Data processing device and method

By separating the calculation and transmission process in the data processing device, using the independent operation of the processing module and the Tx transmitter, the problem of insufficient flexibility and efficiency in various data processing and fast data volume processing is solved, and high efficiency is achieved data processing and parallel processing of computing modules.

CN112817638BActive Publication Date: 2024-11-12STREAM COMPUTING INC
View PDF 11 Cites 0 Cited by

Patent Information

Application Number
CN201911128755.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2019-11-18
Publication Date
2024-11-12
Estimated Expiration
2039-11-18

AI Technical Summary

Technical Problem

Existing chips have insufficient flexibility and processing capabilities when processing data in a variety of different fields and rapidly growing data volumes, resulting in slow computing speed and low efficiency, especially in multi-core architectures that affect the overall performance of the chip.

Method used

Design a data processing device, including a processing module, a computing module and a Tx transmitter, the processing module sends calculation instructions and transmission instructions, the calculation module performs calculations and generates results, and the Tx transmitter sends calculation results independently, without the processing module participating in the calculation process and The result transmission improves the efficiency of the processing module and the utilization of the computing module.

Benefits of technology

It realizes efficient data processing of processing modules, improves overall data processing efficiency and utilization of computing modules, reduces program writing and compilation difficulty, and avoids the impact of data transmission on chip performance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112817638B_ABST
    Figure CN112817638B_ABST
Patent Text Reader

Abstract

The present invention discloses a data processing device and method, the data processing device comprising: a processing module (10) for sending instructions, the instructions comprising calculation instructions and transmission instructions; a calculation module (20) for performing calculations and generating calculation results based on the calculation instructions sent by the processing module; a Tx transmitter (30) for transmitting the calculation results generated by the calculation module based on the transmission instructions. In the data processing device provided by the embodiment of the present invention, the processing module sends the calculation instructions and the transmission instructions, and the Tx transmitter is connected to the calculation module to transmit the calculation results generated by the calculation module. The processing module does not need to participate in the calculation process and the transmission of the calculation results all the time, which further enables the processing module to efficiently process other data, thereby improving the efficiency of the processing module.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of processing core structure, and in particular to a data processing device and method. Background Art

[0002] With the development of science and technology, human society is rapidly entering the intelligent era. An important feature of the intelligent era is that people obtain more and more types of data, the amount of data they obtain is increasing, and the speed of processing data is getting higher and higher.

[0003] Chips are the cornerstone of data processing, and they fundamentally determine people's ability to process data. From the perspective of application areas, there are two main routes for chips: one is the general chip route, such as the central processing unit (CPU), which can provide great flexibility, but the effective computing power is relatively low when processing algorithms in specific fields; the other is the dedicated chip route, such as the tensor processing unit (TPU), which can exert high effective computing power in certain specific fields, but in the face of flexible and changeable more general fields, their processing capabilities are relatively poor or even unable to handle.

[0004] Since data in the intelligent era are of many types and huge amounts, chips are required to have both extremely high flexibility to handle ever-changing algorithms in different fields and extremely strong processing capabilities to quickly process huge and rapidly growing amounts of data. Summary of the invention

[0005] (I) Purpose of the invention

[0006] The object of the present invention is to provide a data processing device and method, wherein the data processing device is provided with a processing module, a computing module and a Tx transmitter, the processing module sends a computing instruction and a transmission result, the computing module performs a calculation based on the computing instruction and generates a computing result, and the Tx transmitter sends the computing result generated by the computing module based on the transmission instruction.

[0007] In the data processing device provided by the present invention, the processing module sends calculation instructions and transmission instructions, and the Tx transmitter is connected to the calculation module to send the calculation results generated by the calculation module. The processing module does not need to participate in the calculation process and the transmission of the calculation results all the time, which further enables the processing module to efficiently process other data, thereby improving the efficiency of the processing module.

[0008] (II) Technical solution

[0009] To solve the above problems, the first aspect of the present invention provides a data processing device, including: a processing module, used to send instructions, the instructions include calculation instructions and transmission instructions; a calculation module, used to perform calculations based on the calculation instructions and generate calculation results; a Tx transmitter, connected to the calculation module, used to send the calculation results generated by the calculation module based on the transmission instructions.

[0010] In the data processing device provided by the embodiment of the present invention, the processing module sends calculation instructions and transmission instructions, and the Tx transmitter is connected to the calculation module to send the calculation results generated by the calculation module. The processing module does not need to participate in the calculation process and the transmission of the calculation results all the time, which further enables the processing module to efficiently process other data, thereby improving the efficiency of the processing module.

