5G downlink RB scheduling device and method for DSP
By designing devices and methods for 5G downlink RB scheduling in DSP, using RB type index values to obtain fill information, the problem of low DSP processing efficiency is solved, and unified processing and efficient flow processing of 5G downlink channel data is realized.
Patent Information
- Application Number
- CN202411671609.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-21
- Publication Date
- 2025-02-18
AI Technical Summary
The prior art is difficult to improve the processing efficiency of DSP, especially in 5G downlink RB scheduling.
Provide a 5G downlink RB scheduling device and method for DSP. Through the storage module, acquisition module, data processing module and output module, RB filling information is obtained from the local storage space using the RB type index value, and the address of the acquisition data is determined. Fill the value to realize unified processing and flow processing.
This method allows 5G downlink channel data to be processed uniformly using a piece of code, reducing the number of calls to start and end of the DSP processing flow, and improving the processing efficiency of DSP.
Smart Images

Figure CN119485712A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of communication technology, and in particular to a 5G downlink RB scheduling device and method for a vector processor (DSP). Background Art
[0002] Compared with previous generations of communication technology, 5G technology has made significant improvements in several aspects. The 5G baseband subsystem needs to handle larger bandwidth and higher MIMO levels, all of which require lower latency characteristics, so the processing platform must be able to handle a large amount of baseband processing.
[0003] The advantages of vector operations mainly include efficient data processing capabilities, reduced code size, and improved code readability and maintainability. Vectorized operations allow programs to perform calculations at a higher level of abstraction, which not only helps to simplify the code structure, but also effectively utilizes the parallel processing capabilities of modern processors. The disadvantage of vector operations is that they can only process the same data in batches or can only perform the same operation on multiple blocks of data according to certain rules. Summary of the invention
[0004] The technical problem to be solved by the present invention is to provide a 5G downlink RB scheduling device and method for DSP to improve the processing efficiency of DSP.
[0005] The technical solution adopted by the present invention to solve the technical problem is: to provide a 5G downlink RB scheduling device for DSP, including:
[0006] A storage module, used to store N types of RB type filling information, where N is the number of all possible combinations of data types that may be placed in 12 REs in an RB;
[0007] An acquisition module, used to acquire RB scheduling information, wherein the RB scheduling information includes a data arrangement index and addresses of different data types;
[0008] A data processing module, used to obtain corresponding RB type filling information from the storage module according to the data arrangement index in the RB scheduling information, and obtain data in corresponding addresses from the RB type filling information according to addresses of different data types, and fill them into corresponding RE memory output space;
[0009] The output module is used to output the data in the corresponding RE memory output space.
[0010] The RB type filling information includes data source information and location information, wherein the data source information is used to determine the address for acquiring data; and the location information is used to determine the offset for acquiring data.
[0011] The storage module is the internal storage space of DSP.
[0012] The RB scheduling information is stored in a storage space accessible to both the CPU and the DSP.
[0013] The technical solution adopted by the present invention to solve the technical problem is: to provide a 5G downlink RB scheduling method, using the above-mentioned 5G downlink RB scheduling device for DSP, comprising the following steps:
[0014] Obtain RB scheduling information, and extract the data arrangement index value and the address of the data type from the RB scheduling information;
[0015] Obtaining corresponding RB type filling information from the storage module according to the extracted data arrangement index value;
[0016] According to the address of the data type in the RB scheduling information, obtain the data in the corresponding address from the RB type filling information, and fill it into the corresponding RE memory output space;
[0017] Determine whether it is the last RB. If it is the last RB, end the scheduling; if it is not the last RB, return to the step of obtaining RB scheduling information for the next RB.
[0018] Before obtaining the RB scheduling information, the method further includes: encoding and storing the index value of the RB scheduled by the CPU, specifically:
[0019] For the RB to be scheduled, an index value is generated for each RB, wherein the index value of each RB is different;
[0020] The prepared RB filling data is stored in the storage module, and the address value of the data to be filled in each RB is assigned to a preset position to form N types of RB filling information.
