Processing chip, design method and electronic equipment
The processing chip design addresses the delay in asynchronous sub-circuit interactions by using a synchronized asynchronous clock interface circuit with dual beat conversion, enhancing chip performance by reducing signal interaction delay across different clock domains.
Patent Information
- Application Number
- CN202310626996.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-30
- Publication Date
- 2024-12-03
AI Technical Summary
The interaction of control signals between asynchronous sub-circuits in ultra-large-scale integrated circuits using asynchronous clock signals is delayed due to the need for separate beating operations by sending and receiving beat conversion circuits, affecting the performance of the chip.
A processing chip design that includes a first subcircuit, an asynchronous clock interface circuit with a first and second beat conversion circuit, and a trigger circuit, where control signals are simultaneously beaten across different clock domains by the two conversion circuits, reducing interaction delay.
This design reduces the delay in asynchronous interaction of control signals, improving the performance of the processing chip by synchronizing signals across different clock domains simultaneously.
Smart Images

Figure CN119066021A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of chip design technology, and in particular to a processing chip, a design method and an electronic device. Background Art
[0002] In the integrated circuit design of chips, global synchronous clock signals are often used. With the development of integrated circuits, the scale of integrated circuits has gradually increased. Global clock integrated circuits using global synchronous clock signals have problems such as high power consumption and complex communication interactions, which limit the expansion of circuit scale. In order to meet the research and expansion of larger-scale integrated circuits, asynchronous clock signals are gradually being used in ultra-large-scale integrated circuits, that is, two sub-circuits in an ultra-large-scale integrated circuit are based on different clock signals. However, how to asynchronously interact control signals between two asynchronous sub-circuits is a difficult problem.
[0003] An interaction method of an asynchronous clock integrated circuit is as follows: an asynchronous clock interface circuit is set between an asynchronous first subcircuit and a second subcircuit. The asynchronous clock interface includes a sending beat conversion circuit and a receiving beat conversion circuit. The first subcircuit realizes the interaction of control signals based on a handshake mechanism between the sending beat conversion circuit and the receiving beat conversion circuit and the second subcircuit, specifically: the first subcircuit sends a first request handshake signal to the receiving beat conversion circuit, and the first request handshake signal in the same clock domain as the first subcircuit is beat by the receiving beat conversion circuit, so that the first request handshake signal is converted into a second request handshake signal in the same clock domain as the second subcircuit and sent to the second subcircuit. The second subcircuit sends a first feedback handshake signal to the sending beat conversion circuit according to the second request handshake signal. The sending beat conversion circuit beats the first feedback handshake signal in the same clock domain as the second subcircuit, so that the first feedback handshake signal is converted into a second feedback handshake signal in the same clock domain as the first subcircuit and sent to the first subcircuit, thereby completing the interaction of control signals between the first subcircuit and the second subcircuit. However, in this implementation, each handshake interaction of the control signal needs to be beaten by the sending beat conversion circuit and the receiving beat conversion circuit respectively, and the beating operations of the sending beat conversion circuit and the receiving beat conversion circuit will bring a certain delay, thereby affecting the performance of the chip based on the asynchronous clock integrated circuit. Summary of the invention
[0004] The embodiments of the present application provide a processing chip, a design method, and an electronic device, which reduce the delay of asynchronous interaction of control signals of a processing chip based on an asynchronous clock integrated circuit, and improve the performance of the processing chip.
[0005] To achieve the above objectives, the embodiments of the present application adopt the following technical solutions:
[0006] In a first aspect, a processing chip is provided, which includes a first subcircuit, an asynchronous clock interface circuit, and a second subcircuit. The asynchronous clock interface circuit includes a first beat conversion circuit, a first trigger circuit, and a second beat conversion circuit. The first signal end of the first subcircuit is coupled to the trigger input end of the first trigger circuit. The trigger output end of the first trigger circuit is coupled to the input end of the first beat conversion circuit and the input end of the second beat conversion circuit respectively. The output end of the first beat conversion circuit is coupled to the third signal end of the first subcircuit. The output end of the second beat conversion circuit is coupled to the second signal end of the second subcircuit. The first clock output end of the first subcircuit is coupled to the first clock end of the first beat conversion circuit. The second clock output end of the second subcircuit is coupled to the second clock end of the second beat conversion circuit.
[0007] In an embodiment of the present application, the first subcircuit and the second subcircuit are two subcircuits based on different clock signals. When the control signals are interacted between the two subcircuits, it is necessary to beat the control signals entering different clock domains. When the first signal end of the first subcircuit sends a first request handshake signal to the trigger input end of the first trigger circuit, the trigger output end of the first trigger circuit is triggered by the first request handshake signal to output a second request handshake signal. The second request handshake signal is simultaneously output to the first beat conversion circuit coupled to the first subcircuit and the second beat conversion circuit coupled to the second subcircuit through the trigger output end of the first trigger circuit. The two beat conversion circuits start to beat the second request handshake signal at the same time. The first beat conversion circuit beats the second request handshake signal to obtain a second feedback handshake signal output to the third signal end of the first subcircuit. The second beat conversion circuit beats the second request handshake signal to obtain a second request handshake signal in the same clock domain as the second subcircuit and outputs it to the second signal end of the second subcircuit, thereby completing the interaction of control signals between the first subcircuit and the second subcircuit. The embodiment of the present application reduces the interaction delay of control signals between sub-circuits in two different clock domains in an asynchronous clock integrated circuit by simultaneously beating the first beat conversion circuit and the second beat conversion circuit.
[0008] In a possible implementation, the asynchronous clock interface circuit further includes a clock generation circuit. A plurality of input terminals of the clock generation circuit are respectively coupled to a first signal terminal of the first subcircuit, a trigger output terminal of the first trigger circuit, and a fourth signal terminal of the second subcircuit. An output terminal of the clock generation circuit is coupled to a first enable terminal of the first trigger circuit. In an embodiment of the present application, a clock enable signal independent of the first subcircuit and the second subcircuit can be generated by the clock generation circuit. The operation of the first trigger circuit is driven and controlled by the clock enable signal. The clock generation circuit can generate the clock enable signal based on a control signal interacting between the first subcircuit and the second subcircuit and a signal output by the first trigger circuit.
[0009] In a possible implementation, the asynchronous clock interface circuit further includes a data buffer. The first data output terminal of the first subcircuit is coupled to the first data input terminal of the data buffer. The second data output terminal of the data buffer is coupled to the second data input terminal of the second subcircuit. The output terminal of the clock generation circuit is also coupled to the second enable terminal of the data buffer and the first enable terminal of the first trigger circuit. In an embodiment of the present application, the first subcircuit and the second subcircuit can also realize asynchronous interaction of data signals through the data buffer. At the same time, in this process, the data buffer can be controlled to input data signals and / or output data signals through the clock enable signal output by the clock generation circuit.
[0010] In a possible implementation, the asynchronous clock interface circuit further includes a first register and a second register. The first data output terminal of the first subcircuit is coupled to the input terminal of the first register. The output terminal of the first register is coupled to the first data input terminal of the data buffer. The clock output terminal of the first subcircuit is also coupled to the third clock terminal of the first register. The input terminal of the second register is coupled to the second data output terminal of the data buffer. The output terminal of the second register is coupled to the second data input terminal of the second subcircuit. The clock output terminal of the second subcircuit is also coupled to the fourth clock terminal of the second register. In an embodiment of the present application, the first register can be used to beat the transmitted data signal between the first subcircuit and the data buffer. The second register can be used to beat the transmitted data signal between the data buffer and the second subcircuit.
[0011] In a possible implementation, the asynchronous clock interface circuit further includes a voltage holding circuit. The output end of the clock generation circuit is coupled to the second enable end of the data buffer through the voltage holding circuit. In an embodiment of the present application, when the data signal transmitted between the first subcircuit and the second subcircuit is a data signal of multiple bits. In order to ensure that the clock enable signal can drive the data buffer to input the data signal of multiple bits output by the first subcircuit, and / or the data buffer can output the data signal of multiple bits to the second subcircuit. A voltage holding circuit can be set to maintain the driving of the data buffer by the clock generation circuit.
[0012] In a possible implementation, the asynchronous clock interface circuit further includes a pulse width delay circuit. The output end of the clock generation circuit is coupled to the first enable end of the first trigger circuit through the pulse width delay circuit. In the embodiment of the present application, the effective pulse width of the clock enable signal can be adjusted through the pulse width delay circuit to adapt to the operation of the first trigger circuit.
[0013] In a possible implementation, the pulse width delay circuit includes a delay circuit and a first AND gate. The output end of the clock generation circuit is coupled to the first input end of the first AND gate. The output end of the clock generation circuit is also coupled to the second input end of the first AND gate through the delay circuit. In an embodiment of the present application, the clock generation circuit outputs a clock enable signal. The clock enable signal can be delayed by the delay circuit to obtain a delayed clock enable signal. Then, a logical AND operation is performed on the delayed clock enable signal and the undelayed clock enable signal through the first AND gate to obtain a clock enable signal after adjusting the effective pulse width. The first trigger circuit is driven by the clock enable signal after adjusting the effective pulse width.
[0014] In a possible implementation, the clock generation circuit includes an XOR gate, an XOR gate, and a second AND gate. The multiple input terminals of the clock generation circuit include a first input terminal of the XOR gate, a second input terminal of the XOR gate, a first input terminal of the XOR gate, and a second input terminal of the XOR gate. The first input terminal of the XOR gate is coupled to the first signal terminal of the first subcircuit. The second input terminal of the XOR gate is coupled to the trigger output terminal of the first trigger circuit. The first input terminal of the XOR gate is coupled to the fourth signal terminal of the second subcircuit. The second input terminal of the XOR gate is coupled to the trigger output terminal of the first trigger circuit. The output terminal of the XOR gate is coupled to the first input terminal of the second AND gate. The output terminal of the XOR gate is coupled to the second input terminal of the second AND gate. The output terminal of the second AND gate is the output terminal of the clock generation circuit. In the embodiment of the present application, the trigger output terminal of the first trigger circuit outputs a second request handshake signal. The first signal terminal of the first subcircuit outputs a first request handshake signal. The second signal terminal of the second subcircuit outputs a first feedback handshake signal. The first request handshake signal output from the first signal terminal and the second request handshake signal output from the trigger output terminal are logically XORed by the XOR gate. The second request handshake signal output from the trigger output terminal and the first feedback handshake signal output from the fourth signal terminal of the second sub-circuit are subjected to a logic XOR operation through an XOR gate. The operation results of the logic XOR operation and the logic XOR operation are subjected to a logic AND operation through a second AND gate to obtain a clock enable signal.