[0011] Furthermore, the calculation module is also used to cache calculation results; and the Tx transmitter is used to send the calculation results cached in the calculation module based on the transmission instruction.

[0012] Furthermore, the calculation module includes an execution unit and an output register group; the execution unit is used to perform calculations based on calculation instructions and generate the calculation results; the output register group is used to cache the calculation results, and the output register group includes at least one output register.

[0013] Furthermore, the output register group is further configured to send a first signal when the output register group is capable of receiving the calculation result, and the execution unit sends the calculation result to the output register group based on the first signal.

[0014] Furthermore, the output register group is further used to send a second signal when the output register group cannot receive the calculation result, and the execution unit suspends sending the calculation result based on the second signal.

[0015] Furthermore, the output register group is also used to send a third signal when the number of the calculation results stored in the output register group is lower than a first preset value, and the Tx transmitter suspends fetching data from the output register group based on the third signal.

[0016] Furthermore, the output register group is further used to send a fourth signal when the number of the calculation results stored in the output register group is higher than a second preset value, and the Tx transmitter fetches data from the output register group based on the fourth signal.

[0017] Furthermore, the calculation module includes one or more multipliers and adders; the number of output registers in the output register group is greater than or equal to the number of multipliers and adders.

[0018] Furthermore, the number of output registers in the output register group is a multiple of the number of multipliers and adders provided in the calculation module.

[0019] Further preferably, the number of output registers provided in the output register group is at least twice the number of multipliers and adders provided in the execution unit.

[0020] Furthermore, the processing module is used to send the calculation instruction and the sending instruction in no particular order.

[0021] The second aspect of the present invention further provides a core structure, comprising the data processing device and storage module provided in the first aspect; the storage module is used to store instructions and data, and the instructions include the calculation instructions and the transmission instructions.

[0022] Furthermore, the storage module includes multiple storage units; the multiple storage units include a first storage unit, a second storage unit and a third storage unit; the first storage unit is used to store the instructions and / or store data for reading and writing by the processing module; the second storage unit is used to store the calculation results generated by the calculation module and / or store data for reading and writing by the calculation module; the third storage unit is used to store data received from the outside.

[0023] The third aspect of the present invention also provides a chip, comprising one or more core structures provided by the second aspect.

[0024] The fourth aspect of the present invention further provides a card board, comprising one or more chips according to the third aspect.

[0025] The fifth aspect of the present invention also provides an electronic device, comprising one or more card boards of the fourth aspect.

[0026] In the sixth aspect of the present invention, a data processing method is also provided, including a processing module sending calculation instructions and transmission instructions; a calculation module performing calculations based on the calculation instructions to generate calculation results; a Tx transmitter is connected to the calculation module and sends the calculation results generated by the calculation module based on the transmission instructions.

[0027] (III) Beneficial effects

[0028] The above technical solution of the present invention has the following beneficial technical effects:

[0029] (1) In the data processing device provided by the present invention, the processing module sends calculation instructions and transmission instructions, the calculation module generates calculation results, and the Tx transmitter is connected to the calculation module to send the calculation results generated by the calculation module. The processing module does not need to participate in the calculation process and the transmission of the calculation results all the time, which further enables the processing module to efficiently process other data, thereby improving the efficiency of the processing module.

[0030] (2) In the data processing device provided by the embodiment of the present invention, the processing module sends the calculation instruction and the transmission instruction in no particular order, and the transmission instruction instructs the Tx transmitter to send the calculation result generated by the calculation module, so that the calculation module executes the calculation and the Tx transmitter sends the calculation result in parallel, thereby improving the utilization rate of the calculation module.

[0031] (3) The data processing device provided in the embodiment of the present invention does not require the processing module to split the task, thereby reducing the difficulty of program writing and compilation. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a schematic diagram of the structure of a data processing device;

[0033] Figure 2 is a schematic diagram of the structure of another data processing device;

[0034] Figure 3 yes Figure 2 A timing diagram of another data processing device performing data processing;

[0035] Figure 4 is a schematic structural diagram of a data processing device according to an embodiment of the present invention;

[0036] Figure 5 is a schematic structural diagram of a data processing device according to an embodiment of the present invention;

[0037] Figure 6 is a schematic structural diagram of a data processing device according to an embodiment of the present invention;

[0038] Figure 7 is a schematic diagram of a core structure according to an embodiment of the present invention;

[0039] Figure 8 is a schematic diagram of a core structure according to an embodiment of the present invention;

[0040] Fig. 9 It is a flowchart of a data processing method according to an embodiment of the present invention.