[0021] Beneficial Effects
[0022] Due to the adoption of the above-mentioned technical scheme, compared with the prior art, the present invention has the following advantages and positive effects: the present invention obtains RB filling information from the local storage space through the RB type index value, and determines the address and specific filling value of the acquired data through the data source information and location information in the RB filling information. This method enables 5G downlink channel data to be uniformly processed using one code without channel processing. The DSP can realize data discharge only according to the RB index value in each cycle, and when the DSP processes the RB information, one cycle can process all RBs within the entire bandwidth of one or more cells without looping by UE. The DSP cycle can realize pipeline processing with a large number of cycles, reducing the number of calls to start and end the DSP processing pipeline, ensuring that the DSP is in a high-efficiency operating state most of the time. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] Figure 1 It is a structural diagram of a 5G downlink RB scheduling device for DSP according to an embodiment of the present invention;
[0024] Figure 2 Format diagram of RB type filling information in an embodiment of the present invention;
[0025] Figure 3 It is a schematic diagram of a 5G downlink RB scheduling method in an implementation manner of the present invention. DETAILED DESCRIPTION
[0026] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms fall within the scope limited by the appended claims of the application equally.
[0027] The embodiments of the present invention relate to a 5G downlink RB scheduling device for DSP, such as Figure 1 As shown, it includes: a storage module, an acquisition module, a data processing module and an output module.
[0028] The storage module is used to store N types of RB type filling information, where N is the number of all possible combinations of data types that can be placed in 12 REs within an RB.
[0029] In the 5G protocol, each RB consists of 12 REs, and PDSCH DATA, PDSCH DMRS, CSI-RS, PTRS and other data can be placed in REs. Each RB can have one or more types of data at the same time. Therefore, this implementation method counts the number N of all possible combinations of data types that can be placed in 12 REs in an RB, and then generates N types of RB type filling information in the internal storage space of the DSP (i.e., the storage module) for storage. The format of a single information is as follows: Figure 2 As shown, Figure 2 In the figure, A and B respectively represent two addresses (ie, data source information) of two data sources (such as PDSCH DATA and CSI-RS) in the current RB, and offset is the position information, wherein offset0 represents that the offset address is 0 data, and offsetx represents that the offset address is x data.
[0030] The acquisition module is used to obtain RB scheduling information, and the RB scheduling information includes a data arrangement index and addresses of different data types. In this implementation, each RB scheduling information defined by the Cpu includes the following information: a data arrangement index [0-(N-1))], and the address of data type A, the address of data type B, and the address of data type C. Data types ABC can represent different data addresses according to different RB indexes, and can be used in full or in part. For example, when PDSCH DATA, CSI-RS, and PTRS exist in an RB at the same time, A represents the storage address of the PDSCH DATA of this RB, B represents the storage address of the CSI-RS of this RB, and C represents the storage address of the PTRS of this RB. The RB scheduling information of the Cpu can be stored in the interface memory of the Cpu and DSP, and the interface memory refers to a storage space that can be accessed by both the Cpu and the Dsp.
[0031] A data processing module is used to obtain the corresponding RB type filling information from the storage module according to the data arrangement index in the RB scheduling information, and obtain the data in the corresponding address from the RB type filling information according to the addresses of different data types, and fill it into the corresponding RE memory output space.
[0032] In this implementation mode, the DSP can process the RB data scheduled by the CPU in a loop according to the RB scheduling information of the CPU, and process one RB block or multiple RB blocks in a loop each time. When processing each RB block, data is obtained from address A, address B, and address C in the RB scheduling information, and then placed in the DSP memory where the output address of the RB is located.
[0033] The output module is used to output the data in the corresponding RE memory output space.
[0034] The following is a specific example to illustrate the calling method of the 5G downlink RB scheduling device for DSP in this embodiment.
[0035] In this embodiment, there are three formats for storing PDSCH channel data in the 5G downlink channel, namely: (1) PDSCH channel RB format 0, 12 REs in RB, even numbers are DMRS, and odd numbers are DATA; (2) PDSCH channel RB format 1, 12 REs in RB, all storing DATA; (3) PDSCH channel RB format 2, indexes 0, 3, 6, and 9 of the 12 REs in RB are DMRS, and indexes 1, 2, 4, 5, 7, 8, 10, and 11 are DATA.