[0015] In a possible implementation, the asynchronous clock interface circuit further includes a first level logic conversion circuit. The fourth signal terminal of the second sub-circuit is coupled to at least one of the multiple input terminals of the clock generation circuit through the first level logic conversion circuit. In an embodiment of the present application, when the triggering modes between the second sub-circuit and the clock generation circuit are different, the triggering mode conversion can be achieved by setting the first level logic conversion circuit. For example, the conversion from level triggering to edge triggering, or the conversion from edge triggering to level triggering.
[0016] In a possible implementation, the asynchronous clock interface circuit further includes at least one of a second level logic conversion circuit, a third level logic conversion circuit, and a fourth level logic conversion circuit. The first signal end of the first subcircuit is coupled to the trigger input end of the first trigger circuit through the second level logic conversion circuit. The output end of the first beat conversion circuit is coupled to the third signal end of the first subcircuit through the third level logic conversion circuit. The output end of the second beat conversion circuit is coupled to the second signal end of the second subcircuit through the fourth level logic conversion circuit. In an embodiment of the present application, when the triggering mode between the first subcircuit and the first trigger circuit is different, the triggering mode can be converted through the second level logic conversion circuit. When the triggering mode between the first subcircuit and the first beat conversion circuit is different, the triggering mode can be converted through the third level logic conversion circuit. When the triggering mode between the second subcircuit and the second beat conversion circuit is different, the triggering mode can be converted through the fourth level logic conversion circuit.
[0017] In a possible implementation, the first subcircuit and / or the second subcircuit is any one of a storage circuit, a logic control circuit, a routing interface circuit, and a neural network processing circuit. In an embodiment of the present application, the processing chip may be a system on chip, a network on chip, etc. The first subcircuit and / or the second subcircuit in the processing chip may be a subcircuit in a very large scale integrated circuit such as a system on chip, a network on chip, such as a storage circuit, a logic control circuit, a routing interface circuit, a neural network processing circuit, etc.
[0018] In a second aspect, an embodiment of the present application further provides a design method, which is used to design a processing chip as described in the second aspect above. The design method includes: combining a first description file and a second description file to obtain a third description file. The first description file is used to describe the logical function of the first chip; the first chip includes multiple sub-circuits. The multiple sub-circuits are based on the same clock domain; or, some or all of the multiple sub-circuits are based on different clock domains. The second description file is used to describe the logical function of the asynchronous clock interface circuit. The third description file is used to describe the logical function of the processing chip. Timing constraints are performed on the third description file. The logical circuit structure of the processing chip is obtained based on the third description file after timing constraints.
[0019] In an embodiment of the present application, in the EDA design for chip manufacturing. The first description file is a description file based on an existing integrated circuit. The first chip can be based on a synchronous integrated circuit or an asynchronous clock integrated circuit. Then, the first description file for the first chip can be a description file related to multiple sub-circuits in a synchronous integrated circuit or a description file related to multiple sub-circuits in an asynchronous clock integrated circuit. The second description file can be generated according to the logic function of the asynchronous clock interface circuit. According to the combination of the first description file and the second description file, a description file about the logic function of the processing chip, i.e., a third description file, is obtained. After the third description file is subjected to timing constraints to ensure that the logic function of the third description file operates reasonably, the logic circuit structure of the processing chip is designed based on the third description file after the timing constraints to obtain a design layout. The processing chip can be manufactured according to the design layout later. By improving the third description file based on the first description file of the first chip and combining the second description file to obtain the third description file, the design process of the processing chip can be greatly shortened, and the design difficulty of the processing chip can be reduced.
[0020] In some possible implementations, the design method further includes: before obtaining the logic circuit structure of the processing chip based on the third description file after the timing constraints, modifying the third description file after the timing constraints, and performing design constraints on the modified third description file. In an embodiment of the present application, the third description file can be adaptively modified according to different actual application scenarios. At the same time, in the process of timing constraints, the content that does not meet the relevant timing constraints can also be modified. Then the modified file is subjected to design constraints. The reliability and applicability of the processing chip can be increased through the embodiment of the present application.
[0021] In a third aspect, an embodiment of the present application further provides an electronic device, the electronic device comprising a circuit board and a processing chip as described in the first aspect above. The processing chip is disposed on the circuit board.
[0022] Regarding the technical principles and beneficial effects of the second and third aspects, reference may be made to the relevant description of the first aspect, which will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] Figure 1 A schematic diagram of a design flow of a synchronous integrated circuit based on an EDA design tool provided in an embodiment of the present application;
[0024] Figure 2 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application;
[0025] Figure 3 A schematic diagram of the structure of a processing chip provided in an embodiment of the present application;
[0026] Figure 4 A schematic diagram of the structure of a first asynchronous clock interface circuit provided in an embodiment of the present application;
[0027] Figure 5 A schematic diagram of the structure of a second asynchronous clock interface circuit provided in an embodiment of the present application;
[0028] Figure 6 A schematic diagram of the structure of a third asynchronous clock interface circuit provided in an embodiment of the present application Figure 1 ;
[0029] Figure 7 A schematic diagram of the structure of another third asynchronous clock interface circuit provided in an embodiment of the present application Figure 2 ;
[0030] Figure 8 A schematic diagram of the structure of another third asynchronous clock interface circuit provided in the embodiment of the present application Figure 3 ;
[0031] Fig. 9 A schematic diagram of the structure of another third asynchronous clock interface circuit provided in the embodiment of the present application Figure 4 ;
[0032] Fig.10 A schematic diagram of the structure of another third asynchronous clock interface circuit provided in the embodiment of the present application Figure 5 ;
[0033] Fig.11 A schematic diagram of the structure of another third asynchronous clock interface circuit provided in the embodiment of the present application Figure 6 ;
[0034] Fig.12 A schematic diagram of the structure of another third asynchronous clock interface circuit provided in the embodiment of the present application Figure 7 ;
[0035] Fig.13 A schematic diagram of a level logic conversion circuit for converting edge triggering into level triggering provided in an embodiment of the present application;
[0036] Fig.14 A schematic diagram of the structure of a level logic conversion circuit for converting a level trigger into an edge trigger provided in an embodiment of the present application;
[0037] Fig.15 A schematic diagram of the structure of another third asynchronous clock interface circuit provided in the embodiment of the present application Figure 8 ;
[0038] Fig.16 A schematic diagram of the structure of a processing chip of a system-on-chip architecture provided in an embodiment of the present application;
[0039] Fig.17 A schematic diagram of the structure of a processing chip for network-on-chip processing provided in an embodiment of the present application;
[0040] Fig.18 A schematic diagram of a flow chart of a design method for an asynchronous clock integrated circuit provided in an embodiment of the present application;
[0041] Fig.19 A schematic diagram of clock signal input of a synchronous integrated circuit described in a first description file provided in an embodiment of the present application;
[0042] Fig. 20 A method based on the embodiment of the present application is provided Fig.19 A flow chart of a design method for a very large-scale asynchronous clock integrated circuit in the scenario shown;
[0043] Fig.21 A schematic diagram of timing constraints of various circuit signals in the design of an asynchronous clock integrated circuit provided in an embodiment of the present application;
[0044] Fig. 22 A schematic diagram of clock signal input of an asynchronous clock integrated circuit described in a first description file provided in an embodiment of the present application;
[0045] Fig.23 A method based on the embodiment of the present application is provided Fig. 22 A flow chart of a design method for a very large-scale asynchronous clock integrated circuit in the scenario shown;
[0046] Fig.24 A signal diagram of a timing constraint check provided in an embodiment of the present application;
[0047] Fig.25 A schematic diagram of a checking module for timing constraint checking provided in an embodiment of the present application;
[0048] Fig.26 A schematic diagram of a maximum delay check module provided in an embodiment of the present application. DETAILED DESCRIPTION
[0049] It should be noted that the terms "first", "second", etc. involved in the embodiments of the present application are only used to distinguish features of the same type and cannot be understood as indicating relative importance, quantity, order, etc.
[0050] The terms "exemplary" or "for example" and the like in the embodiments of the present application are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the terms "exemplary" or "for example" is intended to present the related concepts in a specific way.
[0051] The terms "coupling" and "connection" involved in the embodiments of the present application should be understood in a broad sense. For example, they may refer to a direct physical connection, or an indirect connection achieved through electronic devices, such as a connection achieved through resistors, inductors, capacitors or other electronic devices.
[0052] First, some basic concepts involved in the embodiments of the present application are explained:
[0053] Electronic design automation (EDA) refers to the use of computer-aided design (CAD) software to complete the functional design, synthesis, verification, physical design (including layout, routing, layout, design rule checking, etc.) of very large scale integration circuit (VLSI) chips.
