CN116627894A - A medium access control layer, communication method and system - Google Patents
A medium access control layer, communication method and system Download PDFInfo
- Publication number
- CN116627894A CN116627894A CN202310893529.XA CN202310893529A CN116627894A CN 116627894 A CN116627894 A CN 116627894A CN 202310893529 A CN202310893529 A CN 202310893529A CN 116627894 A CN116627894 A CN 116627894A
- Authority
- CN
- China
- Prior art keywords
- communication interface
- data
- chip communication
- inter
- media access
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F15/00—Digital computers in general; Data processing equipment in general
- G06F15/76—Architectures of general purpose stored program computers
- G06F15/78—Architectures of general purpose stored program computers comprising a single central processing unit
- G06F15/7807—System on chip, i.e. computer system on a single chip; System in package, i.e. computer system on one or more chips in a single package
- G06F15/7825—Globally asynchronous, locally synchronous, e.g. network on chip
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Communication Control (AREA)
Abstract
Description
技术领域technical field
本申请涉及集成电路技术领域,特别是涉及一种介质访问控制层、通信方法和系统。The present application relates to the technical field of integrated circuits, in particular to a medium access control layer, a communication method and a system.
背景技术Background technique
为了应对大芯片制造问题,Chiplet技术应运而生,Chiplet又称“小芯片”或“芯粒”,它是一种功能电路块。Chiplet技术通过把芯片拆分为较小的功能块分别进行制造,再通过Chiplet接口协议和先进封装技术实现功能块之间的互联,从而规避大芯片的可制造性问题、减少专用集成电路(Application Specific Integrated Circuit,ASIC)设计的高成本、长工期的问题。Chiplet技术已经成为一项热门技术。In order to deal with the problem of large chip manufacturing, Chiplet technology came into being. Chiplet, also known as "small chip" or "core particle", is a functional circuit block. Chiplet technology divides the chip into smaller functional blocks for manufacturing, and then realizes the interconnection between functional blocks through the Chiplet interface protocol and advanced packaging technology, thereby avoiding the manufacturability problem of large chips and reducing the number of application-specific integrated circuits (ASICs). Specific Integrated Circuit, ASIC) design problems of high cost and long construction period. Chiplet technology has become a hot technology.
此外,近年来众核架构应用越来越广泛,众核设计的核心就是充分开发片上高性能算力,这也是高性能计算的必然发展趋势。现今拥有的众核处理器,其中包含大量独立的内核和紧密耦合的内存,而它们之间通过高速片上网络(Network on Chip,NoC)基础设施进行互连。从开放式系统互联模型(Open System Interconnect,OSI)的角度来看,芯粒片间接口协议属于物理层协议,而芯粒内部的通信网络则属于网络层组件,两个层级在模型上需要数据链路层作为中介,这样的组件需要提供介质访问控制层(MediaAccessControl,MAC)。In addition, many-core architectures have become more and more widely used in recent years. The core of many-core design is to fully develop on-chip high-performance computing power, which is also an inevitable development trend of high-performance computing. Today's many-core processors contain a large number of independent cores and tightly coupled memory, which are interconnected by a high-speed Network on Chip (NoC) infrastructure. From the perspective of Open System Interconnect (OSI), the inter-chip interface protocol belongs to the physical layer protocol, while the internal communication network of the chip belongs to the network layer components. The two layers require data in the model. Link layer as an intermediary, such a component needs to provide a media access control layer (MediaAccessControl, MAC).
MAC在OSI中属于数据链路层中的一个子层,但它具有与物理层耦合的特点,不同的物理介质对应了完全不同的MAC协议和电路模块设计。现存的用于其他场景的MAC模块(例如IEEE-802. 3所定义以太网MAC)几乎无法直接用于芯粒片间通信场景。同时,市面上具有各种不同规格的片间通信接口,这些接口的位宽、通信速率和时钟频率都有很大的差别,而且这些位于芯粒四周的通信接口一般与芯粒内部的电路处于不同的时钟域,目前提供的介质访问控制层无法兼容市面上的各种不同规格的片间通信接口,介质访问控制层无法复用。MAC belongs to a sublayer in the data link layer in OSI, but it has the characteristics of coupling with the physical layer. Different physical media correspond to completely different MAC protocols and circuit module designs. Existing MAC modules used in other scenarios (such as Ethernet MAC defined by IEEE-802. 3) can hardly be directly used in inter-chip communication scenarios. At the same time, there are various inter-chip communication interfaces with different specifications on the market. The bit width, communication rate and clock frequency of these interfaces are very different, and these communication interfaces located around the chip are generally in the same position as the internal circuits of the chip. For different clock domains, the currently provided media access control layer cannot be compatible with various inter-chip communication interfaces of different specifications on the market, and the media access control layer cannot be reused.
针对相关技术中,介质访问控制层无法兼容市面上的各种不同规格的片间通信接口,导致无法复用的问题,目前尚未提出有效的解决方案。In the related technology, the medium access control layer is not compatible with various inter-chip communication interfaces of different specifications on the market, resulting in the problem of inability to be reused, and no effective solution has been proposed yet.
发明内容Contents of the invention
基于此,有必要针对上述技术问题,提供一种介质访问控制层、通信方法和系统。Based on this, it is necessary to provide a media access control layer, a communication method and a system for the above technical problems.
第一方面,本申请实施例提供了一种介质访问控制层,所述介质访问控制层连接在芯粒的片内通信接口和片间通信接口之间,所述介质访问控制层包括:In the first aspect, the embodiment of the present application provides a media access control layer, the media access control layer is connected between the on-chip communication interface and the inter-chip communication interface of the chip, and the media access control layer includes:
数据链路协议桥模块,用于建立所述片内通信接口和所述片间通信接口之间的逻辑数据链路;A data link protocol bridge module, configured to establish a logical data link between the on-chip communication interface and the inter-chip communication interface;
与所述数据链路协议桥模块连接的介质访问控制器模块,用于接收所述数据链路协议桥模块传输的片内通信接口数据,将所述片内通信接口数据编码为所述片间通信接口的规格所对应的帧数据,并将所述帧数据传输至所述片间通信接口;A media access controller module connected to the data link protocol bridge module, configured to receive the on-chip communication interface data transmitted by the data link protocol bridge module, and encode the on-chip communication interface data into the inter-chip Frame data corresponding to the specification of the communication interface, and transmitting the frame data to the inter-chip communication interface;
与所述数据链路协议桥模块连接的配置模块,用于接收并解析所述数据链路协议桥模块传输的配置包,以配置所述片间通信接口。A configuration module connected to the data link protocol bridge module is used to receive and parse the configuration packet transmitted by the data link protocol bridge module to configure the inter-chip communication interface.
在其中一个实施例中,所述介质访问控制器模块包括第一帧编码器,用于根据所述片间通信接口对应的位宽和时钟频率,将所述片内通信接口数据编码为所述片间通信接口的规格所对应的帧数据。In one of the embodiments, the media access controller module includes a first frame encoder, configured to encode the on-chip communication interface data into the Frame data corresponding to the specification of the inter-chip communication interface.