[0041] Reference numerals:

[0042] 1: core; 10: processing module; 20: computing module; 21: execution unit; 22: output register group; 30: Tx transmitter; 40: Rx receiver; 100: data processing device; 200: storage module. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings. It should be understood that these descriptions are only exemplary and are not intended to limit the scope of the present invention. In addition, in the following description, the description of well-known structures and technologies is omitted to avoid unnecessary confusion of the concept of the present invention.

[0044] In a multi-core architecture, multiple cores often participate in the execution of a task at the same time. At this time, data will be transmitted between the cores. The data transmission may affect the calculation of the core, thereby reducing the effective performance of the entire chip, such as the data processing efficiency of the chip.

[0045] Figure 1 It is a structural schematic diagram of a data processing device.

[0046] like Figure 1 As shown, the data processing device includes a central processing unit CPU, a data transceiver module TR, a processing unit (PU), and a storage module Memory (referred to as Mem), wherein the central processing unit CPU, the transceiver module TR and the processing unit PU share the same storage module mem. Under the control of the CPU, the PU performs calculations and the TR performs data transmission and reception. When the calculation of a task is completed, the result data generated by the calculation of this task is transmitted. The specific process is as follows:

[0047] The CPU reads the computing task from the storage module Mem, and then issues a computing instruction to the PU, which starts the computing. After the PU completes the computing, it writes the computing result into the storage module Mem, and sends a message of the computing completion to the CPU. Based on the message of the computing completion, the CPU issues a transmission instruction to the TR, and the TR reads the computing result from the storage module Mem, and then starts the data transmission. After the TR completes the data transmission, it sends a message of the data transmission completion to the CPU.

[0048] This type of data processing device usually has the following defects:

[0049] 1. Since data calculation and transmission of calculation results are processed serially, the chip calculation speed is slow and the effectiveness is reduced.

[0050] 2. The CPU needs to control the calculation and transmit the calculation results based on the running status of the PU and TR, that is, based on the message of PU calculation completion and the message of TR transmission completion, it instructs the PU to calculate or the TR to transmit, resulting in low CPU data processing efficiency and inability to fully utilize the CPU.

[0051] 3. The calculation result needs to be written into Mem by PU first, and then read out from Mem by TR and sent out. This will consume more time and power consumption, resulting in low data processing efficiency.

[0052] Figure 2 It is a schematic diagram of the structure of another data processing device.

[0053] like Figure 2 As shown, the data processing device includes a central processing unit CPU, a sending and receiving module TR, a computing unit PU, and a storage module Mem, wherein the storage module Mem includes a first storage unit M0 and a second storage unit M1. In the data processing device, two storage units are provided so that PU and TR can read and write data at the same time, realizing parallel processing of computing data and transmitting computing results. The process of completing the calculation of a task and transmitting the result data generated by the calculation of this task is as follows:

[0054] First, the CPU reads the computing task from the storage module Mem and splits the task into several subtasks; then the CPU gives a computing instruction to the PU, the PU starts the calculation for the first time and writes the calculation result into the first storage unit M0; when the PU completes the calculation, the PU sends the information that the calculation is completed to the CPU; the CPU gives a second computing instruction to the PU, the PU starts the second calculation and stores the calculation result in the second storage unit M1. At the same time, the CPU gives a transmission instruction to the TR, instructing the TR to read the data stored in the first storage unit M0 and transmit it. When TR transmits the data stored in MO, the PU completes the second calculation, writes the calculation result into the first storage unit M1, and sends a signal that the second calculation is completed to the CPU; then the CPU instructs the PU to perform the third calculation, the PU starts the third calculation and writes the result of the third calculation into MO. At the same time, the CPU instructs the TR to transmit the result data of the second calculation stored in the second calculation unit M1, and so on, until the entire computing task is completed. Figure 3 yes Figure 2 Another data processing device performs data processing as shown in the timing diagram. Figure 3 As shown, in the data processing device, the CPU splits the tasks. Figure 3 The example in the figure is to split the task into 2N times.

[0055] Usually, this data transmission device usually has the following defects:

[0056] 1. In order to make PU and TR process in parallel, the task needs to be divided. PU needs to write the calculation results into the corresponding storage unit based on different calculation times and different times, which will make program writing and compilation complicated.

[0057] 2. The CPU needs to control the calculation and transmit the calculation results based on the running status of the PU and TR, that is, based on the message of PU calculation completion and TR transmission completion, the CPU is instructed to calculate again or the TR to transmit again, which reduces the efficiency of the CPU, resulting in low CPU data processing efficiency and inability to fully utilize the CPU.

[0058] 3. The calculation result needs to be written into Mem by PU first, and then read out from Mem by TR and sent out, which consumes time and power, resulting in low data processing efficiency.