[0054] Intellectual property (IP) cores play a very important role in EDA. They can be understood as pre-designed circuit function modules in integrated circuits. Specifically, IP cores can be divided into soft IP cores, solid IP cores, and hard IP cores. Soft IP cores are software function blocks described in some hardware description languages, but do not involve what specific circuit components are used to implement these functions. To a certain extent, it makes the subsequent processes unable to adapt to the overall design, thus requiring a certain degree of soft IP correction. Hard IP cores are used to provide the final stage product of the design, that is, the mask. The mask is represented by a netlist that has been completely laid out and routed. Solid IP cores are a combination of soft IP cores and hard IP cores. Soft IP cores are usually provided in an encrypted form, so that the actual register transfer level (RTL) description is invisible to the user, but the layout and routing are flexible. In these encrypted soft IP cores, if the core is parameterized, the user can easily adjust the parameters through the header file or graphical user interface (GUI). For cores that have strict timing requirements, specific signals can be pre-routed or specific routing resources can be allocated to meet timing requirements. These cores can be classified as fixed IP cores, which may affect the overall design containing the core because the core is a pre-designed code module.
[0055] Very large scale integration circuit (VLSI) is an integrated circuit that combines a large number of transistors into a single chip, and its integration is greater than that of LSI. VLSI design is usually carried out by electronic design automation (EDA). The number of integrated transistors varies in different standards. With the development of complex semiconductors and communication technologies, the research and development of integrated circuits has gradually unfolded. The rapid improvement of chip integration has enabled more IP cores to be integrated on a single chip, thus forming a system on chip (SoC). SoC may integrate homogeneous or heterogeneous IP cores such as digital signal processor (DSP), micro controller unit (MCU), memory, analog to digital converter (ADC), digital to analog converter (DAC), universal asynchronous receiver / transmitter (UART), and software and hardware modules that implement special customized functions. Therefore, data communication and exchange between different IP cores in the chip has become one of the key technologies in SoC chip design. The existing communication structures mainly include point-to-point communication, bus-based communication, cross-switch-based communication and network-based communication. However, with the improvement of the integration of integrated circuits, the number of IP cores in integrated circuits has also increased. In the above communication methods, there are problems such as design layout difficulties, communication interaction conflicts, increased communication power consumption, poor scalability, reduced communication reliability, and reduced throughput. In order to solve the respective problems existing in the above communication methods in SoC, the industry has proposed a design architecture of network on chip (NoC).
[0056] Metastability refers to the inability of a trigger to reach a verifiable state (such as 0 level or 1 level) within a specified period of time. When a trigger enters a metastable state, it is impossible to predict the output level of the circuit unit, nor when the output will stabilize at a correct level. During this stable period, the trigger outputs some intermediate level, or may be in an oscillating state, and this useless output level can be cascaded along the various triggers on the signal path.
[0057] Global synchronous clock means that all circuits in an integrated circuit implement circuit design based on the same clock signal. This integrated circuit based on global synchronous clock can also be called synchronous clock integrated circuit. Synchronous clock integrated circuit is based on two basic assumptions: the signals in the synchronous clock integrated circuit are all in binary form; and all components in the synchronous clock integrated circuit share a common discrete timing, which is defined by the global synchronous clock signal distributed throughout the synchronous clock integrated circuit. In the synchronous clock integrated circuit, data processing is completed by combinational logic. Combinational logic has no storage function, and data signals are all in the form of linear variables. At this time, the change of the input state of the synchronous clock integrated circuit is immediately reflected in the output signal, regardless of the previous state of the synchronous clock integrated circuit. The storage of data is the responsibility of sequential logic, which has a memory function. The output at any time is jointly determined by the current input and the previous state of the synchronous clock integrated circuit. The global synchronous clock signal is used to standardize the timing of the synchronous clock integrated circuit to ensure the correct operation of the synchronous clock integrated circuit. Since all input signals in the synchronous clock integrated circuit only contain two logical states, 0 and 1, they are controlled by a global synchronous clock signal. This control method can effectively avoid the metastable state of the synchronous clock integrated circuit, and filter out the glitches caused by combinational logic competition and risk, so as to hide the complex intermediate states in the integrated circuit, as long as it meets the requirement of "establishing a certain hold time". This method greatly simplifies the design and verification of synchronous circuits. Therefore, in the long-term development, the design of synchronous clock integrated circuits based on global synchronous clock signals has established a complete set of top-down design processes and supporting EDA design tools, allowing designers to simply and effectively design the required ultra-large-scale integrated circuit chips. In the vast majority of integrated circuit designs in the past, global synchronous clock signals were used to design synchronous clock integrated circuits. The synchronous clock integrated circuit design method has always occupied a mainstream position in the industry. Such as Figure 1As shown, it is a flow chart of the design of synchronous clock integrated circuit based on EDA design tools, and the specific process includes the following steps: S1, input synchronization description file. The synchronization description file is an RTL description file used to describe the logic function of the synchronous clock integrated circuit. S2, modify and adjust the synchronization description file after timing constraint. S3, perform design constraints on the modified and adjusted synchronization description file. S4, perform synthesis processing on the synchronization description file after design constraints to obtain the synthesis structure of the corresponding logic gate circuit. S5, map the synthesis structure of the logic gate circuit to the network table (netlist). S6, perform back-end design (physical design) layout according to the mapped network table. S7, perform static timing analysis (static timing analysis, STA) on the circuit structure after the back-end design. S8, obtain the final design layout of the synchronous clock integrated circuit after completing the static timing analysis.
[0058] With the increase in the functional integration and number of devices of integrated circuits, more and more problems will arise when using global synchronous clock signals in very large-scale integrated circuits. For example: First, when using the control method of global synchronous clock signals, the clock cycle will be determined by the critical path with the longest delay, and the performance of the chip will be the worst performance under the constraints of process, temperature, voltage and other factors. Secondly, the use of global synchronous clock signals in very large-scale integrated circuits requires higher and higher design complexity of clock circuits. Due to the continuous improvement of chip integration, the design of clock circuits without clock skew is becoming more and more difficult, and factors such as thermal noise, power supply noise and crosstalk in the chip will also cause clock jitter, which affects the correctness of the timing logic. Thirdly, the global synchronous clock signal controls all timing elements in the chip. Even if the transmitted data signal does not change, the clock signal will generate unnecessary power consumption when it is flipped. In addition, with the increase in integration and the increase in chip area, there are long wires in the clock distribution circuit. The clock signal will generate additional power consumption for driving the long wires, especially leakage current power consumption. Therefore, in the design of synchronous clock integrated circuits, this power consumption of the clock signal will account for a large part of the power consumption waste of the chip. For example, in a synchronous NoC using a global synchronous clock signal, since the NoC contains homogeneous or heterogeneous IP cores, each IP core may have different operating frequencies, electrical characteristics, and timing constraints. If the control method of the global synchronous clock signal is adopted, in the process of designing a synchronous clock integrated circuit, once a new IP core is added, the global synchronous clock network needs to be adjusted or even redesigned, which will increase the chip design cycle and complexity. For this reason, some researchers have proposed to design an asynchronous clock integrated circuit using an asynchronous clock signal. A handshaking mechanism can be used in an asynchronous clock integrated circuit to replace the global synchronous clock signal. Under the control of the handshake signal, data streams are transmitted between subcircuits with two different clock signals in a pipeline manner. The handshake signal includes a request signal and a feedback signal. The specific handshake method is: the first subcircuit at the sending end sends a request signal and data, and the second subcircuit at the receiving end feeds back a response signal to the sending end after receiving the data to start the next data cycle. The asynchronous clock integrated circuit is data-driven, and the next data processing can be started after the current data cycle ends, without having to wait for a fixed clock cycle like a synchronous integrated circuit. Due to the removal of the global clock, asynchronous circuits have the following advantages over synchronous circuits: (1) The global timing is eliminated, the circuit is easy to modularize, and has greater design flexibility; (2) The global clock is removed to avoid problems such as clock skew; (3) It has low power consumption characteristics; (4) It has potential high performance characteristics; (5) It is not sensitive to signal delays and has stronger process adaptability; (6) It has better electromagnetic compatibility.
[0059] Although asynchronous clock integrated circuits have many advantages, they also have some problems compared with synchronous clock integrated circuits. Asynchronous clock integrated circuits do not have a global synchronous clock, so there is no shielding effect of synchronous pulses, and it is impossible to filter out glitches caused by combinational logic competition and risk. In addition, the complex intermediate states (i.e., metastable states) that may appear in synchronous clock integrated circuits show an exponential growth, which is very likely to cause great harm to the circuit, and need to be discovered and removed as much as possible to ensure the correct function of the circuit. Therefore, when designing circuits for very large-scale integrated circuits under asynchronous integration, it is necessary to face more serious state failure problems than synchronous clock integrated circuits during the design verification process. For complex systems, these failure problems in design verification are difficult to completely solve manually. In the design of asynchronous clock integrated circuits, the industry has the following design methods: Design method 1, designing and synthesizing asynchronous bundled data (BD) controller circuits based on signal transition graphs (STG). However, it is extremely difficult to draw a perfect STG when the integration scale is large. Design method 2: Based on physical design tools, the layout and routing of asynchronous clock integrated circuits are optimized and analyzed to design asynchronous clock integrated circuits. However, this method requires manual design and adjustment of the layout and routing of circuits, which is also difficult to apply in the design of ultra-large-scale integrated circuits. Design method 3: Based on open source EDA design tools for asynchronous clock integrated circuits, syntax-oriented asynchronous clock integrated circuits are described and synthesized through EDA design tools. However, in the development of integrated circuits over the years, the industry has focused on the research and design of synchronous clock integrated circuits, and open source EDA design tools are all developed for synchronous circuit systems. The development of asynchronous clock integrated circuits has stagnated for many years, and it does not have a complete design method, process and complete EDA design tools like the design of synchronous clock integrated circuits. The imperfection of the design process, the lack of design tools and the complexity of verification work have increased the difficulty of asynchronous clock integrated circuit design to a certain extent.
[0060] The present application embodiment provides an electronic device, such as Figure 2 As shown, the electronic device 1000 includes a circuit board and a processing chip 100. The processing chip 100 is arranged on the circuit board. The processing chip 100 includes a plurality of sub-circuits. The plurality of sub-circuits include a first sub-circuit and a second sub-circuit, wherein the first sub-circuit and the second sub-circuit are sub-circuits based on different clock signals. Figure 3 As shown, the processing chip 100 further includes an asynchronous clock interface circuit 20. The first sub-circuit 10 and the second sub-circuit 30 are coupled respectively through the asynchronous clock interface circuit 20 to implement asynchronous interaction of data signals and / or asynchronous interaction of control signals.