在其中一个实施例中,所述介质访问控制器模块还包括第一仲裁器,用于根据所述片内通信接口数据对应的子网优先级以及所述片内通信接口数据对应的数据链路协议桥模块优先级对所述片内通信接口数据进行仲裁。In one of the embodiments, the media access controller module further includes a first arbiter, configured to select the subnet priority corresponding to the on-chip communication interface data and the data link corresponding to the on-chip communication interface data. The priority of the protocol bridge module arbitrates the data of the on-chip communication interface.
在其中一个实施例中,所述数据链路协议桥模块包括:In one of the embodiments, the data link protocol bridge module includes:
第一数据处理模块,用于接收并处理所述片内通信接口数据,并将处理后的片内通信接口数据传输至所述介质访问控制器模块;A first data processing module, configured to receive and process the on-chip communication interface data, and transmit the processed on-chip communication interface data to the media access controller module;
第二仲裁器,用于根据所述片内通信接口数据对应的子网是否持有下游信用以及所述片内通信接口数据对应的子网优先级,对所述片内通信接口数据进行仲裁。The second arbiter is configured to arbitrate the on-chip communication interface data according to whether the subnet corresponding to the on-chip communication interface data holds downstream credits and the subnet priority corresponding to the on-chip communication interface data.
在其中一个实施例中,所述数据处理模块包括:In one of the embodiments, the data processing module includes:
第一存储器,用于存放所述片内通信接口数据;The first memory is used to store the on-chip communication interface data;
复用器,用于根据所述第二仲裁器获得的仲裁结果,输出对应的片内通信接口数据;a multiplexer, configured to output corresponding on-chip communication interface data according to the arbitration result obtained by the second arbiter;
第二帧编码器,用于将所述片内通信接口数据编码为所述片间通信接口的规格所对应的帧数据。The second frame encoder is configured to encode the intra-chip communication interface data into frame data corresponding to the specification of the inter-chip communication interface.
在其中一个实施例中,所述数据链路协议桥模块还包括第二数据处理模块,用于接收并处理所述介质访问控制器模块传输的片间通信接口数据,并将处理后的所述片间通信接口数据传输至所述片内通信接口。In one of the embodiments, the data link protocol bridge module further includes a second data processing module, configured to receive and process the inter-chip communication interface data transmitted by the media access controller module, and process the processed The inter-chip communication interface data is transmitted to the on-chip communication interface.
在其中一个实施例中,所述第二数据处理模块包括:帧解码器,解复用器,第二存储器,所述帧解码器,用于将所述介质访问控制器模块传输过来的片间通信接口数据进行解码;所述解复用器,用于根据所述片间接口数据的来源将所述片间通信接口数据输入至所述第二存储器进行存放。In one of the embodiments, the second data processing module includes: a frame decoder, a demultiplexer, a second memory, the frame decoder, an inter-chip for transmitting the media access controller module The communication interface data is decoded; the demultiplexer is configured to input the inter-chip communication interface data into the second memory for storage according to the source of the inter-chip interface data.
第二方面,本申请实施例还提供了一种基于介质访问控制层的通信方法,所述方法包括:In the second aspect, the embodiment of the present application also provides a medium access control layer-based communication method, the method including:
接收所述数据链路协议桥模块传输的片内通信接口数据;receiving the on-chip communication interface data transmitted by the data link protocol bridge module;
将所述片内通信接口数据编码为所述片间通信接口的规格所对应的帧数据;Encoding the on-chip communication interface data into frame data corresponding to the specification of the inter-chip communication interface;
将所述帧数据传输至所述片间通信接口。transmitting the frame data to the inter-chip communication interface.
在其中一个实施例中,所述方法还包括:In one embodiment, the method also includes:
接收所述介质访问控制模块传输的片间通信接口数据;receiving the inter-chip communication interface data transmitted by the media access control module;
将所述片间通信接口数据解码并根据所述片间通信接口数据对应的来源传输至所述片内通信接口。The inter-chip communication interface data is decoded and transmitted to the on-chip communication interface according to the source corresponding to the inter-chip communication interface data.
第三方面,本申请实施例还提供了芯粒片间通信系统,包括第一介质访问控制层以及第二介质访问控制层,所述第一介质访问控制层与所述第二介质访问控制层通信连接;所述第一介质访问控制层和/或所述第二介质访问控制层实现如上述第二方面所述的方法。In a third aspect, the embodiment of the present application further provides an inter-chip communication system, including a first media access control layer and a second media access control layer, the first media access control layer and the second media access control layer A communication connection: the first media access control layer and/or the second media access control layer implements the method described in the second aspect above.
上述介质访问控制层、通信方法和系统,通过数据链路协议桥模块,用于建立所述片内通信接口和所述片间通信接口之间的逻辑数据链路;与所述数据链路协议桥模块连接的介质访问控制器模块,用于接收所述数据链路协议桥模块传输的片内通信接口数据,将所述片内通信接口数据编码为所述片间通信接口的规格所对应的帧数据,并将所述帧数据传输至所述片间通信接口;与所述数据链路协议桥模块连接的配置模块,用于接收并解析所述数据链路协议桥模块传输的配置包,以配置所述片间通信接口。解决了介质访问控制层无法兼容市面上的各种不同规格的片间通信接口,导致介质访问控制层无法复用的问题。The above media access control layer, communication method and system are used to establish a logical data link between the intra-chip communication interface and the inter-chip communication interface through the data link protocol bridge module; and the data link protocol The media access controller module connected to the bridge module is used to receive the on-chip communication interface data transmitted by the data link protocol bridge module, and encode the on-chip communication interface data into the specification corresponding to the inter-chip communication interface. frame data, and transmit the frame data to the inter-chip communication interface; the configuration module connected to the data link protocol bridge module is used to receive and analyze the configuration packet transmitted by the data link protocol bridge module, to configure the inter-chip communication interface. It solves the problem that the media access control layer cannot be compatible with various inter-chip communication interfaces of different specifications on the market, resulting in the problem that the media access control layer cannot be reused.
本申请的一个或多个实施例的细节在以下附图和描述中提出,以使本申请的其他特征、目的和优点更加简明易懂。The details of one or more embodiments of the application are set forth in the accompanying drawings and the description below, so as to make other features, objects, and advantages of the application more comprehensible.
附图说明Description of drawings
此处所说明的附图用来提供对本申请的进一步理解,构成本申请的一部分,本申请的示意性实施例及其说明用于解释本申请,并不构成对本申请的不当限定。在附图中:The drawings described here are used to provide a further understanding of the application and constitute a part of the application. The schematic embodiments and descriptions of the application are used to explain the application and do not constitute an improper limitation to the application. In the attached picture:
图1是根据本申请实施例的介质访问控制层结构图;FIG. 1 is a structural diagram of a media access control layer according to an embodiment of the present application;
图2是根据本申请另一实施例的介质访问控制层结构图;FIG. 2 is a structural diagram of a media access control layer according to another embodiment of the present application;
图3是根据本申请实施例的介质访问控制器模块的结构图;FIG. 3 is a structural diagram of a media access controller module according to an embodiment of the present application;
图4是根据本申请实施例的片间通接口对应的帧格式示意图;4 is a schematic diagram of a frame format corresponding to an inter-chip communication interface according to an embodiment of the present application;
图5是根据本申请另一实施例的介质访问控制器模块的结构图;FIG. 5 is a structural diagram of a media access controller module according to another embodiment of the present application;
图6是根据本申请实施例的数据链路协议桥模块的结构图;6 is a structural diagram of a data link protocol bridge module according to an embodiment of the present application;
图7是根据本申请实施例的第一数据处理模块的结构图;FIG. 7 is a structural diagram of a first data processing module according to an embodiment of the present application;
图8是根据本申请实施例的片内通接口对应的帧格式示意图;FIG. 8 is a schematic diagram of a frame format corresponding to an intra-chip communication interface according to an embodiment of the present application;
图9是根据本申请实施例的数据链路协议桥结构图;FIG. 9 is a structural diagram of a data link protocol bridge according to an embodiment of the present application;
图10是根据本申请实施例的基于介质访问控制层的通信方法流程示意图;FIG. 10 is a schematic flowchart of a communication method based on a media access control layer according to an embodiment of the present application;
图11是根据本申请实施例的芯粒片间通信系统结构结构图。FIG. 11 is a structural diagram of an inter-chip communication system according to an embodiment of the present application.