[0059] The data transceiver device provided by one embodiment of the present application will be described in detail below. In the description of the present invention, it should be noted that the terms "first", "second", "third", and "fourth" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0060] Figure 4 is a schematic structural diagram of a data processing device according to a first embodiment of the present invention.

[0061] like Figure 4 As shown, the data processing device includes a processing module 10 , a calculation module 20 and a Tx transmitter 30 .

[0062] Optionally, the processing module 10 may be a central processing unit CPU, and the computing module 20 may be one or more data processing units PU.

[0063] The processing module 10 is used to send instructions, and the instructions include calculation instructions and transmission instructions.

[0064] The calculation module 20 is used to perform calculations based on the calculation instructions sent by the processing module 10 to generate calculation results.

[0065] The Tx transmitter 30 is connected to the calculation module 20 and is used to send the calculation result generated by the calculation module 20 based on the transmission instruction.

[0066] In one embodiment, the calculation module 20 is further used to cache the calculation results; the Tx transmitter 30 is used to send the calculation results cached in the calculation module 20 based on the transmission instruction.

[0067] In the data processing device provided by the above-mentioned embodiment of the present invention, the processing module sends calculation instructions and transmission instructions, and the Tx transmitter is connected to the calculation module to send the calculation results generated by the calculation module. There is no need for the processing module 10 to participate in the calculation process all the time, which further enables the processing module to efficiently process other data, thereby improving the efficiency of the processing module.

[0068] Figure 5 It is a schematic diagram of the structure of a data processing device provided by one embodiment of the present invention.

[0069] like Figure 5 As shown, the computing module 20 includes an execution unit 21 and an output register group 22 which are connected to each other.

[0070] The execution unit 21 is used to perform calculations based on the calculation instructions sent by the processing module 10 and generate calculation results.

[0071] The output register group 22 is used to cache the calculation result, wherein the output register group 22 includes at least one output register.

[0072] In one embodiment, the output register set 22 is further configured to send a first signal when the output register set 22 is capable of receiving the calculation result. The execution unit 21 sends the calculation result to the output register set 22 based on the first signal.

[0073] It should be noted that in the above embodiment, the output register group 22 can receive the calculation result, which may be that there is storage space in the output register group 22, or the storage space in the output register group 22 is sufficient to accommodate the amount of data obtained by the execution unit 21 performing a calculation.

[0074] In one embodiment, the processing module 10 sends a calculation instruction to the execution unit 21 of the calculation module 20. When the execution unit 21 receives the calculation instruction sent by the processing module 10, it starts to calculate and obtains the calculation result. When the execution unit 21 receives the first signal sent by the output register group, it caches the calculation result in the output register group 22.

[0075] In a preferred embodiment, when the execution unit 21 receives the first signal sent by the output register group 22, it determines whether the first signal indicates that there is storage space in the output register group 22; if so, the execution unit 21 sends the calculation result to the output register group 22.

[0076] In a preferred embodiment, when the execution unit 21 receives the first signal sent by the output register group 22, it determines whether the first signal indicates that the execution unit 21 can store the generated calculation result in the output register group 22; if so, the execution unit 21 sends the calculation result to the output register group 22.

[0077] In one embodiment, if the amount of calculation corresponding to the calculation instruction sent by the processing module 10 is large, a large number of calculation results will be generated, and the execution unit 21 will continuously write the calculation results into the output memory group 22, which may cause the output register group 22 to be unable to receive the calculation results. For example, the cache space in the output register group 22 is full, or the remaining cache space in the output register group 22 is not enough to store the amount of data generated by the execution unit 21 in one calculation.

[0078] The output memory group 22 is also used to send the second signal to the execution unit 21. The execution unit 21 enters a waiting state based on the second signal, that is, the execution unit 21 suspends sending the calculation result based on the second signal. At this time, if the execution unit 21 cannot transmit the calculated data to the output register group 22, the calculation is suspended. Until the execution unit 21 receives the first signal sent by the output register group 22 again, the execution unit 21 sends the calculated result to the output register group and continues to calculate based on the calculation instruction.

[0079] In an optional embodiment, when there is no calculation result in the output register group 22 or the number of calculation results is lower than the first preset value, the output register group generates a third signal, and the Tx transmitter 30 enters a waiting state based on the third signal, that is, the Tx transmitter 30 suspends fetching data from the output register group 22 based on the third signal.

[0080] In an optional embodiment, if there is a calculation result in the output register group 22, or the number of calculation results in the output register group 22 is higher than the second preset value, the output register group 22 sends a fourth signal, and the Tx transmitter 30 takes data from the output register group 22 based on the fourth signal. After receiving the fourth signal, the Tx transmitter 30 can first determine whether the fourth signal indicates that the Tx transmitter 30 can take out the calculation result from the cache of the output register group 22. If so, the Tx transmitter 30 takes out the calculation result from the cache of the output register group 22 and sends it out.