[0061] In some possible implementations, such as Figure 3 The asynchronous clock interface circuit 20 in the illustrated embodiment is a first asynchronous clock interface circuit based on an asynchronous first input first output (FIFO) memory. The first asynchronous clock interface circuit is used to implement asynchronous interaction of data signals. Figure 4 As shown, the first asynchronous clock interface circuit 20A includes a write controller 21A, a data transmission memory 22A, a read controller 23A, a first Gray code conversion circuit 24A, a second Gray code conversion circuit 25A, a first synchronization circuit 26A, a second synchronization circuit 27A, a full state circuit 28A and an empty state circuit 29A. The data transmission memory 22A is a FIFO memory. Among them, the first data output terminal data1 of the first sub-circuit 10 is coupled with the data input terminal wdata of the data transmission memory 22A. The data output terminal rdata of the data transmission memory 22A is coupled with the second data input terminal data4 of the second sub-circuit 30. The first signal terminal X1 of the first sub-circuit 10 is coupled with the first write enable terminal wen of the write controller 21A. The second write enable terminal ram_wen1 of the write controller 21A is coupled with the storage write enable terminal ram_wen2 of the data transmission memory 22A. The second signal terminal X2 of the second sub-circuit 30 is coupled with the first read enable terminal ren of the read controller 23A. The second read enable terminal ram_ren1 of the read controller 23A is coupled to the storage read enable terminal ram_ren2 of the data transmission memory 22A. The write pointer input terminal w_ptr of the write controller 21A is coupled to the input terminal of the first Gray code conversion circuit 24A and the first operation address terminal waddr of the data transmission memory 22A respectively. The read pointer input terminal r_ptr of the read controller 23A is coupled to the input terminal of the second Gray code conversion circuit 25A and the second operation address terminal raddr of the data transmission memory 22A respectively. The output terminal of the first Gray code conversion circuit 24A is coupled to the input terminal of the empty state circuit 29A through the first synchronization circuit 26A. The output terminal of the empty state circuit 29A is coupled to the empty state terminal empty of the read controller 23A. The output terminal of the second Gray code conversion circuit 25A is coupled to the input terminal of the full state circuit 28A through the second synchronization circuit 27A. The output terminal of the full state circuit 28A is coupled to the full state terminal full of the write controller 21A.
[0062] In such Figure 4In the illustrated embodiment, the first clock output terminal clk1 of the first subcircuit 10 is coupled to the first read clock control terminal wclk1 of the write controller 21A and the second read clock control terminal wclk2 of the data transmission memory 22A. The clock output terminal of the second subcircuit 30 is coupled to the first write clock control terminal rclk1 of the read controller 23A and the second write clock control terminal rclk2 of the data transmission memory 22A. The first subcircuit 10 and the second subcircuit 30 are based on different clock signals. The binary code is converted into the Gray code by the first Gray code conversion circuit 24A and the second Gray code conversion circuit 25A, and the Gray code is used to synchronize the pointer of the asynchronous FIFO buffer. In the embodiment of the present application, it is necessary to set the data transmission memory 22A as the FIFO memory, and the data transmission memory 22A requires a larger chip size to store data. At the same time, it is also necessary to set up additional circuits such as a write controller 21A, a read controller 23A, a first Gray code conversion circuit 24A, a second Gray code conversion circuit 25A, a first synchronization circuit 26A, a second synchronization circuit 27A, a full state circuit 28A, and an empty state circuit 29A to form a read-write control circuit, and the additional read-write control area is also very large. In addition, in order to solve the problem of crossing clock domains between the first sub-circuit 10 and the second sub-circuit 30, the first Gray code conversion circuit 24A, the second Gray code conversion circuit 25A, the first synchronization circuit 26A, the second synchronization circuit 27A, etc. are used for pointer synchronization, and pointer synchronization will limit the data transmission performance of the data transmission memory 22A, thereby causing area and performance losses.
[0063] In some possible implementations, such as Figure 3 The asynchronous clock interface circuit 20 in the embodiment shown is a second asynchronous clock interface circuit. The asynchronous interaction of control signals is realized by the second asynchronous clock interface circuit. Figure 5 As shown, the second asynchronous clock interface circuit 20B includes a sending beat conversion circuit 21B and a receiving beat conversion circuit 22B. The first signal terminal X1 of the first sub-circuit 10 is coupled to the input terminal of the receiving beat conversion circuit 22B. The output terminal of the receiving beat conversion circuit 22B is coupled to the second signal terminal X2 of the second sub-circuit 30. The fourth signal terminal X4 of the second sub-circuit 30 is coupled to the input terminal of the sending beat conversion circuit 21B. The output terminal of the sending beat conversion circuit 21B is coupled to the third signal terminal X3 of the first sub-circuit 10. The first clock output terminal clk1 of the first sub-circuit 10 is coupled to the enable terminal of the sending beat conversion circuit 21B. The second clock output terminal clk2 of the second sub-circuit 30 is coupled to the enable terminal of the receiving beat conversion circuit 22B.
[0064] In this application Figure 5In the illustrated embodiment, the first subcircuit 10 is based on the first clock signal, and the second subcircuit 30 is based on the second clock signal. When the control signal as the handshake signal is interacted between the two subcircuits, in order to avoid the metastable problem caused by the direct interaction of the control signals under different clock signals. When the control signal enters different clock domains, it is necessary to set the sending beat conversion circuit 21B and the receiving beat conversion circuit 22B to perform the beat operation under different clock signals. The first subcircuit 10 sends a request handshake signal to the input end of the receiving beat conversion circuit 22B through the first signal terminal X1, and the request handshake signal is based on the first clock signal of the first subcircuit 10. The receiving beat conversion circuit 22B works based on the second clock signal of the second subcircuit 30. After the receiving beat conversion circuit 22B inputs the request handshake signal, it beats the request handshake signal for N beats based on the second clock signal of the second subcircuit 30 to obtain a request handshake signal based on the second clock signal. Then the receiving beat conversion circuit 22B sends the request handshake signal based on the second clock signal to the second signal terminal X2 of the second subcircuit 30. After inputting the request handshake signal based on the second clock signal, the second subcircuit 30 outputs a feedback handshake signal, which is based on the second clock signal. The second subcircuit 30 sends a feedback handshake signal based on the second clock signal to the input end of the sending beat conversion circuit 21B through the fourth signal terminal X4. The sending beat conversion circuit 21B beats the feedback handshake signal based on the second clock signal for M beats based on the first clock signal to obtain a feedback handshake signal based on the first clock signal. Then the feedback handshake signal based on the first distorted signal is sent to the third signal terminal X3 of the first subcircuit 10. The process from the first subcircuit 10 outputting the request handshake signal to the first subcircuit 10 receiving the feedback handshake signal is a handshake between the first subcircuit 10 and the second subcircuit 30. In this process, the beating operation of the sending beat conversion circuit 21B and the receiving beat conversion circuit 22B will cause a delay. Taking N and M as 2 as an example, there is a delay of 4 beats in a handshake operation. If the values of N and M are larger, the delay of a handshake operation will also be larger. This will make the interactive performance under the asynchronous clock integrated circuit significantly lower than the interactive performance under the synchronous integrated circuit.
[0065] In some possible implementations, in order to reduce the delay of handshake operation in asynchronous clock integrated circuits. Figure 3 The asynchronous clock interface circuit 20 in the embodiment shown is a third asynchronous clock interface circuit. The asynchronous interaction of control signals is realized by the third asynchronous clock interface circuit. Figure 6As shown, the third asynchronous clock interface circuit 20C includes a sending interface, an asynchronous conversion interface and a receiving interface. The sending interface includes a first beat conversion circuit 21C. The receiving interface includes a second beat conversion circuit 22C. The asynchronous conversion interface includes a first trigger circuit 23C. The first signal terminal X1 of the first sub-circuit 10 is coupled to the trigger input terminal D1 of the first trigger circuit 23C. The trigger output terminal Q1 of the first trigger circuit 23C is coupled to the input terminal of the first beat conversion circuit 21C and the input terminal of the second beat conversion circuit 22C respectively. The output terminal of the first beat conversion circuit 21C is coupled to the third signal terminal X3 of the first sub-circuit 10. The output terminal of the second beat conversion circuit 22C is coupled to the second signal terminal X2 of the second sub-circuit 30. The first clock output terminal clk1 of the first sub-circuit 10 is coupled to the first clock terminal CK1 of the first beat conversion circuit 21C. The second clock output terminal clk2 of the second sub-circuit 30 is coupled to the second clock terminal CK2 of the second beat conversion circuit 22C.
[0066] In this application Figure 6 In the embodiment shown, the first signal terminal X1 of the first sub-circuit 10 outputs a first request handshake signal to the trigger input terminal D1 of the first trigger circuit 23C. After the first trigger circuit 23C receives the first request handshake signal, the first request handshake signal triggers the first trigger circuit 23C to send a second request handshake signal to the input terminal of the first beat conversion circuit 21C and the input terminal of the second beat conversion circuit 22C through the trigger output terminal Q1. The second beat conversion circuit 22C obtains a second request handshake signal based on the second clock signal after tapping the second request handshake signal for M beats, and sends the second clock handshake signal to the second signal terminal X2 of the second sub-circuit 30. After the first beat conversion circuit 21C taps the second request handshake signal for N beats, it obtains a second feedback handshake signal based on the first clock signal, and sends the second feedback handshake signal to the third signal terminal X3 of the first sub-circuit 10. In the embodiment of the present application, the first trigger circuit 23C sends the second request handshake signal to the first beat conversion circuit 21C and the second beat conversion circuit 22C at the same time for beat tapping. Figure 5 In the embodiment shown, the sending beat conversion circuit 21B and the receiving beat conversion circuit 22B perform beats respectively, and the beat delay is N+M. Figure 6 In the embodiment shown, the first beat conversion circuit 21C and the second beat conversion circuit 22C perform beat tapping at the same time, and the tapping delay at this time is N or M (the larger delay is selected between N and M). Figure 5 The embodiment shown, as Figure 6 The illustrated embodiment can greatly reduce the delay caused by the beat rhythm, thereby improving the interactive performance of the asynchronous clock integrated circuit.