其中,11、数据链路协议桥模块;12、介质访问控制器模块;13、配置模块;1110、数据链路协议桥;1210、第一帧编码器;1220、第一仲裁器;21、第一数据处理模块;22、第二仲裁器;31、第一存储器;32、复用器;33、第二帧编码器;23、第二数据处理模块;34、帧解码器;35、解复用器;36、第二存储器;51、第一介质访问控制层;52、第二介质访问控制层。Among them, 11. Data link protocol bridge module; 12. Media access controller module; 13. Configuration module; 1110, Data link protocol bridge; 1210, First frame encoder; 1220, First arbiter; 21, The first A data processing module; 22, a second arbiter; 31, a first memory; 32, a multiplexer; 33, a second frame encoder; 23, a second data processing module; 34, a frame decoder; 35, demultiplexing 36, the second memory; 51, the first media access control layer; 52, the second media access control layer.
具体实施方式Detailed ways
为了使本申请的目的、技术方案及优点更加清楚明白,以下结合附图及实施例,对本申请进行描述和说明。应当理解,此处所描述的具体实施例仅仅用以解释本申请,并不用于限定本申请。基于本申请提供的实施例,本领域普通技术人员在没有作出创造性劳动的前提下所获得的所有其他实施例,都属于本申请保护的范围。In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be described and illustrated below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application, and are not intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by persons of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.
显而易见地,下面描述中的附图仅仅是本申请的一些示例或实施例,对于本领域的普通技术人员而言,在不付出创造性劳动的前提下,还可以根据这些附图将本申请应用于其他类似情景。此外,还可以理解的是,虽然这种开发过程中所作出的努力可能是复杂并且冗长的,然而对于与本申请公开的内容相关的本领域的普通技术人员而言,在本申请揭露的技术内容的基础上进行的一些设计,制造或者生产等变更只是常规的技术手段,不应当理解为本申请公开的内容不充分。Obviously, the accompanying drawings in the following description are only some examples or embodiments of the present application, and those skilled in the art can also apply the present application to other similar scenarios. In addition, it can also be understood that although such development efforts may be complex and lengthy, for those of ordinary skill in the art relevant to the content disclosed in this application, the technology disclosed in this application Some design, manufacturing or production changes based on the content are just conventional technical means, and should not be understood as insufficient content disclosed in this application.
在本申请中提及“实施例”意味着,结合实施例描述的特定特征、结构或特性可以包含在本申请的至少一个实施例中。在说明书中的各个位置出现该短语并不一定均是指相同的实施例,也不是与其它实施例互斥的独立的或备选的实施例。本领域普通技术人员显式地和隐式地理解的是,本申请所描述的实施例在不冲突的情况下,可以与其它实施例相结合。Reference in this application to an "embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The occurrences of this phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is understood explicitly and implicitly by those of ordinary skill in the art that the embodiments described in this application can be combined with other embodiments without conflict.
除非另作定义,本申请所涉及的技术术语或者科学术语应当为本申请所属技术领域内具有一般技能的人士所理解的通常意义。本申请所涉及的“一”、“一个”、“一种”、“该”等类似词语并不表示数量限制,可表示单数或复数。本申请所涉及的术语“包括”、“包含”、“具有”以及它们任何变形,意图在于覆盖不排他的包含;例如包含了一系列步骤或模块(单元)的过程、方法、系统、产品或设备没有限定于已列出的步骤或单元,而是可以还包括没有列出的步骤或单元,或可以还包括对于这些过程、方法、产品或设备固有的其它步骤或单元。本申请所涉及的“连接”、“相连”、“耦接”等类似的词语并非限定于物理的或者机械的连接,而是可以包括电气的连接,不管是直接的还是间接的。本申请所涉及的“多个”是指两个或两个以上。“和/或”描述关联对象的关联关系,表示可以存在三种关系,例如,“A和/或B”可以表示:单独存在A,同时存在A和B,单独存在B这三种情况。字符“/”一般表示前后关联对象是一种“或”的关系。本申请所涉及的术语“第一”、“第二”、“第三”等仅仅是区别类似的对象,不代表针对对象的特定排序。Unless otherwise defined, the technical terms or scientific terms involved in the application shall have the usual meanings understood by those with ordinary skill in the technical field to which the application belongs. Words such as "a", "an", "an" and "the" involved in this application do not indicate a limitation on quantity, and may indicate singular or plural numbers. The terms "comprising", "comprising", "having" and any variations thereof involved in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product or process that includes a series of steps or modules (units). The apparatus is not limited to the listed steps or units, but may further include steps or units not listed, or may further include other steps or units inherent to the process, method, product or apparatus. The words "connected", "connected", "coupled" and similar words mentioned in this application are not limited to physical or mechanical connection, but may include electrical connection, no matter it is direct or indirect. The "plurality" involved in this application refers to two or more than two. "And/or" describes the association relationship of associated objects, indicating that there may be three types of relationships. For example, "A and/or B" may indicate: A exists alone, A and B exist simultaneously, and B exists independently. The character "/" generally indicates that the contextual objects are an "or" relationship. The terms "first", "second", "third" and the like involved in this application are only used to distinguish similar objects, and do not represent a specific ordering of objects.
本实施例提供了一种介质访问控制层,如图1所示,所述介质访问控制层连接在芯粒的片内通信接口和片间通信接口之间,所述介质访问控制层包括:This embodiment provides a media access control layer. As shown in FIG. 1, the media access control layer is connected between the on-chip communication interface and the inter-chip communication interface of the chip, and the media access control layer includes:
数据链路协议桥模块11,用于建立所述片内通信接口和所述片间通信接口之间的逻辑数据链路;A data link protocol bridge module 11, configured to establish a logical data link between the on-chip communication interface and the inter-chip communication interface;
与所述数据链路协议桥模块11连接的介质访问控制器模块12,用于接收所述数据链路协议桥模块11传输的片内通信接口数据,将所述片内通信接口数据编码为所述片间通信接口的规格所对应的帧数据,并将所述帧数据传输至所述片间通信接口;The media access controller module 12 connected with the data link protocol bridge module 11 is used to receive the on-chip communication interface data transmitted by the data link protocol bridge module 11, and encode the on-chip communication interface data into the frame data corresponding to the specifications of the inter-chip communication interface, and transmit the frame data to the inter-chip communication interface;
与所述数据链路协议桥模块11连接的配置模块13,用于接收并解析所述数据链路协议桥模块11传输的配置包,以配置所述片间通信接口。The configuration module 13 connected to the data link protocol bridge module 11 is configured to receive and parse the configuration packet transmitted by the data link protocol bridge module 11 to configure the inter-chip communication interface.