[0036] like Figure 3 As shown, before scheduling, the index values of the RBs scheduled by the three CPUs need to be encoded and stored, including:
[0037] 1) The CPU generates an index value of 1 for RB0 scheduling format 1, an index value of 2 for RB1 scheduling RB format 2, and an index value of 0 for RB3 scheduling RB format 0.
[0038] 2) The CPU stores the prepared RB filling data (DATA and DMRS) in the storage space of the DSP, and then assigns the address values of the data to be filled in RB0, RB1 and RB2 to address A and address B of RB0, RB1 and RB2 respectively.
[0039] When scheduling, the DSP runs the processing code to schedule the required RB data. The specific steps are as follows:
[0040] 1) Initialize the RB sequence number to rbIdx=0.
[0041] 2) DSP obtains the RB scheduling information with RB number rbIdx from the interface memory between CPU and DSP, and extracts the RB type index value n and the address of the data type therefrom.
[0042] 3) The DSP obtains the RB type filling information from the storage space of the DSP according to the RB type index value n, thereby obtaining the corresponding data source information and location information.
[0043] 4) DSP obtains data according to the corresponding data source information and location information, and fills the data into the corresponding RE memory output space according to the emission format.
[0044] 5) Determine whether it is the last RB according to rbIdx. If it is the last RB, exit the loop and the DSP scheduling process ends. If it is not the last RB, increase the value of rbIdx by 1 and jump to step 2) to continue scheduling the next RB.
[0045] It is not difficult to find that the present invention obtains RB filling information from the local storage space through the RB type index value, and determines the address and specific filling value of the acquired data through the data source information and location information in the RB filling information. This method enables 5G downlink channel data to be uniformly processed using one code without channel processing. The DSP can realize data discharge only according to the RB index value in each cycle, and when the DSP processes the RB information, one cycle can process all RBs within the entire bandwidth of a cell or multiple cells without looping according to the UE. The DSP cycle can realize pipeline processing with a large number of cycles, reducing the number of calls to start and end the DSP processing pipeline, ensuring that the DSP is in a high-efficiency operating state most of the time.
Claims
1. A 5G downlink RB scheduling device for DSP, characterized in that: include: A storage module, used to store N types of RB type filling information, where N is the number of all possible combinations of data types that may be placed in 12 REs in an RB; An acquisition module, used to acquire RB scheduling information, wherein the RB scheduling information includes a data arrangement index and addresses of different data types; A data processing module, used to obtain corresponding RB type filling information from the storage module according to the data arrangement index in the RB scheduling information, and obtain data in corresponding addresses from the RB type filling information according to addresses of different data types, and fill them into corresponding RE memory output space; The output module is used to output the data in the corresponding RE memory output space.
2. The 5G downlink RB scheduling device for DSP according to claim 1, characterized in that: The RB type filling information includes data source information and location information, and the data source information is used to determine the address for obtaining data; The position information is used to determine the offset for acquiring data.
3. The 5G downlink RB scheduling device for DSP according to claim 1, characterized in that: The storage module is the internal storage space of DSP.
4. The 5G downlink RB scheduling device for DSP according to claim 1, characterized in that: The RB scheduling information is stored in a storage space accessible to both the CPU and the DSP.
5. A 5G downlink RB scheduling method, characterized in that: The 5G downlink RB scheduling device for DSP as claimed in any one of claims 1 to 4 is adopted, comprising the following steps: Obtain RB scheduling information, and extract the data arrangement index value and the address of the data type from the RB scheduling information; Obtaining corresponding RB type filling information from the storage module according to the extracted data arrangement index value; According to the address of the data type in the RB scheduling information, obtain the data in the corresponding address from the RB type filling information, and fill it into the corresponding RE memory output space; Determine whether it is the last RB. If it is the last RB, end the scheduling; if it is not the last RB, return to the step of obtaining RB scheduling information for the next RB.
6. The 5G downlink RB scheduling method for DSP according to claim 5, characterized in that: Before obtaining the RB scheduling information, the method further includes: encoding and storing the index value of the RB scheduled by the CPU, specifically: For the RB to be scheduled, an index value is generated for each RB, wherein the index value of each RB is different; The prepared RB filling data is stored in the storage module, and the address value of the data to be filled in each RB is assigned to a preset position to form N types of RB filling information.