[0081] It can be understood that the first preset value may be the same as or different from the second preset value, and the present invention is not limited thereto.

[0082] It should be noted that the number of calculation results mentioned here can be the amount of data calculated by the execution unit once, or the amount of data calculated multiple times, and the amount of data calculated once can be the calculation result of the entire task, or the calculation result of part of the task, which is not limited here. The first preset value refers to the size of the number of calculation results transmitted at one time. For example, if the first preset value is 8 bits, then when the number of calculation results in the output register group is less than 8 bits, the output register group sends a third signal. Of course, the second preset value is also the same. This embodiment does not limit this, and the preset value can be set according to actual needs.

[0083] Further, the execution unit 21 includes one or more multipliers and adders. The output register group 22 includes one or more output registers, and the number of the output registers set in the output register group 22 is greater than or equal to the number of the multipliers and adders.

[0084] Preferably, the number of output registers provided in the output register group 22 is a multiple of the number of multipliers and adders provided in the execution unit 21 .

[0085] Further preferably, the number of output registers provided in the output register group 22 is at least twice the number of multipliers and adders provided in the execution unit 21 .

[0086] The number of output registers set in the output register group 22 is at least twice the number of multipliers and adders set in the execution unit 21. For example, the output register group 22 includes a first output register and a second output register, and there is one multiplier and adder in the execution unit 21, wherein the first output register is used to store the calculation result generated by the execution unit 21 last time, and the second output register is used to store the calculation result generated by the execution unit 21 this time. In this way, the output register group 22 can provide more storage space, so that the execution unit 21 performs calculations, the output register group 22 stores calculation results, and the Tx transmitter obtains data from the output register group. These three can be carried out in parallel, thereby improving data processing efficiency.

[0087] For another example, the execution unit 21 is provided with two multipliers and adders, and the output calculation result of each multiplier and adder in 2s is 2 bytes. Then, the two multipliers and adders in 2s generate a total of 4 bytes of calculation results. Four output registers can be provided in the output register group, wherein two output registers store the 4-byte calculation results generated by the previous 2s execution unit 21, and the other two output registers store the 4-byte data generated by the next 2s execution unit 21. In this way, the output registers can continuously store the data generated by the execution unit 21, and the execution unit 21 can be efficiently used for data calculation. In this embodiment, one output register is used to store the output result of one calculation of a multiplier and adder, that is, one output register is used to store the calculation result of 2 bytes.

[0088] In one embodiment, the processing module 10 is used to send a calculation instruction to the calculation module 20 and a transmission instruction to the Tx transmitter 30 in no particular order.

[0089] In this embodiment, after receiving the transmission instruction sent by the processing module 10, the Tx transmitter 30 obtains the information of the data to be transmitted according to the setting of the processing module 10. Optionally, the information of the data to be transmitted may be parameters such as the type of data, the size of the data to be transmitted, and the destination address to which the data is to be transmitted.

[0090] When the Tx transmitter 30 receives the transmission instruction and when the Tx transmitter 30 receives the fourth signal sent by the output register group 22 , the Tx transmitter 30 will fetch data from the output register group 22 and send the data to the destination address.

[0091] The embodiment of the present invention takes the computing module 20 as a PU, the processing module 10 as a CPU, the execution unit 21 as an execution unit (EU), and the output register group 22 as a Rout Array as an example to describe in detail the data processing process of the above data processing device:

[0092] The CPU extracts the calculation instruction and the transmission instruction, issues the calculation instruction to the EU in the PU, and sends the transmission instruction to the Tx transmitter 30 in no particular order.

[0093] After receiving the calculation instruction, EU starts the calculation and generates the calculation result.

[0094] When the Rout Array is able to receive the calculation result, a first signal is sent to the EU.

[0095] When EU receives the first signal W_rdy sent by Rout Array, EU determines whether the first signal W_rdy indicates that Rout Array can store the calculation result. If so, the calculation result is stored in Rout Array.

[0096] When there are calculation results in the Rout Array or the number of calculation results is higher than the second preset value, a fourth signal is sent to the Tx transmitter, and the Tx transmitter 30 takes out the calculation results from the cache of the output register group 22 and sends them out.

[0097] When the Rout Array cannot receive the calculation result, a second signal is sent to the EU. After receiving the second signal, the EU stops sending the calculation result to the Rout Array.