数据链路协议桥模块11,如图2所示,桥接了两种数据链路协议,一种是片内链路协议,用于与片内组件进行互联,另一种是一侧MAC到另一侧MAC的数据链路协议。在OSI模型中,Chiplet片间通信接口属于物理层,而介质访问控制器模块12属于数据链路协议桥模块11的介质访问控制子层,而数据链路协议桥模块11的逻辑链路控制子层则由数据链路协议桥1110实现,以建立一侧MAC到另一侧MAC的逻辑链路。The data link protocol bridge module 11, as shown in Figure 2, bridges two kinds of data link protocols, one is an on-chip link protocol for interconnecting with on-chip components, and the other is one side MAC to another The data link protocol of the MAC on one side. In the OSI model, the Chiplet inter-chip communication interface belongs to the physical layer, and the media access controller module 12 belongs to the media access control sublayer of the data link protocol bridge module 11, and the logical link control sublayer of the data link protocol bridge module 11 The layer is implemented by the data link protocol bridge 1110 to establish a logical link from the MAC on one side to the MAC on the other side.
片内链路协议使用NoC领域常用的基于信用的流量控制方法,基于信用的流量的控制是实现逐链路中每个虚拟回路流量控制的有效方式,在通过连接发送数据之前,发送端需要接收接收端通过虚拟回路发送的信用值,在不同时期,接收端发送信用额到发送端,说明接收端可用的缓冲区大小。当接收到信用额,发送端就按照信用额发送数据到接收端,每次发送端发送数据后,相应的信用额减少,这样可以有效减少失败重传造成网络阻塞。如表1所示,一组片内通信接口具有以下信号:The on-chip link protocol uses the credit-based flow control method commonly used in the NoC field. Credit-based flow control is an effective way to realize the flow control of each virtual circuit in each link. Before sending data through the connection, the sender needs to receive The credit value sent by the receiving end through the virtual circuit. In different periods, the receiving end sends the credit amount to the sending end, indicating the buffer size available at the receiving end. When the credit amount is received, the sending end sends data to the receiving end according to the credit amount. Each time the sending end sends data, the corresponding credit amount decreases, which can effectively reduce network congestion caused by failed retransmissions. As shown in Table 1, a set of on-chip communication interfaces has the following signals:
其中data_in,valid_in,credit_in这些信号用于MAC从片内通信接口接收的数据和相关流控;data_out,valid_out,credit_out这些信号用于MAC向片内通信接口发送的数据和相关流控。对于图2中标注的每个片内链路协议,可以存在多于一组上述接口,其中每组接口用于服务一个物理子网或一个抽象优先级。本实施例中,片内链路协议侧每组数据线(单个方向上)的位宽N定为128。Among them, the data_in, valid_in, and credit_in signals are used for the data and related flow control received by the MAC from the on-chip communication interface; the data_out, valid_out, and credit_out signals are used for the data and related flow control sent by the MAC to the on-chip communication interface. For each intra-chip link protocol marked in FIG. 2, there may be more than one set of the above-mentioned interfaces, wherein each set of interfaces is used to serve a physical subnet or an abstract priority. In this embodiment, the bit width N of each group of data lines (in a single direction) on the on-chip link protocol side is set to 128.
为了使片间通信接口可以被应用于多种场合,提高接口设计的灵活性,在接口内加入可配置选项,在接口正常工作前,按照设计需求对其进行配置。整个配置过程分为两步,第一步对片内传过来的配置包进行解析,第二步根据解析出来的配置地址和配置数据进行寄存器配置。In order to enable the inter-chip communication interface to be used in various occasions and improve the flexibility of the interface design, configurable options are added to the interface, and the interface is configured according to the design requirements before it works normally. The entire configuration process is divided into two steps. The first step is to analyze the configuration package transmitted from the chip, and the second step is to configure the registers according to the resolved configuration address and configuration data.
本实施例中的配置包具有专有格式,如表2所示。配置包有若干个连续的flit组成,每个flit的位宽为128位。第一个flit为数据包的包头,通过其中msgtype的数值可以确认此数据包为配置包。The configuration package in this embodiment has a proprietary format, as shown in Table 2. The configuration package consists of several consecutive flits, and the bit width of each flit is 128 bits. The first flit is the packet header of the data packet, through which the value of msgtype can confirm that this data packet is a configuration packet.
配置模块13通过读取数据链路协议桥模块11的数据判定当前为配置包时,将整个MAC电路转入配置状态,数据链路协议桥模块11当前数据及其后的若干条数据(由第一个flit中的packetSize的值标识数量)流入配置模块,而非像普通模式一样流过介质访问控制器模块12。Configuration module 13 by reading the data of data link protocol bridge module 11 and judging that it is currently a configuration packet, the whole MAC circuit is transferred to the configuration state, the current data of data link protocol bridge module 11 and several pieces of data thereafter (by the first The value of packetSize in a flit identifies the quantity) flows into the configuration module instead of flowing through the media access controller module 12 as in the normal mode.
对于配置包的前两个flit,配置模块取出其中的haddr,adX和dataX信号(X为0,1,2,3,次数由size所标识)。通过拼接haddr与adX得出完整地址,从而得出一次写入的所需要的信息:(寄存器地址_X,寄存器数据_X)。通过任意常用配置接口都可以写入这类数据,具体使用何种接口取决于Chiplet片间通信接口提供何种配置接口。对于后续的每个flit,可以最多装载3条配置信息,处理方法和上述方法相同。当配置过程完成后,MAC回到普通状态,进行数据传输,本实施例中的配置模块为一个状态机电路。For the first two flits of the configuration package, the configuration module extracts haddr, adX and dataX signals (X is 0, 1, 2, 3, and the number of times is identified by size). The complete address is obtained by splicing haddr and adX, so as to obtain the information required for one write: (register address _X, register data _X). This type of data can be written through any commonly used configuration interface. The specific interface used depends on the configuration interface provided by the chiplet inter-chip communication interface. For each subsequent flit, a maximum of 3 pieces of configuration information can be loaded, and the processing method is the same as the above method. After the configuration process is completed, the MAC returns to the normal state for data transmission. The configuration module in this embodiment is a state machine circuit.
本实施例通过数据链路协议桥模块,用于建立所述片内通信接口和所述片间通信接口之间的逻辑数据链路;与所述数据链路协议桥模块连接的介质访问控制器模块,用于接收所述数据链路协议桥模块传输的片内通信接口数据,将所述片内通信接口数据编码为所述片间通信接口的规格所对应的帧数据,并将所述帧数据传输至所述片间通信接口;与所述数据链路协议桥模块连接的配置模块,用于接收并解析所述数据链路协议桥模块传输的配置包,以配置所述片间通信接口。解决了介质访问控制层无法兼容市面上的各种不同规格的片间通信接口,导致介质访问控制层无法复用的问题。In this embodiment, a data link protocol bridge module is used to establish a logical data link between the on-chip communication interface and the inter-chip communication interface; a media access controller connected to the data link protocol bridge module A module, configured to receive the on-chip communication interface data transmitted by the data link protocol bridge module, encode the on-chip communication interface data into frame data corresponding to the specifications of the inter-chip communication interface, and encode the frame data into The data is transmitted to the inter-chip communication interface; the configuration module connected to the data link protocol bridge module is used to receive and analyze the configuration packet transmitted by the data link protocol bridge module to configure the inter-chip communication interface . It solves the problem that the media access control layer cannot be compatible with various inter-chip communication interfaces of different specifications on the market, resulting in the problem that the media access control layer cannot be reused.