[0067] Exemplarily, the first trigger circuit 23C may be a D flip-flop, a latch, or the like.
[0068] In some possible implementations, such as Figure 7 As shown, the first beat conversion circuit 21C includes a plurality of first metronomes CB1 connected in series. The second beat conversion circuit 22C includes a plurality of second metronomes CB2. The first clock output terminal clk1 of the first subcircuit 10 is coupled to the first clock terminal CK1 of the plurality of first metronomes CB1 respectively. The second clock output terminal clk2 of the second subcircuit 30 is coupled to the second clock terminal CK2 of the plurality of second metronomes CB2 respectively. Optionally, the first metronome CB1 and / or the second metronome CB1 can be a D flip-flop or the like. Optionally, the number of the plurality of first metronomes CB1 and the number of the plurality of second metronomes CB2 are equal or unequal.
[0069] In some possible implementations, the operation of the first trigger circuit 23C may be controlled by a third clock signal that is independent of the first clock signal and the second clock signal. Figure 7 As shown, the asynchronous conversion interface of the third asynchronous clock interface circuit 20C also includes a clock generation circuit 24C. Multiple input terminals of the clock generation circuit 24C are respectively coupled to the first signal terminal X1 of the first sub-circuit 10, the trigger output terminal Q1 of the first trigger circuit 23C, and the fourth signal terminal X4 of the second sub-circuit 30. The output terminal of the clock generation circuit 24C is coupled to the first enable terminal EN1 of the first trigger circuit 23C. In the embodiment of the present application, the first signal terminal X1 of the first sub-circuit 10 outputs a first request handshake signal based on the first clock signal. The trigger output terminal Q1 of the first trigger circuit 23C outputs a second request handshake signal. The fourth signal terminal X4 of the second sub-circuit 30 outputs a first feedback handshake signal based on the second clock signal. The clock generation circuit 24C generates a clock enable signal based on the third clock signal according to the first request handshake signal, the second request handshake signal, and the first feedback handshake signal. The clock generation circuit 24C outputs a clock enable signal to the first enable terminal EN1 of the first trigger circuit 23C. The second request handshake signal output by the first trigger circuit 23C is based on the clock enable signal.
[0070] In some possible implementations, asynchronous interaction of data signals may also be performed based on the third asynchronous clock interface circuit 20C. Figure 7As shown, the asynchronous conversion interface of the third asynchronous clock interface circuit 20C also includes a data buffer 25C. The first data output terminal data1 of the first sub-circuit 10 is coupled to the first data input terminal data2 of the data buffer 25C. The second data output terminal data3 of the data buffer 25C is coupled to the second data input terminal data4 of the second sub-circuit 30. The output terminals of the clock generation circuit 24C are respectively coupled to the second enable terminal EN2 of the data buffer 25C. In an embodiment of the present application, the clock generation circuit 24C also outputs a clock enable signal to the second enable terminal EN2 of the data buffer 25C. Based on the clock enable signal, the data buffer 25C is controlled to output the latched data signal to the second data input terminal data4 of the second sub-circuit 30 to complete the asynchronous interactive transmission of the data signal between the first sub-circuit 10 and the second sub-circuit 30. This application is as shown in FIG. Figure 7 In the embodiment shown, compared with Figure 4 In the embodiment shown in the figure, there is no need to set up a large FIFO storage device, nor is there any need to set up an additional large area and synchronous control circuit. Figure 6 The asynchronous interaction circuit of the control signal shown can control the asynchronous interaction of the data signal. Figure 6 The asynchronous interaction of the control signals shown has a small delay, which can improve the Figure 7 The illustrated embodiment provides interactive capabilities for asynchronous interaction of data information.
[0071] Exemplarily, the data buffer 25C may be a register, a latch, etc.
[0072] In some possible implementations, such as Figure 8As shown, the clock generation circuit 24C includes an XOR gate XOR1, an XNOR gate XNOR1, and a second AND gate AND2. The multiple input terminals of the clock generation circuit 24C include a first input terminal of the XOR gate XOR1, a second input terminal of the XOR gate XOR1, a first input terminal of the XNOR gate XNOR1, and a second input terminal of the XNOR gate XNOR1. The first input terminal of the XOR gate XOR1 is coupled to the first signal terminal X1 of the first sub-circuit 10. The second input terminal of the XOR gate XOR1 is coupled to the trigger output terminal Q1 of the first trigger circuit 23C. The first input terminal of the XNOR gate XNOR1 is coupled to the fourth signal terminal X4 of the second sub-circuit 30. The second input terminal of the XNOR gate XNOR1 is coupled to the trigger output terminal Q1 of the first trigger circuit 23C. The output terminal of the XOR gate XOR1 is coupled to the first input terminal of the second AND gate AND2. The output terminal of the XNOR gate XNOR1 is coupled to the second input terminal of the second AND gate AND2. The output terminal of the second AND gate AND2 is the output terminal of the clock generation circuit 24C. In an embodiment of the present application, an XOR logic processing is performed on the first request handshake signal based on the first clock signal and the second request handshake signal based on the clock enable signal through an XOR gate XOR1, specifically: when the first request handshake signal and the second request handshake signal are at different levels, the XOR gate XOR1 outputs a high level. An XNOR gate XNOR1 performs an XOR-N (i.e., XOR) logic processing on the first feedback handshake signal based on the second clock signal and the second request handshake signal based on the clock enable signal, specifically: when the first feedback handshake signal and the second request handshake signal are at different levels, the XNOR gate XNOR1 outputs a high level. Then, a logic AND operation is performed on the level signal output by the XOR gate XOR1 and the level signal output by the XNOR gate XNOR1 through the second AND gate AND2, and the clock enable signal can be obtained.
[0073] In some possible implementations, such as Figure 8 As shown, the asynchronous conversion interface of the third asynchronous clock interface circuit 20C also includes a pulse width delay circuit 26C. The output end of the clock generation circuit 24C is coupled to the first enable end EN1 of the first trigger circuit 23C through the pulse width delay circuit 26C. In the embodiment of the present application, in some actual application scenarios, the pulse width of the clock enable signal required by the first trigger circuit 23C is relatively large. The pulse width of the clock enable signal required by the first trigger circuit 23C can be increased by setting the pulse width delay circuit 26C to meet the needs of the corresponding scenario.
[0074] For example, Fig. 9As shown, the pulse width delay circuit 26C includes a delay circuit DL1 and a first AND gate AND1. The output end of the clock generation circuit 24C is coupled to the first input end of the first AND gate AND1. The output end of the clock generation circuit 24C is also coupled to the second input end of the first AND gate AND1 through the delay circuit DL1. In the embodiment of the present application, the input clock enable signal is delayed by the delay circuit DL1, so that the effective pulse width (i.e., a high level of 1) of the clock enable signal is delayed, and the delayed clock enable signal is output to the second input end of the first AND gate AND1. At the same time, the first input end of the first AND gate AND1 is also input with an undelayed clock enable signal. The input undelayed clock enable signal and the delayed clock enable signal are logically ANDed by the first AND gate AND1 to adjust the effective pulse width of the signal output to the first enable end EN1 of the first trigger circuit 23C.
[0075] Exemplarily, the delay circuit DL1 includes an even number of cascaded inverters. In the embodiment of the present application, the output signal logic of the even number of cascaded inverters is equivalent to that of no inverter, but the even number of inverters can constitute an oscillator so that the output voltage signal reaches a steady state and avoids voltage competition. At the same time, the cascaded multiple inverters can increase a certain delay of the clock enable signal, and the delayed clock enable signal and the undelayed clock enable signal can obtain clock enable signals with different pulse widths after a logical AND operation is performed at the first AND gate AND1.
[0076] In some possible implementations, the data signal exchanged between the first sub-circuit 10 and the second sub-circuit 30 includes a plurality of bits. Fig.10 As shown, the data buffer 25C includes a plurality of data buffer units 25C1. In the embodiment of the present application, when the data buffer 25C latches a data signal of a plurality of bits, a plurality of data buffer units 25C1 may be provided in the data buffer 25C. The data signal of a plurality of bits is latched by the plurality of data buffer units 25C1 and output to the second sub-circuit 30.
[0077] Exemplarily, each data cache unit 25C1 may be a register or a latch.
[0078] For example, Fig.10As shown, the asynchronous conversion interface of the third asynchronous clock interface circuit 20C also includes a voltage holding circuit 27C. The output end of the clock generation circuit 24C is coupled to the second enable end EN2 of the data buffer 25C through the voltage holding circuit 27C. In the embodiment of the present application, when more data signals are required, it may happen that the duration of the effective pulse width (i.e., high level) of the clock enable signal is insufficient to enable all the data cache units 25C1 to output data signals to the second sub-circuit 30. At this time, the voltage holding circuit 27C can be set to extend the enabling time of the clock enable signal to multiple data cache units 25C1.
[0079] In some possible implementations, such as Fig.11 As shown, the transmission interface of the third asynchronous clock interface circuit 20C also includes a first register 28C. The first data output terminal data1 of the first sub-circuit 10 is coupled to the input terminal of the first register 28C. The output terminal of the first register 28C is coupled to the first data input terminal data2 of the data buffer 25C. The first clock output terminal clk1 of the first sub-circuit 10 is also coupled to the third clock terminal CK3 of the first register 28C. In the embodiment of the present application, the data signal output by the first sub-circuit 10 can be beat by the first register 28C. At the same time, when there are many data signals that need to be interacted, the first sub-circuit 10 can cache the output data signal to the first register 28C first, and then output it to the data buffer 25C through the first register 28C.