[0098] When there is no calculation result in the Rout Array or the number of calculation results is lower than the first preset value, the Rout Array sends a third signal to the Tx transmitter 30, and the Tx transmitter 30 temporarily stops fetching data from the Rout Array.

[0099] And so on, until the Tx transmitter 30 sends out all the calculation results corresponding to the transmission instruction issued by the CPU this time.

[0100] In the data processing device of the above embodiment of the present invention, the CPU sends computing tasks and transmission tasks to the EU and the Tx transmitter in no particular order, and there is no need to split the tasks. The calculation results generated by the EU are stored through the Rout Array, and the Tx transmitter extracts the calculation results from the Rout Array. On the one hand, there is no need to store the calculation results in the Mem, and the Mem will not be occupied, thereby increasing power consumption; on the other hand, the EU calculation and the Tx transmitter sending the calculation results are processed in parallel, without the need for the CPU to be involved all the time, thereby improving the efficiency of the CPU.

[0101] Further, such as Figure 6 As shown, the output register group 22 is, for example, a first in first out output register group (FIFO), W_rdy represents the first signal, Full represents the second signal, Empty represents the third signal, and R_rdy represents the fourth signal. The symbols between the FIFO and the Tx transmitter and between the FIFO and the EU represent inverters. Wt represents the line for the EU to write data, and Rd represents the line for the Tx transmitter to read data.

[0102] In one embodiment, Full means that the FIFO memory is full and has no storage space, indicating that the Rout Array cannot receive the calculation result, and Full = 1 can be set. Empty means that the FIFO is empty and has no calculation result, and Empty = 1 can be set.

[0103] The CPU sends the calculation instructions and transmission instructions to the EU and Tx transmitter in no particular order.

[0104] EU performs calculation based on the calculation instruction and generates a calculation result. If there is no storage space in FIFO, FIFO sends Full=1 to EU. When sending, it passes through the inverter, that is, after passing through the inverter, W_rdy=0 is obtained. Then, when EU receives W_rdy=0, it knows that FIFO is full. EU enters a waiting state until EU receives the first signal sent by FIFO, that is, when it receives the W_rdy=1 signal, it indicates that there is storage space in FIFO, and then controls Wt to ​​write the calculation result into FIFO.

[0105] If the FIFO is empty and no calculation results are stored, the FIFO will send Empty = 1 to the Tx transmitter. When sending or passing through the inverter, that is, after passing through the inverter, R_rdy = 0. At this time, even if the Tx transmitter has received the transmission instruction sent by the CPU, Tx enters the waiting state. Until the FIFO is not empty, the Tx transmitter will receive R_rdy = 1, indicating that there is data to be sent. If the Tx transmitter has received the transmission instruction sent by the CPU, Tx controls Rd to read the data from the FIFO and then send it out.

[0106] In this embodiment, the output register group and the execution unit in the data processing device can realize parallel processing during the calculation process, and the processing module CPU sends the calculation instructions and the transmission instructions to the calculation module and the Tx transmitter in no particular order, so that the parallel processing of data calculation and transmission can be realized, and the calculation results generated by the execution unit are stored in the output register group, and the Tx transmitter extracts the data from the output register group and sends the calculation results. There is no need for the CPU to participate in the calculation all the time, which can enable the CPU to process more other data and improve the effective performance of the CPU.

[0107] Figure 7 is a schematic diagram of a core structure according to an embodiment of the present invention.

[0108] exist Figure 7 The core 1 of the embodiment shown includes the data processing device 100 and the storage module 200 provided in the above embodiment. In the data processing device 100, the computing module 20 is a PU, the processing module 10 is a CPU, the execution unit 21 is an execution unit (EU), and the output register group 22 is a Rout Array. W_rdy represents the first signal, and R_rdy represents the fourth signal.

[0109] The storage module is used to store instructions and data, and the instructions include calculation instructions sent by the CPU to the execution unit 21 and transmission instructions sent to the Tx transmitter 30.

[0110] In the above embodiment, the storage module 200 may be provided with one storage unit.

[0111] exist Figure 7 In the illustrated implementation, the data processing process of the core structure is as follows:

[0112] The CPU extracts the calculation instruction and the transmission instruction in the storage module 200, issues the calculation instruction to the EU in the PU, and sends the transmission instruction to the Tx transmitter 30 in no particular order. After receiving the calculation instruction, the EU starts to calculate and generates a calculation result. When the EU receives the first signal W_rdy sent by the Rout Array, the calculation result is stored in the Rout Array. When the calculation result is stored in the Rout Array or the number of calculation results is higher than the second preset value, the fourth signal R_rdy is sent. The Tx transmitter extracts the calculation result from the Rout Array and sends the calculation result out.