在其中一个实施例中,如图3所示,所述介质访问控制器模块12包括第一帧编码器1210,用于根据所述片间通信接口对应的位宽和时钟频率,将所述片内通信接口数据编码为所述片间通信接口的规格所对应的帧数据。In one embodiment, as shown in FIG. 3 , the media access controller module 12 includes a first frame encoder 1210, configured to convert the slices to The internal communication interface data is encoded as frame data corresponding to the specification of the inter-chip communication interface.
不同规格下片间通信接口所提供的片间传输带宽总是一样的,但是受到工艺限制等多方面的原因,向片内提供的接口具有不同的时钟频率和位宽,本实施例中,片内链路协议侧每组数据线(单个方向上)的位宽为128。MAC所支持的子网数量为2,片间通信接口对应有以下三种规格:The inter-chip transmission bandwidth provided by the inter-chip communication interface under different specifications is always the same, but due to various reasons such as process limitations, the interfaces provided to the chip have different clock frequencies and bit widths. In this embodiment, the chip The bit width of each group of data lines (in a single direction) on the inner link protocol side is 128. The number of subnets supported by the MAC is 2, and the inter-chip communication interface corresponds to the following three specifications:
(1)4X规格(1) 4X specification
此规格下的介质访问控制器模块12针对大位宽、低时钟频率设计。片间通接口提供的位宽为640位,MAC将此位宽看作是四组160位的数据位宽,每组160位可以容纳来自片内接口的128位宽的数据和一些附加数据。The MAC module 12 under this specification is designed for large bit width and low clock frequency. The bit width provided by the inter-chip communication interface is 640 bits. MAC regards this bit width as four groups of 160-bit data bits. Each group of 160 bits can accommodate 128-bit wide data and some additional data from the on-chip interface.
(2)2X规格(2) 2X specification
此规格下片间通接口提供的位宽为320位,MAC将此位宽看作是两组160位的数据位宽,每组160位可以容纳来自片内接口的128位宽的数据和一些附加数据。时钟频率为4X规格的两倍。The bit width provided by the inter-chip communication interface under this specification is 320 bits. MAC regards this bit width as two sets of 160-bit data bit width. Each set of 160 bits can accommodate 128-bit wide data from the on-chip interface and some Additional data. The clock frequency is twice that of the 4X specification.
(3)1X规格(3) 1X specification
此规格下的介质访问控制器针对小位宽、高时钟频率设计。片间通接口提供为位宽为160位,MAC将此位宽看作是一组160位的数据位宽,此位宽仅可容纳一个128位的数据及几位附加信息。时钟频率为4X规格的四倍。Media access controllers under this specification are designed for small bit width and high clock frequency. The inter-chip communication interface provides a bit width of 160 bits. MAC regards this bit width as a set of 160-bit data bit width, which can only accommodate a 128-bit data and several additional information. The clock frequency is four times that of the 4X specification.
不同规格下提供的数据位宽不同,1X规格下位宽为160位,帧格式如图4所示,sel位表示数据来源于哪个子网路由,val位表示数据段数据是否有效,dat位表示传输数据段,QID标识了被仲裁出的数据来自于哪一个数据链路协议桥1110,yms标识流控信息。对于2X规格和4X规格编码得出的帧,分别具有320位和640位,其中319~160位,639~480位,479~320位的位段定义与低160位一致,但是不必填写任何流控信息(段内偏移23~16位置为0即可)。The data bit width provided under different specifications is different. The bit width of the 1X specification is 160 bits. The frame format is shown in Figure 4. The sel bit indicates which subnet route the data comes from, the val bit indicates whether the data segment data is valid, and the dat bit indicates the transmission In the data segment, QID identifies which data link protocol bridge 1110 the arbitrated data comes from, and yms identifies flow control information. For the frames encoded by the 2X specification and the 4X specification, they have 320 bits and 640 bits respectively, of which 319~160 bits, 639~480 bits, and 479~320 bits are defined to be consistent with the lower 160 bits, but there is no need to fill in any stream Control information (the position of offset 23~16 in the segment is 0).
在其中一个实施例中,如图5所示,所述介质访问控制器模块12还包括第一仲裁器1220,用于根据所述片内通信接口数据对应的子网优先级以及所述片内通信接口数据对应的数据链路协议桥模块优先级对所述片内通信接口数据进行仲裁。第一仲裁器1220负责对来自数据链路协议桥1110的待仲裁数据进行仲裁。对于每组待仲裁数据接口,如表3所示,有以下信号:In one of the embodiments, as shown in FIG. 5 , the media access controller module 12 further includes a first arbiter 1220, configured to use the subnet priority corresponding to the on-chip communication interface data and the on-chip The priority of the data link protocol bridge module corresponding to the communication interface data arbitrates the on-chip communication interface data. The first arbiter 1220 is responsible for arbitrating the data to be arbitrated from the data link protocol bridge 1110 . For each group of data interfaces to be arbitrated, as shown in Table 3, there are the following signals:
介质访问控制器模块12中第一仲裁器1220具有如下仲裁规则:对于来自数据链路协议桥模块11的待仲裁数据,优先根据所述片内通信接口数据对应的子网优先级进行仲裁,仲裁规则为应答子网的优先级总高于请求子网,对于具有同样优先级的待仲裁数据,采用片内通信接口数据对应的数据链路协议桥模块11优先级,具体为数据链路协议桥1110的优先级进行仲裁。例如本实施例中数据链路协议桥模块11具有四个数据链路协议桥1110,对应的ID为0、1、2、3,可采用固定优先级从高到低依次为0、1、2、3。The first arbitrator 1220 in the media access controller module 12 has the following arbitration rules: for the data to be arbitrated from the data link protocol bridge module 11, the arbitration is preferentially performed according to the subnet priority corresponding to the on-chip communication interface data, and the arbitration The rule is that the priority of the response subnet is always higher than that of the request subnet. For data to be arbitrated with the same priority, the priority of the data link protocol bridge module 11 corresponding to the on-chip communication interface data is used, specifically the data link protocol bridge 1110 priority for arbitration. For example, in this embodiment, the data link protocol bridge module 11 has four data link protocol bridges 1110, the corresponding IDs are 0, 1, 2, and 3, and the fixed priorities can be 0, 1, and 2 from high to low. , 3.