[0080] In some possible implementations, such as Fig.11 As shown, the receiving interface of the third asynchronous clock interface circuit 20C also includes a second register 29C. The input end of the second register 29C is coupled to the second data output end data3 of the data buffer 25C. The output end of the second register 29C is coupled to the second data input end data4 of the second sub-circuit 30. The second clock output end clk2 of the second sub-circuit 30 is also coupled to the fourth clock end CK4 of the second register 29C. In the embodiment of the present application, the data signal output by the data buffer 25C is beat by the second register 29C, and then output to the second sub-circuit 30. At the same time, when there are many data signals that need to be interacted, the data signal output by the data buffer 25C can also be cached to the second register 29C first, and then output to the second sub-circuit 30 through the second register 29C.
[0081] In some possible implementations, such as Fig.12As shown, the receiving interface of the third asynchronous clock interface circuit 20C further includes a first level logic conversion circuit C1. The fourth signal terminal X4 of the second sub-circuit 30 is coupled to at least one of the multiple input terminals of the clock generation circuit 24C through the first level logic conversion circuit C1. In the embodiment of the present application, when the level triggering mode on the second sub-circuit 30 side is different from the level triggering mode of the clock generation circuit 24C on the third asynchronous clock interface circuit 20C side, the first feedback handshake signal outputted from the fourth signal terminal X4 of the second sub-circuit 30 can be logically converted (for example, from an edge-triggered level signal to a level-triggered level signal) through the first level logic conversion circuit C1 to adapt to the corresponding triggering mode.
[0082] In some possible implementations, such as Fig.12 As shown, the transmission interface of the third asynchronous clock interface circuit 20C also includes at least one of a second level logic conversion circuit C2, a third level logic conversion circuit C3 and a fourth level logic conversion circuit C4. Among them, the first signal terminal X1 of the first sub-circuit 10 is coupled to the trigger input terminal D1 of the first trigger circuit 23C through the second level logic conversion circuit C2. The output terminal of the first beat conversion circuit 21C is coupled to the third signal terminal X3 of the first sub-circuit 10 through the third level logic conversion circuit C3. The output terminal of the second beat conversion circuit 22C is coupled to the second signal terminal X2 of the second sub-circuit 30 through the fourth level logic conversion circuit C4. In the embodiment of the present application, the relevant description of the second level logic conversion circuit C2, the third level logic conversion circuit C3 and the fourth level logic conversion circuit C4 can refer to the relevant description of the first level logic conversion circuit C1 in the above embodiment.
[0083] For example, Fig.13 As shown, it is a level logic conversion circuit for converting edge triggering into level triggering. The level logic conversion circuit includes a third AND gate AND3 and a first D flip-flop 1D. The first input terminal of the third AND gate AND3 and the trigger input terminal D of the first D flip-flop 1D respectively input the edge trigger control signal before conversion. The trigger output terminal Q of the first D flip-flop 1D is coupled to the second input terminal of the third AND gate AND3. The clock terminal CK of the first D flip-flop 1D is used to input the corresponding clock signal. Exemplarily, when the level logic conversion circuit is the first level logic conversion circuit C1 and the fourth level logic conversion circuit C4, its clock terminal CK inputs the second clock signal of the second sub-circuit 30. When the level logic conversion circuit is the second level logic conversion circuit C2 and the third level logic conversion circuit C3, its clock terminal CK inputs the first clock signal of the first sub-circuit 10.
[0084] For example, Fig.14As shown, it is a level logic conversion circuit for converting a level trigger into an edge trigger. The level logic conversion circuit may include a first inverter INV1 and a second D flip-flop 2D. The input end of the first inverter INV1 inputs a control signal before conversion and a control signal after conversion respectively. The output end of the first inverter INV1 is coupled to the input end of the second D flip-flop 2D. The clock end CK of the second D flip-flop 2D inputs a corresponding clock signal. Fig.14 The corresponding relationship between the clock signal input to the level logic conversion circuit shown in FIG. Fig.13 The description related to the corresponding relationship of the clock signal input to the level logic conversion circuit shown is not repeated here.
[0085] For example, the first sub-circuit 10 and the second sub-circuit 30 are level-triggered, and the third asynchronous clock interface circuit 20C is edge-triggered. Fig.15 As shown, the first level logic conversion circuit C1 and the second level logic conversion circuit C2 are level triggers converted into voltage-changing triggers, which can be used as follows Fig.14 The third level logic conversion circuit C3 and the fourth level logic conversion circuit C4 are edge-triggered to level-triggered conversions, which can be used as follows: Fig.13 The level logic conversion circuit shown.
[0086] Optionally, the processing chip 100 is a large-scale SoC. Fig.16As shown, the SoC includes a plurality of processing cores core1, a system bus SBus, a high speed peripheral bus Bus1, a low speed peripheral bus Bus2, a control bus Bus3, a first storage circuit M1, a second storage circuit M2, a control circuit Cont1, and a third asynchronous clock interface circuit 20C. Exemplarily, the third asynchronous clock interface circuit 20C can be coupled between the system bus SBus and the high speed peripheral bus Bus1, in which case the first sub-circuit 10 and the second sub-circuit 30 can be the processing core core1 and the first storage circuit M1, respectively. Exemplarily, the third asynchronous clock interface circuit 20C can be coupled between the system bus SBus and the low speed peripheral bus Bus2, in which case the first sub-circuit 10 and the second sub-circuit 30 can be the processing core core1 and the second storage circuit M2, respectively. Exemplarily, the third asynchronous clock interface circuit 20C can be coupled between the system bus SBus and the control bus Bus3, in which case the first sub-circuit 10 and the second sub-circuit 30 can be the processing core core1 and the control circuit Cont1, respectively. In the embodiment of the present application, the system bus SBus and other buses are coupled through the third asynchronous clock interface circuit 20C. The system bus SBus of the SoC and other buses can be set as an asynchronous circuit. This avoids the problem that the SoC under the synchronous integrated circuit, with the increase of the integration area, etc., leads to the problem of delayed synchronous clock signals due to excessively long signal lines, and in the case of ultra-large scale, the SoC design layout is difficult, communication interaction conflicts, increased communication power consumption, poor scalability, reduced communication reliability, reduced throughput and other problems. The scale of the SoC under the asynchronous clock integrated circuit can be expanded more easily through the third asynchronous clock interface circuit 20C.
[0087] Optionally, the processing chip 100 is a NoC. Fig.17As shown, it is a schematic diagram of the structure of a NoC with a two-dimensional grid (2D-mesh) structure. The NOC includes multiple routing nodes R1 and multiple processing nodes N1. Multiple routing nodes R1 are connected to form an array structure, and each routing node R1 is coupled to a processing node N1 respectively. Asynchronous interaction of data signals and asynchronous interaction of control signals are realized between multiple processing nodes N1 through corresponding routing nodes R1. Optionally, three sub-areas of different specifications can be divided on the NoC, and a third asynchronous clock interface circuit 20C can be set in each of the three sub-areas of different specifications. The three sub-areas of different regular types are: sub-area one formed by a single routing node R1 or a single processing node N1; sub-area two including a pair of interconnected routing nodes R1 and processing nodes N1; and sub-area three including multiple pairs of interconnected routing nodes R1 and processing nodes N1. Exemplarily, for multiple sub-areas one, asynchronous interaction can be realized between sub-area one for routing node R1 and sub-area one for processing node N1 through a third asynchronous clock interface circuit 20C. Exemplarily, for two adjacent sub-areas 2, routing nodes R1 in the two sub-areas 2 may be interconnected through a third asynchronous clock interface circuit 20C to achieve direct asynchronous interaction between the two sub-areas 2. Exemplarily, for two adjacent sub-areas 3, adjacent routing nodes R1 in the two sub-areas 3 may be interconnected through at least one third asynchronous clock interface circuit 20C to achieve asynchronous interaction between the two sub-areas 3.
[0088] In some possible implementations, the processing chip 100 further includes a fourth asynchronous clock interface circuit. The second sub-circuit 30 is coupled to the first sub-circuit 10 via the fourth asynchronous clock circuit. The second sub-circuit 30 can perform data signal interaction and control signal interaction with the first sub-circuit 10 via the fourth asynchronous clock circuit. In the embodiment of the present application, the technical principle and beneficial effects of the second sub-circuit 30 performing signal interaction with the first sub-circuit 10 via the fourth asynchronous clock circuit can be referred to the relevant description of the third asynchronous clock circuit 20C in the above embodiment, which will not be repeated here.
[0089] The embodiment of the present application also provides a design method, by which the design method is based on the existing EDA design tool to design the following Figure 6 , Figure 7 , Figure 8 , Fig. 9 , Fig.10 , Fig.11 , Fig.12 , Fig.13 , Fig.14 , Fig.15 , Fig.16 and Fig.17 The processing chip of the structure shown in FIG. Fig.18As shown, the design method specifically includes the following contents recorded in steps S110 to S140:
[0090] S110: Combine the first description file and the second description file to obtain a third description file.
[0091] The first description file is used to describe the logic function of the first chip. The first chip includes a plurality of sub-circuits. The second description file is used to describe the logic function of the asynchronous clock interface circuit. The third description file is used to describe the logic function of the processing chip; the second description file is used to describe the logic function of the third asynchronous clock interface circuit 20C. The third description file is used to describe the logic function of the processing chip 100.
[0092] In an embodiment of the present application, a first description file about the first chip has been generated in an existing EDA design tool. Both the first description file and the second description file are RTL description files, wherein the first description file is an original description file used to describe the logic function of the first chip in the existing EDA design tool. The second description file is an RTL description file obtained after the logic function of the third asynchronous clock interface circuit 20C is integrated into an IP core. In actual applications, the third asynchronous clock interface circuit 20C for an asynchronous clock integrated circuit is integrated into a basic IP core. In subsequent designs, the corresponding second description file and the first description file can be directly called to be combined to design an asynchronous clock integrated circuit.