[0113] In the core structure provided by the above implementation, since EU stores the calculation results in Rout Array, and Tx takes the data from Rout Array and sends it, there is no need to split the tasks. When a storage unit is set, parallel processing of calculation and sending can also be achieved, reducing the complexity of program writing and compilation.

[0114] Figure 8 is a schematic diagram of a core structure according to an embodiment of the present invention.

[0115] like Figure 8 As shown, the core 1 includes the data processing device 100 and the storage module 200 provided in the above embodiment. The storage module 200 includes one or more storage units. The data processing device also includes an Rx receiver 40.

[0116] Optionally, the Rx receiver 40 and the Tx receiver 30 may be referred to as a TR data transceiver module.

[0117] The plurality of storage units include a first storage unit 210 , a second storage unit 220 , and a third storage unit 230 .

[0118] The first storage unit 210 is used to store the calculation instructions sent by the CPU to the execution unit 21, and can also be used to store data for the CPU to read and write.

[0119] The second storage unit 220 is used to store the calculation results generated by the calculation module 20, and can also be used to store data for the calculation module 20 to read and write data.

[0120] The third storage unit 230 is used to store data received from the outside by the Rx receiver 40. The Tx transmitter 30 can read data from the third storage unit 230 and then send the data.

[0121] In this embodiment, the output register group 22 and the execution unit 21 in the data processing device 100 can realize parallel processing during the calculation process, and the processing module CPU sends the calculation instructions and transmission instructions to the calculation module and the Tx transmitter in no particular order, which can realize the parallel processing of data calculation and transmission, and by storing the calculation results generated by the execution unit in the output register group, the Tx transmitter takes the data from the output register group and sends it, without the constant participation of the CPU, which can enable the CPU to process more other data and improve the effective performance of the CPU. In addition, when the storage module 200 is provided with three storage units, the TR data transceiver module and the processing module CPU in the data processing device can realize full parallelism when reading the required data during the data processing process. Therefore, in this embodiment, the core has a faster processing speed and better timeliness.

[0122] An embodiment of the present invention further provides a chip, comprising one or more core structures of the above embodiments.

[0123] An embodiment of the present invention further provides a card board, comprising one or more chips of the above embodiments.

[0124] An embodiment of the present invention further provides an electronic device, comprising one or more card boards of the above embodiments.

[0125] Fig. 9 It is a flow chart of a data processing method provided by one embodiment of the present invention.

[0126] like Fig. 9 As shown, the method includes steps S101 to S103:

[0127] Step S101: The processing module sends instructions, which include calculation instructions and transmission instructions.

[0128] In a preferred embodiment, the processing module sends a calculation instruction to the calculation module and a transmission instruction to the Tx transmitter in no particular order.

[0129] Step S102: The calculation module performs calculations based on the calculation instructions sent by the processing module and generates calculation results.

[0130] In one embodiment, the computing module includes an execution unit and an output register set.

[0131] The step in which the calculation module performs calculations based on the calculation instructions sent by the processing module and generates calculation results, and the step in which the calculation module sends a transmission instruction based on the calculation results includes:

[0132] The execution unit performs calculation based on the calculation instruction sent by the processing module and generates the calculation result.

[0133] The output register group caches the calculation result, and the output register group includes at least one output register.

[0134] In one embodiment, a first signal is sent to the execution unit when the output register set is capable of receiving the calculation result.

[0135] Step S103: The Tx transmitter is connected to the output register group, and sends the calculation result generated by the calculation module based on the transmission instruction.

[0136] In an optional embodiment, the steps further include:

[0137] When the execution unit receives the first signal sent by the output register group, the execution unit will determine whether the first signal indicates that the output register group can store the calculation result. If so, the calculation result is stored in the output register group.

[0138] When there are calculation results in the output register group or the number of calculation results is higher than the second preset value, a fourth signal is sent to the Tx transmitter, and the Tx transmitter takes out the calculation results from the cache of the output register group and sends them out.

[0139] When the output register group cannot receive the calculation result, a second signal is sent to the execution unit. After receiving the second signal, the execution unit stops sending the calculation result to the output register group.

[0140] When there is no calculation result in the output register group or the number of calculation results is lower than the first preset value, the output register group sends a third signal to the Tx transmitter, and the Tx transmitter temporarily stops fetching data from the output register group.

[0141] The above technical solution of the present invention has the following beneficial technical effects:

[0142] (1) The data processing method provided by the present invention sends calculation instructions and transmission instructions through the processing module, the calculation module generates calculation results, and the Tx transmitter is connected to the calculation module to send the calculation results generated by the calculation module. The processing module does not need to participate in the calculation process and the transmission of the calculation results all the time, which further enables the processing module to efficiently process other data, thereby improving the efficiency of the processing module.