根据片间通信接口对应的规格,实施例中以四个数据链路协议桥1110为例,第一仲裁器1220进行如下仲裁:对于4X规格,无需仲裁,四个数据链路协议桥1110的数据都作为输出数据;对于2X规格,第一仲裁器1220从四个数据链路协议桥1110的数据中仲裁出两个作为输出数据;对于1X规格,第一仲裁器1220从四个数据链路协议桥1110的数据中仲裁出一个作为输出数据。According to the specifications corresponding to the inter-chip communication interface, four data link protocol bridges 1110 are taken as an example in the embodiment, and the first arbiter 1220 performs the following arbitration: For the 4X specification, no arbitration is required, and the data of the four data link protocol bridges 1110 are all used as output data; for 2X specification, the first arbiter 1220 arbitrates two data from the four data link protocol bridges 1110 as output data; One of the data of the bridge 1110 is arbitrated as the output data.
对于众核架构的芯片,常采用NoC作为片内互联系统。有些情况下为了避免发生协议级的死锁,常常根据NoC上承载的上层协议的事务的不同阶段,配备不同的NoC物理子网。这些物理NoC子网在拓扑上和路由各方面一般都是一样的,但是承载不同的数据包。例如划分为请求子网和应答子网,它们分别只传输请求数据包和应答数据包,通过这种方式将不同优先级在物理上分开,避免了死锁。For chips with many-core architecture, NoC is often used as the on-chip interconnection system. In some cases, in order to avoid protocol-level deadlocks, different NoC physical subnets are often configured according to the different stages of the upper-layer protocol transactions carried by the NoC. These physical NoC subnets are generally identical in topology and routing aspects, but carry different packets. For example, it is divided into a request subnet and a response subnet, which only transmit request packets and response packets respectively. In this way, different priorities are physically separated and deadlocks are avoided.
本实施例中的介质访问控制层支持芯片内部具有多个不同优先级的物理子网使用片间通信接口进行通信,并且保证多个物理子网同时共享同一个Chiplet片间接口的前提下,依然维持物理子网之间的优先级关系,从而把死锁避免的性质通过介质访问控制层延拓到整个芯粒互联系统。The media access control layer in this embodiment supports multiple physical subnets with different priorities inside the chip to communicate using the inter-chip communication interface, and ensures that multiple physical subnets share the same Chiplet inter-chip interface at the same time. The priority relationship between physical subnets is maintained, so that the nature of deadlock avoidance is extended to the entire chip interconnection system through the media access control layer.
在其中一个实施例中,如图6所示,所述数据链路协议桥模块11包括:In one of the embodiments, as shown in FIG. 6, the data link protocol bridge module 11 includes:
第一数据处理模块21,用于接收并处理所述片内通信接口数据,并将处理后的片内通信接口数据传输至所述介质访问控制器模块12;The first data processing module 21 is configured to receive and process the on-chip communication interface data, and transmit the processed on-chip communication interface data to the media access controller module 12;
第二仲裁器22,用于根据所述片内通信接口数据对应的子网是否持有下游信用以及所述片内通信接口数据对应的子网优先级,对所述片内通信接口数据进行仲裁。The second arbiter 22 is configured to arbitrate the on-chip communication interface data according to whether the subnet corresponding to the on-chip communication interface data holds downstream credits and the subnet priority corresponding to the on-chip communication interface data .
第一数据处理模块21从片内接收数据后,数据向第二仲裁器22提出请求,仲裁器根据数据对应的子网是否持有下游信用决定数据是否具有发送资格,下游信用指的是从另一侧MAC发出的,被本侧MAC接收的信用,从所有具有发送请求和发送资格的数据中按照子网优先级得出仲裁结果,第一数据处理模块21根据仲裁结果对数据进行处理并传输至介质访问控制器模块12。After the first data processing module 21 receives data from the chip, the data makes a request to the second arbiter 22, and the arbitrator decides whether the data is qualified to send according to whether the subnet corresponding to the data holds downstream credit. The credit sent by the MAC on one side and received by the MAC on this side obtains the arbitration result from all the data with sending request and sending qualification according to the subnet priority, and the first data processing module 21 processes and transmits the data according to the arbitration result to the media access controller module 12.
在其中一个实施例中,如图7所示,所述第一数据处理模块21包括:In one of the embodiments, as shown in FIG. 7, the first data processing module 21 includes:
第一存储器31,用于存放所述片内通信接口数据;The first memory 31 is used to store the on-chip communication interface data;
复用器32,用于根据所述第二仲裁器22获得的仲裁结果,输出对应的片内通信接口数据;A multiplexer 32, configured to output corresponding on-chip communication interface data according to the arbitration result obtained by the second arbiter 22;
第二帧编码器33,用于将所述片内通信接口数据编码为所述片间通信接口的规格所对应的帧数据。The second frame encoder 33 is configured to encode the intra-chip communication interface data into frame data corresponding to the specification of the inter-chip communication interface.
本实施例中的第一存储器31为FIFO(First In First Out),FIFO是一种先进先出的数据缓存器,在与片内接口相邻处,使用FIFO暂存接口数据,进行流量控制,并划分片内时钟域与MAC时钟域。从片内接收的数据,被存储器存放后,由第二仲裁器22进行仲裁,决定存放于哪一个FIFO中的数据可以通过。每个非空FIFO都可以向仲裁器提出请求,第二仲裁器22根据FIFO对应的子网是否持有下游信用决定FIFO是否具有发送资格。复用器32用于根据所述第二仲裁器22获得的仲裁结果,输出对应的数据。The first memory 31 in this embodiment is FIFO (First In First Out), and FIFO is a first-in-first-out data buffer. Adjacent to the on-chip interface, FIFO is used to temporarily store interface data for flow control. And divide the on-chip clock domain and the MAC clock domain. After the data received from the chip is stored in the memory, the second arbiter 22 performs arbitration to determine which FIFO the data stored in can pass through. Each non-empty FIFO can submit a request to the arbiter, and the second arbiter 22 determines whether the FIFO has the sending qualification according to whether the subnet corresponding to the FIFO holds downstream credit. The multiplexer 32 is configured to output corresponding data according to the arbitration result obtained by the second arbiter 22 .
第二帧编码器33将输出的数据编码为所述片间通信接口的规格所对应的数据链路帧,用于承载来自片内的128位数据和附加信息为例,帧格式如图8所示。The second frame encoder 33 encodes the output data into a data link frame corresponding to the specification of the inter-chip communication interface, which is used to carry 128-bit data and additional information from the chip as an example. The frame format is as shown in FIG. 8 Show.
其中,subnet位标识数据来自于哪个物理NoC子网,vaild位表示数据段是否有效,dat位为传输数据段,ym1和ym0分别为需要传送给对端MAC的流控信用(前文所称credit_in或credit_out信号线)信号。在数据链路协议桥1110与介质访问控制器模块12的传输中,无论输入输出,帧格式都是这样。Among them, the subnet bit identifies which physical NoC subnet the data comes from, the vaild bit indicates whether the data segment is valid, the dat bit indicates the transmission data segment, and ym1 and ym0 are the flow control credits that need to be transmitted to the peer MAC (referred to as credit_in or credit_out signal line) signal. In the transmission between the data link protocol bridge 1110 and the media access controller module 12, regardless of input and output, the frame format is the same.