[0093] In some possible implementations, when multiple sub-circuits in the first chip are based on the same clock domain, the multiple sub-circuits in the first chip receive the same clock signal. Fig.19 As shown, the multiple sub-circuits in the first chip include a sending circuit sen1 and a receiving circuit rec1 as an example: at this time, the sending circuit sen1 and the receiving circuit rec1 are based on the same clock signal. Step S110 may include the following: Fig. 20 The operations of steps S111 to S113 shown are as follows:
[0094] S111 , setting the sending circuit sen1 and the receiving circuit rec1 described in the first description file to different local clock domains to obtain the first sub-circuit 10 and the second sub-circuit 30 .
[0095] S112: Construct a second description file.
[0096] In the embodiment of the present application, when the first description file is an RTL description file for a synchronous integrated circuit, the synchronous integrated circuit does not include the logic function of the third asynchronous clock interface circuit 20C used for asynchronous interaction. At this time, it is necessary to first design the logic function of the third asynchronous interface circuit 20C and generate the corresponding second description file. The specific operations are as follows:
[0097] First, the logic functions of each circuit in the sending interface and the receiving interface are generated, for example, the logic functions of the first beat conversion circuit 21C and the second beat conversion circuit 22C. Optionally, according to the actual application scenario, the following can also be generated: Fig.12 The logic function of at least one of the first level logic conversion circuit C1, the second level logic conversion circuit C2, the third level logic conversion circuit C3 and the fourth level logic conversion circuit C4 is shown.
[0098] Secondly, the logic functions of the circuits in the asynchronous conversion circuit are generated, such as the logic functions of the first trigger circuit 23C. Optionally, the logic functions of the data buffer 25C, the clock generation circuit 24C, the pulse width delay circuit 26C, the voltage holding circuit 27C and other circuits can also be generated.
[0099] Next, a timing cell library of the third asynchronous clock interface circuit 20C is constructed, and timing constraints are imposed on the logic functions of the generated third asynchronous clock interface circuit 20C. Fig.21 As shown, the interactive transmission of the control signal of the third asynchronous clock interface circuit 20C adopts the protocol mechanism of the handshake signal. In the figure, Δt1 is the trigger time of the clock enable signal to the data buffer 25C. Δt2 is the preparation time for the clock enable signal to drive the data buffer 25C to transmit the data signal. Δt3 is the holding time after the clock enable signal drives the data buffer 25C. In the scenario where a single-bit data signal needs to be transmitted, due to the mutually exclusive structure of the circuit, the duration Δt2 is required to meet the minimum valid level constraint of not less than the data buffer 25C, and the first request handshake signal and the clock enable signal after the pulse width delay have a set timing requirement. Since the clock enable signal after the pulse width delay and the first request handshake signal are both generated through multi-stage gate circuits, the timing requirement is easy to meet. For the scenario where multiple bits of data signals need to be transmitted, the data signal needs to be bound to the first request handshake signal, and Δt2 is required to meet the minimum valid level constraint of not less than the data buffer 25C, while Δt1 also needs to meet the establishment time greater than the data buffer 25C. The third asynchronous clock interface circuit 20C is customized into a standard unit using these constraints, and key signal timing characteristics are extracted.
[0100] S113: Combine the first description file with the second description file to obtain a third description file.
[0101] Exemplarily, the second description file is inserted into the interface position specified in the first description file by automatic insertion, and the transmission path of the original sending circuit sen1 and the receiving circuit rec1 in the original first description file is interrupted, and a new transmission path of the first sub-circuit 10 and the second sub-circuit 30 is generated by the second description file, thereby generating a third description file.
[0102] In the case where the first description file is an RTL description file for describing a synchronous integrated circuit, a third description file may be obtained according to the operations of the above steps S111 to S113.
[0103] In some possible implementations, when some or all of the sub-circuits in the first chip are based on different clock domains, the first description file is an RTL description file related to the asynchronous clock integrated circuit. Fig. 22 As shown, the first chip of the asynchronous clock integrated circuit already includes a first sub-circuit 10 and a second sub-circuit 30 based on different clock domains. Step S110 may include the following steps: Fig.23 The operations of step S111' to step S112' are as follows:
[0104] S111', construct a second description file.
[0105] In some possible implementations, the first description file in some existing EDA design tools already has description language for the related logical functions of the first beat conversion circuit 21C and the second beat conversion circuit 22C. At this time, it is only necessary to design the design of the logical functions of the first trigger circuit 23C, the data buffer 25C, the clock generation circuit 24C, the pulse width delay circuit 26C, the voltage holding circuit 27C and other related circuits to produce the second description file. For the description of the design of the logical functions of the first trigger circuit 23C, the data buffer 25C, the clock generation circuit 24C, the pulse width delay circuit 26C, the voltage holding circuit 27C and other related circuits, reference can be made to the relevant description in the above-mentioned embodiment of step S112, which will not be repeated here.
[0106] S112', combining the first description file with the second description file to obtain a third description file.
[0107] For the relevant description of step S112', reference may be made to the relevant description in the above-mentioned embodiment of step S113, which will not be repeated here.
[0108] S120: Perform a timing constraint check on the third description file.
[0109] In the embodiment of the present application, the data signal is bound to the first request handshake signal, and the delay between the data signal and the first request handshake signal is required to meet the establishment time of the asynchronous interface. The third asynchronous clock interface circuit 20C obtained in step S110 needs to meet the following constraints:
[0110] Constraint 1: According to the actual application scenario, the first request handshake signal, the second request handshake signal, the first feedback handshake signal and the second feedback handshake signal are set with specific timing relationship requirements, and it is necessary to ensure that the delay between the data signal and the first request handshake signal should meet the setup time of the third asynchronous clock interface circuit 20C. In the actual application scenario, the first clock signal or the second clock signal can be used as the reference to check the setting constraints for timing constraint checking. Fig.24 and Fig.25 As shown, the first request handshake signal and data signal from the transmitting end to the input end of the third asynchronous clock interface circuit 20C are output from the register of the same clock domain, and the second request handshake signal and data signal from the input end of the third asynchronous clock interface circuit 20C to the receiving end enter the register of the same clock domain, so a data to data setup check or a data to data hold check can be performed to ensure that no error occurs in data transmission when the request handshake signal and the data signal enter the third asynchronous clock interface circuit 20C across different clock domains. Fig.25 As shown in , AsyncIf_0 corresponds to the third asynchronous clock interface circuit 20C; AsyncSend_0 corresponds to the first sub-circuit 10; and AsyncRec_0 corresponds to the second sub-circuit 30. The specific command format on the EDA design tool can be:
[0111] Command 1: set_data_check-from xxx / AsyncIf_0 / X1-to xxx / AsyncIf_0 / Data_A-setup value-hold value;
[0112] Command one: set_data_check-from xxx / AsyncRec_0 / X2-to xxx / AsynRec_0 / Data_B-setup value-hold value.
[0113] Constraint 2: The timing of the two ends of the third asynchronous clock interface circuit 20C (i.e., the output of the first sub-circuit 10 at the transmitting end to the input of the third asynchronous clock interface circuit 20C; and the output of the third asynchronous clock interface circuit 20C to the input of the second sub-circuit 30) needs to be optimized to the shortest delay to avoid performance loss caused by excessive routing delay. Fig.26 As shown, a first maximum delay max delay1 from the first sub-circuit 10 to the third asynchronous clock interface circuit 20C and a second maximum delay max delay2 from the third asynchronous clock interface circuit 20C to the second sub-circuit 30 are set, i.e., the maximum value of the data transmission delay at both ends of the third asynchronous clock interface circuit 20C, to ensure the highest performance of data transmission. Fig.26 In the example, max_delay1 is the first maximum delay. max_delay2 is the second maximum delay. Reg_0 represents the register that outputs the data signal in the first sub-circuit 10. AsyncIf_0 is the third asynchronous clock interface circuit 20C coupled to the first sub-circuit 10. Reg_1 represents the register that receives the data signal in the second sub-circuit 30. AsyncIf_1 is the third asynchronous clock interface circuit 20C coupled to the second sub-circuit 30. The specific command format on the EDA design tool can be:
[0114] Command 1: set_max_delay-datapath_only-from xxx / Reg_0 / Q-to xxx / AsyncIf_0 / Din delay_value;
[0115] Command 2: set_max_delay-datapath_only-from yyy / AsyncIf_1 / Dout-to yyy / Reg_1 / Ddelay_value.
[0116] S130 , modifying the third description file after the timing constraints are imposed, and applying design constraints to the modified third description file.
[0117] In an embodiment of the present application, the third description file in the above step S110 can be modified. The paths that do not meet the timing constraint check requirements in the above step S120 can also be repaired. In the repair process, a certain buffer can be inserted into the path corresponding to the violation that violates the timing constraint to make timing adjustments, and the operations of step S120 and step S130 are iterated until all timing paths meet the timing constraint check conditions, thereby obtaining the final third description file after the timing constraints.
[0118] S140 , obtaining a design layout of the processing chip 100 based on the third description file after the timing constraints.
[0119] In the embodiment of the present application, after the confirmed third description file is obtained, the logic circuit structure is designed according to the third description file to obtain a design layout, and the processing chip 100 is manufactured in a subsequent process according to the design layout.
[0120] In some possible implementations, step S140 includes: Fig. 20 and Fig.23 The operations of the following steps S141-S145 are shown:
[0121] S141, performing synthesis processing of logic gate circuits on the third description file after timing constraints to obtain a component structure of the logic circuit.
[0122] In the embodiment of the present application, the corresponding logic functions in the third description file are matched with the logic gate circuits, and the corresponding logic gate circuit structure is synthesized.
[0123] Step S142: Map the synthesized logic gate circuit structure to obtain a corresponding netlist.
[0124] Step S143: Perform backend design (physical design) layout according to the mapped network table.
[0125] Step S144: Perform static timing analysis (STA) on the circuit structure after back-end design.
[0126] Step S145 , obtaining the final design layout of the processing chip 100 after completing the static timing analysis.
[0127] In the embodiment of the present application, after the design layout is obtained, in the subsequent process, a corresponding mask and the like can be obtained according to the design layout to perform manufacturing and processing of the processing chip 100 .
[0128] In the embodiment of the present application, based on the design method including steps S110 to S140, the processing chip 100 including the third asynchronous clock interface circuit 20C can be quickly designed based on the existing EDA design tools. Compared with the design based on STG, the design based on physical design tools and the design based on asynchronous EDA design tools, the design method of the embodiment of the present application can quickly, efficiently and reliably complete the design of ultra-large-scale asynchronous clock integrated circuits.
[0129] The embodiment of the present application improves a processing chip, a design method and an electronic device. The processing chip includes a first subcircuit, a third asynchronous clock interface circuit and a second subcircuit. The third asynchronous clock interface circuit includes a first beat conversion circuit, a second beat conversion circuit and a first trigger circuit. The first subcircuit performs asynchronous interaction of control signals between the third asynchronous clock interface circuit and the second subcircuit, specifically: the first subcircuit sends a first request handshake signal to the first trigger circuit, and after the first trigger circuit receives the first request handshake signal, it triggers the output of the second request handshake signal, and outputs the second request handshake signal to the first beat conversion circuit and the second beat conversion circuit respectively. Then the first beat conversion circuit converts the second request handshake signal into a second feedback handshake signal in the same clock domain as the first subcircuit after beating multiple beats and outputs it to the first subcircuit. The second beat conversion circuit converts the received second request handshake signal into a second request handshake signal in the same clock domain as the second subcircuit after beating multiple beats and outputs it to the second subcircuit. In the embodiment of the present application, by setting corresponding beats for the first beat conversion circuit and the second beat conversion circuit, a determined timing can be obtained. At this time, the first trigger circuit outputs signals to the first beat conversion circuit and the second beat conversion circuit at the same time, and the first beat conversion circuit and the second beat conversion circuit beat the beat at the same time. Compared with the implementation method in which the two beat conversion circuits beat the beat separately, the embodiment of the present application can reduce the delay caused by the beat conversion circuit beating the beat, and in the circuit in which the beat is beaten approximately multiple times, the more the delay can be reduced.
[0130] The processor involved in the embodiments of the present application may be a chip. For example, it may be a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on chip (SoC), a central processor unit (CPU), a network processor (NP), a digital signal processor (DSP), a microcontroller unit (MCU), a programmable logic device (PLD), or other integrated chips.
[0131] The memory involved in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), and direct RAM (DR RAM). It should be noted that the memory of the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.
[0132] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.
[0133] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.
[0134] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and modules described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.
[0135] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.
[0136] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one device or distributed on multiple devices. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment.
[0137] In addition, each functional module in each embodiment of the present application may be integrated into one device, or each module may exist physically separately, or two or more modules may be integrated into one device.
[0138] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using a software program, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loading and executing a computer program instruction on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from a website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (Digital Subscriber Line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium may be any available medium that a computer can access or may contain one or more servers, data centers and other data storage devices that can be integrated with a medium. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a DVD), or a semiconductor medium (eg, a solid state disk (SSD)).
[0139] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A processing chip, characterized in that: It includes a first sub-circuit, an asynchronous clock interface circuit and a second sub-circuit; the asynchronous clock interface circuit includes a first beat conversion circuit, a first trigger circuit and a second beat conversion circuit; The first signal terminal of the first subcircuit is coupled to the trigger input terminal of the first trigger circuit; the trigger output terminal of the first trigger circuit is coupled to the input terminal of the first beat conversion circuit and the input terminal of the second beat conversion circuit respectively; the output terminal of the first beat conversion circuit is coupled to the third signal terminal of the first subcircuit; the output terminal of the second beat conversion circuit is coupled to the second signal terminal of the second subcircuit; The first clock output terminal of the first sub-circuit is coupled to the first clock terminal of the first beat conversion circuit; the second clock output terminal of the second sub-circuit is coupled to the second clock terminal of the second beat conversion circuit.
2. The processing chip according to claim 1, characterized in that: The asynchronous clock interface circuit also includes a clock generation circuit; The multiple input terminals of the clock generation circuit are respectively coupled to the first signal terminal of the first sub-circuit, the trigger output terminal of the first trigger circuit and the fourth signal terminal of the second sub-circuit; the output terminal of the clock generation circuit is coupled to the first enable terminal of the first trigger circuit.
3. The processing chip according to claim 2, characterized in that: The asynchronous clock interface circuit also includes a data buffer; The first data output terminal of the first sub-circuit is coupled to the first data input terminal of the data buffer; the second data output terminal of the data buffer is coupled to the second data input terminal of the second sub-circuit; The output terminal of the clock generating circuit is also coupled to the second enable terminal of the data buffer and the first enable terminal of the first trigger circuit.
4. The processing chip according to claim 3, characterized in that: The asynchronous clock interface circuit also includes a first register and a second register; The first data output terminal of the first subcircuit is coupled to the input terminal of the first register; the output terminal of the first register is coupled to the first data input terminal of the data buffer; the clock output terminal of the first subcircuit is also coupled to the third clock terminal of the first register; The input end of the second register is coupled to the second data output end of the data buffer; the output end of the second register is coupled to the second data input end of the second sub-circuit; and the clock output end of the second sub-circuit is also coupled to the fourth clock end of the second register.
5. The processing chip according to claim 3 or 4, characterized in that: The asynchronous clock interface circuit also includes a voltage holding circuit; the output end of the clock generation circuit is coupled to the second enable end of the data buffer through the voltage holding circuit.
6. The processing chip according to any one of claims 2 to 5, characterized in that: The asynchronous clock interface circuit also includes a pulse width delay circuit; the output end of the clock generation circuit is coupled to the first enable end of the first trigger circuit through the pulse width delay circuit.
7. The processing chip according to claim 6, characterized in that: The pulse width delay circuit includes a delay circuit and a first AND gate; the output end of the clock generation circuit is coupled to the first input end of the first AND gate; the output end of the clock generation circuit is also coupled to the second input end of the first AND gate through the delay circuit.
8. The processing chip according to any one of claims 2 to 7, characterized in that: The clock generation circuit comprises an XOR gate, an XOR gate and a second AND gate; the multiple input terminals of the clock generation circuit comprise a first input terminal of the XOR gate, a second input terminal of the XOR gate, a first input terminal of the XOR gate and a second input terminal of the XOR gate; The first input terminal of the XOR gate is coupled to the first signal terminal of the first sub-circuit; the second input terminal of the XOR gate is coupled to the trigger output terminal of the first trigger circuit; The first input terminal of the XNOR gate is coupled to the fourth signal terminal of the second sub-circuit; the second input terminal of the XNOR gate is coupled to the trigger output terminal of the first trigger circuit; The output end of the XOR gate is coupled to the first input end of the second AND gate; the output end of the XOR gate is coupled to the second input end of the second AND gate; the output end of the second AND gate is the output end of the clock generation circuit.
9. The processing chip according to any one of claims 2 to 8, characterized in that: The asynchronous clock interface circuit also includes a first level logic conversion circuit; the fourth signal terminal of the second sub-circuit is coupled to at least one of the multiple input terminals of the clock generation circuit through the first level logic conversion circuit.
10. The processing chip according to any one of claims 1 to 9, characterized in that: The asynchronous clock interface circuit further includes at least one of a second level logic conversion circuit, a third level logic conversion circuit, and a fourth level logic conversion circuit; The first signal terminal of the first sub-circuit is coupled to the trigger input terminal of the first trigger circuit through the second level logic conversion circuit; The output end of the first beat conversion circuit is coupled to the third signal end of the first sub-circuit through the third level logic conversion circuit; The output end of the second beat conversion circuit is coupled to the second signal end of the second sub-circuit through the fourth level logic conversion circuit.
11. The processing chip according to any one of claims 1 to 10, characterized in that: The first sub-circuit and / or the second sub-circuit is any one of a storage circuit, a logic control circuit, a routing interface circuit, and a neural network processing circuit.
12. A design method, characterized in that: Used to design a processing chip according to any one of claims 1 to 11; the design method comprises: The first description file and the second description file are combined to obtain a third description file; the first description file is used to describe the logic function of the first chip; the first chip includes multiple sub-circuits; the multiple sub-circuits are based on the same clock domain; or some or all of the multiple sub-circuits are based on different clock domains; the second description file is used to describe the logic function of the asynchronous clock interface circuit; the third description file is used to describe the logic function of the processing chip; Performing timing constraints on the third description file; The logic circuit structure of the processing chip is obtained based on the third description file after the timing constraints.
13. The design method according to claim 12, characterized in that: The design method further comprises: Before obtaining the logic circuit structure of the processing chip based on the third description file after the timing constraints, the third description file after the timing constraints is modified, and the modified third description file is subjected to design constraints.
14. An electronic device, characterized in that: It comprises a circuit board and a processing chip as described in any one of claims 1 to 11; the processing chip is arranged on the circuit board.
Citation Information
Patent Citations
Synchronous circuit for asynchronous signals
CN104767516A
Data read-write synchronization circuit and data read-write method
CN110045782A
Low-power clock-gated synchronizer, a data processing system that incorporates the same and a synchronization method
US10075153B2
Low latency asynchronous interface circuits
US9685953B1
Method for designing asynchronous circuit, and electronic device
WO2023279341A1
Cited By
Processing chip, design method, and electronic device
EP4693060A1
Processing chip, design method, and electronic device
WO2024245162A1
