitable manner if necessary, and then stored in a computer memory.

[0060] It should be understood that the various parts of the present disclosure can be implemented in hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0061] A person skilled in the art may understand that all or part of the steps in the above-mentioned embodiment method may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0062] In addition, each functional unit in each embodiment of the present disclosure may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0063] The storage medium mentioned above may be a read-only memory, a disk or an optical disk, etc. Although the embodiments of the present disclosure have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations of the present disclosure. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present disclosure.

Claims

1. A transmission and computing integrated intelligent optical computing chip architecture, characterized in that: include: A transmitting module, used to receive an optical input signal corresponding to a target task, and load the optical input signal onto an initial carrier, to obtain and transmit a set of signals to be transmitted; The transmission-computation integrated intelligent optical computing chip architecture controls the set of signals to be transmitted to perform optical calculations corresponding to the target task during the propagation process to obtain a set of signals to be received; The receiving module is used to receive the set of signals to be received, and decode the set of signals to be received to obtain an optical output signal, wherein the optical output signal is proportional to the optical input signal.

2. The architecture according to claim 1, characterized in that The set of signals to be transmitted transmitted by the transmitting module is infinitely propagated through the air to the receiving module.

3. The architecture according to claim 1, characterized in that When the transmission-computation integrated intelligent optical computing chip architecture controls the set of signals to be transmitted to perform the optical computing corresponding to the target task during the propagation process, it is specifically used to: The set of signals to be transmitted is controlled to perform optical calculations corresponding to the target task based on the optical propagation matrix corresponding to the target task during the propagation process.

4. The architecture according to claim 3, characterized in that Before controlling the transmission of the signal set to be transmitted to perform the optical calculation corresponding to the target task based on the optical propagation matrix corresponding to the target task during the propagation process, the transmission-computation integrated intelligent optical computing chip architecture is further used to: The initial optical propagation matrix is ​​trained according to the target task to obtain an optical propagation matrix corresponding to the target task.

5. The architecture according to claim 4, characterized in that The transmission-computation integrated intelligent optical computing chip architecture is used to train the initial optical propagation matrix according to the target task, specifically for: According to the target task, initial matrix parameters in the initial optical propagation matrix are trained, wherein the initial matrix parameters include at least one of a phase parameter and an amplitude parameter.

6. The architecture according to claim 1, characterized in that: The transmitting module is used to load the optical input signal onto the initial carrier, obtain and transmit the set of signals to be transmitted, and is specifically used to: Loading the optical input signal onto an initial carrier to obtain a loaded carrier; Modulating the loaded carrier to obtain a set of signals to be transmitted; The modulation transmitter is controlled to transmit the set of signals to be transmitted.

7. The architecture according to claim 6, characterized in that When the transmitting module is used to load the optical input signal onto the initial carrier, it is specifically used to: The optical input signal is loaded onto the initial carrier by means of amplitude modulation.

8. The architecture according to claim 6, characterized in that When the transmitting module is used to modulate the loaded carrier, it is specifically used to: The electro-optic modulator and / or the thermal modulator is controlled to modulate the loaded carrier.

9. The architecture according to claim 6, characterized in that: When the transmitting module is used to modulate the loaded carrier, it is specifically used to: The loaded carrier is modulated by modulating the phase and / or the amplitude.

10. A transmission and computing integrated intelligent optical computing chip system, characterized in that: include: The integrated transmission and computing intelligent optical computing chip architecture as described in any one of claims 1 to 9.

Images