所述数据链路协议桥模块11还包括第二数据处理模块23,用于接收并处理所述介质访问控制器模块传输的片间通信接口数据,并将处理后的所述片间通信接口数据传输至所述片内通信接口。图9是片内存在两个物理NoC子网情况下的数据链路协议桥1110的结构。实线表示数据流向,虚线表示控制流向。第二数据处理模块23包括帧解码器34、解复用器35、第二存储器36,帧解码器34用于将介质访问控制器模块12传输过来的片间通信接口数据进行解码、解复用器35用于根据数据的来源将其存放至第二存储器36中,本实施例中的第二存储器也为FIFO。The data link protocol bridge module 11 also includes a second data processing module 23, configured to receive and process the inter-chip communication interface data transmitted by the media access controller module, and process the inter-chip communication interface data transmitted to the on-chip communication interface. FIG. 9 shows the structure of the data link protocol bridge 1110 in the case that there are two physical NoC subnets in the chip. Solid lines indicate data flow, and dashed lines indicate control flow. The second data processing module 23 includes a frame decoder 34, a demultiplexer 35, and a second memory 36. The frame decoder 34 is used to decode and demultiplex the inter-chip communication interface data transmitted by the media access controller module 12. The device 35 is used to store the data in the second memory 36 according to the source of the data, and the second memory in this embodiment is also a FIFO.
本申请实施例还提供了一种基于介质访问控制层的通信方法,如图10所示,该方法包括以下步骤:The embodiment of the present application also provides a communication method based on the media access control layer, as shown in FIG. 10 , the method includes the following steps:
步骤S401,接收所述数据链路协议桥模块传输的片内通信接口数据;Step S401, receiving the on-chip communication interface data transmitted by the data link protocol bridge module;
步骤S402,将所述片内通信接口数据编码为所述片间通信接口的规格所对应的帧数据;Step S402, encoding the on-chip communication interface data into frame data corresponding to the specification of the inter-chip communication interface;
步骤S403,将所述帧数据传输至所述片间通信接口。Step S403, transmitting the frame data to the inter-chip communication interface.
本实施例,通过接收所述数据链路协议桥模块传输的片内通信接口数据;将所述片内通信接口数据编码为所述片间通信接口的规格所对应的帧数据;将所述帧数据传输至所述片间通信接口。解决了介质访问控制层无法兼容市面上的各种不同规格的片间通信接口,导致介质访问控制层无法复用的问题。In this embodiment, by receiving the on-chip communication interface data transmitted by the data link protocol bridge module; encoding the on-chip communication interface data into frame data corresponding to the specification of the inter-chip communication interface; encoding the frame Data is transmitted to the inter-chip communication interface. It solves the problem that the media access control layer cannot be compatible with various inter-chip communication interfaces of different specifications on the market, resulting in the problem that the media access control layer cannot be reused.
在其中一个实施例中,所述方法还包括以下步骤:In one embodiment, the method also includes the following steps:
步骤1,接收所述介质访问控制模块传输的片间通信接口数据;Step 1, receiving the inter-chip communication interface data transmitted by the media access control module;
步骤2,将所述片间通信接口数据解码并根据所述片间通信接口数据对应的来源传输至所述片内通信接口。Step 2, decoding the inter-chip communication interface data and transmitting it to the on-chip communication interface according to the source corresponding to the inter-chip communication interface data.
本申请实施例还提供了芯粒片间通信系统,如图11所示,所述系统包括第一介质访问控制层51以及第二介质访问控制层52,所述第一介质访问控制层51与所述第二介质访问控制层52通信连接;所述第一介质访问控制层51和/或所述第二介质访问控制层52实现如上述基于介质访问控制层的通信方法。The embodiment of the present application also provides an inter-chip communication system, as shown in FIG. 11 , the system includes a first media access control layer 51 and a second media access control layer 52, and the first media access control layer 51 and The second media access control layer 52 is connected in communication; the first media access control layer 51 and/or the second media access control layer 52 implement the communication method based on the above media access control layer.
当整个系统上电复位时,第一介质访问控制层51以及第二介质访问控制层52需要从片内接收配置数据包,据此配置片间通信接口,并等待片间通信接口建立与对面片间通信接口的连接,随后可以进入正常使用。当配置阶段完成后进入普通阶段,可以进行跨芯粒的通信。下面以具体的例子说明数据在整个数据通路上的传输过程。When the entire system is powered on and reset, the first media access control layer 51 and the second media access control layer 52 need to receive configuration data packets from the chip, configure the inter-chip communication interface accordingly, and wait for the inter-chip communication interface to be established and communicate with the opposite chip. The connection between the communication interfaces can then be put into normal use. After the configuration phase is completed, it enters the normal phase, where cross-chip communication can be performed. The following uses specific examples to illustrate the data transmission process on the entire data path.
首先,来自片内的数据被传输到了第一介质访问控制层51片内接口协议处,通过片间通信接口,数据被接收写入数据链路桥的子网对应FIFO中。一旦FIFO中存有数据,便可以向第二仲裁器22提出请求,第二仲裁器22从所有具有发送请求和发送资格的FIFO中按照子网优先级得出仲裁结果。随后根据此步仲裁中获胜的请求方的数据构造一个数据链路帧,并向介质访问控制器模块12提出仲裁请求,与来自其他数据链路桥的请求进行竞争。介质访问控制器模块12根据规格不同,选中不同的数据,构造新的数据帧,发送给片间通信接口。片间通信接口将数据传输到对面的第二介质访问控制层52对应的片间通信接口并将数据递交第二介质访问控制层52处理。First, the data from the chip is transmitted to the first medium access control layer 51 on-chip interface protocol, and through the inter-chip communication interface, the data is received and written into the FIFO corresponding to the subnet of the data link bridge. Once data is stored in the FIFO, a request can be made to the second arbiter 22, and the second arbiter 22 obtains an arbitration result from all FIFOs that have the sending request and sending qualification according to the priority of the subnetwork. Then construct a data link frame according to the data of the winning requester in this step of arbitration, and submit an arbitration request to the media access controller module 12 to compete with requests from other data link bridges. The media access controller module 12 selects different data according to different specifications, constructs a new data frame, and sends it to the inter-chip communication interface. The inter-chip communication interface transmits the data to the corresponding inter-chip communication interface of the opposite second media access control layer 52 and submits the data to the second media access control layer 52 for processing.
第二介质访问控制层52将接收到的数据位宽每160位看作一个部分,分别进行处理。并取出最低160位数据中的所有流量控制信号。第二介质访问控制层52根据数据位段中的QID号码,把数据分派给对应的数据链路桥。也把流控信号分派给数据链路桥。The second medium access control layer 52 regards each 160-bit bit width of the received data as a part and processes them respectively. And take out all flow control signals in the lowest 160 bits of data. The second media access control layer 52 assigns the data to the corresponding data link bridge according to the QID number in the data segment. Flow control signals are also dispatched to the data link bridge.
每个数据链路桥接收到数据帧后,根据数据帧中的子网号码(subnet号或sel号)分派到对应的子网FIFO。随后通过片内接口协议将数据递交片内组件。以上为完成一次端到端传输的过程。After receiving the data frame, each data link bridge assigns it to the corresponding subnet FIFO according to the subnet number (subnet number or sel number) in the data frame. Then the data is delivered to the on-chip components through the on-chip interface protocol. The above is the process of completing an end-to-end transmission.