[0143] (2) In the data processing method provided in the embodiment of the present invention, the processing module sends the calculation instruction and the transmission instruction in no particular order, and the transmission instruction instructs the Tx transmitter to send the calculation result generated by the calculation module, so that the calculation module executes the calculation and the Tx transmitter sends the calculation result in parallel, thereby improving the utilization rate of the calculation module.

[0144] (3) The data processing device provided in the embodiment of the present invention does not require the processing module to split the task, thereby reducing the difficulty of program writing and compilation.

[0145] It should be understood that the above specific embodiments of the present invention are only used to illustrate or explain the principles of the present invention, and do not constitute a limitation of the present invention. Therefore, any modifications, equivalent substitutions, improvements, etc. made without departing from the spirit and scope of the present invention should be included in the protection scope of the present invention. In addition, the appended claims of the present invention are intended to cover all changes and modifications that fall within the scope and boundaries of the appended claims, or the equivalent forms of such scope and boundaries.

Claims

1. A data processing device, characterized in that: include: A processing module (10) is used to send instructions, the instructions including: a calculation instruction and a transmission instruction, the processing module sends the calculation instruction and the transmission instruction in no particular order; A calculation module (20), connected to the processing module (10), configured to receive the calculation instruction, perform calculation based on the calculation instruction and generate a calculation result; A Tx transmitter (30) is connected to the processing module (10) and is used to receive the transmission instruction, and is connected to the calculation module and is used to send the calculation result generated by the calculation module (20) based on the transmission instruction.

2. The data processing device according to claim 1, characterized in that: The calculation module (20) is further used to cache the calculation result; The Tx transmitter (30) is used to transmit the calculation result cached in the calculation module (20) based on the transmission instruction.

3. The data processing device according to claim 1 or 2, characterized in that: The calculation module (20) comprises: An execution unit (21), configured to execute calculation based on the calculation instruction and generate the calculation result; An output register group (22) is used to cache the calculation result, and the output register group (22) includes at least one output register.

4. The data processing device according to claim 3, characterized in that: The output register group (22) is further configured to send a first signal when the output register group (22) is capable of receiving the calculation result, and the execution unit (21) sends the calculation result to the output register group (22) based on the first signal; and / or The output register group (22) is further configured to send a second signal when the output register group (22) cannot receive the calculation result, and the execution unit (21) suspends sending the calculation result based on the second signal.

5. The data processing device according to claim 4, characterized in that: The output register group (22) is further configured to send a third signal when the number of the calculation results stored in the output register group (22) is lower than a first preset value, and the Tx transmitter (30) suspends fetching data from the output register group (22) based on the third signal; and / or The output register group (22) is further used to send a fourth signal when the number of the calculation results stored in the output register (22) is higher than a second preset value, and the Tx transmitter (30) fetches data from the output register group (22) based on the fourth signal.

6. The data processing device according to any one of claims 4 or 5, characterized in that: The execution unit (21) includes one or more multipliers and adders; The number of the output registers in the output register group (22) is greater than or equal to the number of the multipliers and adders; preferably, the number of the output registers provided in the output register group (22) is a multiple of the number of the multipliers and adders provided in the execution unit (21); further preferably, the number of the output registers provided in the output register group (22) is at least twice the number of the multipliers and adders provided in the execution unit (21).

7. The data processing device according to any one of claims 1 to 2, 4 to 5, characterized in that: The processing module (10) is used to send the calculation instruction and the transmission instruction in no particular order.

8. A nuclear structure, characterized in that comprising a data processing device and a storage module as described in any one of claims 1 to 7; The storage module is used to store instructions and data, and the instructions include the calculation instructions and the transmission instructions.

9. A chip, characterized in that: Comprising one or more core structures as claimed in claim 8.

10. A data processing method, characterized in that: include: The processing module (10) sends the calculation instruction and the transmission instruction in no particular order; The calculation module (20) receives the calculation instruction, performs calculation based on the calculation instruction and generates a calculation result; A Tx transmitter (30) receives the transmission instruction and transmits the calculation result generated by the calculation module (20) based on the transmission instruction.

Citation Information

Patent Citations

  • Arithmetic logic unit using asynchronous circuit to implement

    CN101303643A

  • High-speed operation method and system for spacecraft

    CN102346719A

  • Battle field real-time information sharing method based on cloud computing

    CN103118096A

  • High-performance heterogeneous computing platform based on x86 architecture processor and FPGA (Field Programmable Gate Array)

    CN104657330A

  • A data processing method and a related device

    CN108984235A