本实施例的芯粒片间通信系统,通过第一介质访问控制层51以及第二介质访问控制层52,可实现专门用于芯粒片间通信场景芯粒片间通信系统,解决了介质访问控制层无法兼容市面上的各种不同规格的片间通信接口,导致介质访问控制层无法复用的问题。The chip-to-chip communication system in this embodiment, through the first media access control layer 51 and the second media access control layer 52, can realize the chip-to-chip communication system specially used in the chip-to-chip communication scenario, and solve the problem of medium access The control layer cannot be compatible with various inter-chip communication interfaces of different specifications on the market, resulting in the problem that the media access control layer cannot be reused.
以上所述实施例的各技术特征可以进行任意的组合,为使描述简洁,未对上述实施例中的各个技术特征所有可能的组合都进行描述,然而,只要这些技术特征的组合不存在矛盾,都应当认为是本说明书记载的范围。The technical features of the above-mentioned embodiments can be combined arbitrarily. To make the description concise, all possible combinations of the technical features in the above-mentioned embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, should be considered as within the scope of this specification.
以上所述实施例仅表达了本申请的几种实施方式,其描述较为具体和详细,但并不能因此而理解为对发明专利范围的限制。应当指出的是,对于本领域的普通技术人员来说,在不脱离本申请构思的前提下,还可以做出若干变形和改进,这些都属于本申请的保护范围。因此,本申请专利的保护范围应以所附权利要求为准。The above-mentioned embodiments only represent several implementation modes of the present application, and the description thereof is relatively specific and detailed, but it should not be construed as limiting the scope of the patent for the invention. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the scope of protection of the patent application should be based on the appended claims.
Claims (10)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310893529.XA CN116627894B (en) | 2023-07-20 | 2023-07-20 | Media access control layer, communication method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310893529.XA CN116627894B (en) | 2023-07-20 | 2023-07-20 | Media access control layer, communication method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN116627894A true CN116627894A (en) | 2023-08-22 |
CN116627894B CN116627894B (en) | 2023-10-20 |
Family
ID=87642170
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310893529.XA Active CN116627894B (en) | 2023-07-20 | 2023-07-20 | Media access control layer, communication method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116627894B (en) |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1432139A2 (en) * | 2002-12-19 | 2004-06-23 | Laboratoire Europeen ADSL Leacom Fastnet | Digital data distribution system |
US20060103535A1 (en) * | 2004-11-15 | 2006-05-18 | Kourosh Pahlaven | Radio frequency tag and reader with asymmetric communication bandwidth |
CN104699654A (en) * | 2015-03-02 | 2015-06-10 | 福州瑞芯微电子有限公司 | Interconnection adapting system and method based on CHI on-chip interaction bus and QPI inter-chip interaction bus |
CN106571903A (en) * | 2015-10-08 | 2017-04-19 | 深圳市中兴微电子技术有限公司 | Communication method and system between chips |
CN111092773A (en) * | 2019-12-25 | 2020-05-01 | 成都九芯微科技有限公司 | PCIE (peripheral component interface express) switching chip port configuration system and method supporting virtual switching |
CN114844740A (en) * | 2022-04-29 | 2022-08-02 | 东风电驱动系统有限公司 | Communication method between double chips and chip |
CN115794731A (en) * | 2023-01-29 | 2023-03-14 | 北京超摩科技有限公司 | Decoupling control method for transmission of multi-channel data link between core particles |
CN115914134A (en) * | 2022-11-14 | 2023-04-04 | 中国第一汽车股份有限公司 | Vehicle-mounted Ethernet data transmission method and device based on TSN frame preemption mechanism |
-
2023
- 2023-07-20 CN CN202310893529.XA patent/CN116627894B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1432139A2 (en) * | 2002-12-19 | 2004-06-23 | Laboratoire Europeen ADSL Leacom Fastnet | Digital data distribution system |
US20060103535A1 (en) * | 2004-11-15 | 2006-05-18 | Kourosh Pahlaven | Radio frequency tag and reader with asymmetric communication bandwidth |
CN104699654A (en) * | 2015-03-02 | 2015-06-10 | 福州瑞芯微电子有限公司 | Interconnection adapting system and method based on CHI on-chip interaction bus and QPI inter-chip interaction bus |
CN106571903A (en) * | 2015-10-08 | 2017-04-19 | 深圳市中兴微电子技术有限公司 | Communication method and system between chips |
CN111092773A (en) * | 2019-12-25 | 2020-05-01 | 成都九芯微科技有限公司 | PCIE (peripheral component interface express) switching chip port configuration system and method supporting virtual switching |
CN114844740A (en) * | 2022-04-29 | 2022-08-02 | 东风电驱动系统有限公司 | Communication method between double chips and chip |
CN115914134A (en) * | 2022-11-14 | 2023-04-04 | 中国第一汽车股份有限公司 | Vehicle-mounted Ethernet data transmission method and device based on TSN frame preemption mechanism |
CN115794731A (en) * | 2023-01-29 | 2023-03-14 | 北京超摩科技有限公司 | Decoupling control method for transmission of multi-channel data link between core particles |
Non-Patent Citations (4)
Title |
---|
HANG XIAO 等: "Fast and High-Accuracy Approximate MAC Unit Design for CNN Computing", 《IEEE EMBEDDED SYSTEMS LETTERS》, vol. 14, no. 3, pages 155 - 158 * |
TRIO ADIONO 等: "Flexible data sharing architecture of WiMAX heterogeneous Multiprocessor System on Chip", 《2015 INTERNATIONAL SOC DESIGN CONFERENCE (ISOCC)》, pages 131 - 132 * |
岑姣: "Chiplet片间并行接口", 《中国优秀硕士学位论文全文数据库 信息科技辑》, no. 7, pages 135 - 48 * |
袁文燕: "FPGA 片间万兆可靠通信的设计与实现", 《电视技术》, vol. 38, no. 1, pages 43 - 46 * |
Also Published As
Publication number | Publication date |
---|---|
CN116627894B (en) | 2023-10-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10084692B2 (en) | Streaming bridge design with host interfaces and network on chip (NoC) layers | |
US8718065B2 (en) | Transmission using multiple physical interface | |
US6157972A (en) | Apparatus and method for processing packetized information over a serial bus | |
US9736071B2 (en) | General input/output architecture, protocol and related methods to implement flow control | |
US9836424B2 (en) | General input/output architecture, protocol and related methods to implement flow control | |
JP3412825B2 (en) | Method and apparatus for switching data packets over a data network | |
KR100611268B1 (en) | Enhanced General I / O Architecture and Related Methods for Virtual Channel Setup | |
US20050132089A1 (en) | Directly connected low latency network and interface | |
US7596148B2 (en) | Receiving data from virtual channels | |
WO2003019392A1 (en) | A general input/output architecture, protocol and related methods to provide isochronous channels | |
CN116250220A (en) | A multi-chip packaging structure and switch | |
US20060056424A1 (en) | Packet transmission using output buffer | |
CN116627894B (en) | Media access control layer, communication method and system | |
US20060184710A1 (en) | Bridge between a single channel high speed bus and a multiple channel low speed bus | |
US11860811B2 (en) | Message protocol for a data processing system | |
US20030065869A1 (en) | PCI/LVDS half bridge | |
US6857033B1 (en) | I/O node for a computer system including an integrated graphics engine and an integrated I/O hub | |
WO2024102916A1 (en) | Root complex switching across inter-die data interface to multiple endpoints | |
EP1625504A1 (en) | Processing system and method for communicating data |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |