CN112732612A - Double-side read address event representation transmission protocol circuit - Google Patents
Double-side read address event representation transmission protocol circuit Download PDFInfo
- Publication number
- CN112732612A CN112732612A CN201911033054.7A CN201911033054A CN112732612A CN 112732612 A CN112732612 A CN 112732612A CN 201911033054 A CN201911033054 A CN 201911033054A CN 112732612 A CN112732612 A CN 112732612A
- Authority
- CN
- China
- Prior art keywords
- column
- row
- response
- arbiter
- request
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F13/00—Interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
- G06F13/38—Information transfer, e.g. on bus
- G06F13/42—Bus transfer protocol, e.g. handshake; Synchronisation
- G06F13/4282—Bus transfer protocol, e.g. handshake; Synchronisation on a serial bus, e.g. I2C bus, SPI bus
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F13/00—Interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
- G06F13/14—Handling requests for interconnection or transfer
- G06F13/16—Handling requests for interconnection or transfer for access to memory bus
- G06F13/1605—Handling requests for interconnection or transfer for access to memory bus based on arbitration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2213/00—Indexing scheme relating to interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
- G06F2213/0056—Use of address and non-data lines as data lines for specific data transfers to temporarily enlarge the data bus and increase information transfer rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2213/00—Indexing scheme relating to interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
- G06F2213/36—Arbitration
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Bus Control (AREA)
Abstract
A bilateral read address event representation transmission protocol circuit comprises a row arbiter, a row address encoder, a column arbiter 1, a column arbiter 2, a column address encoder 1, a column address encoder 2, an input 2jIndividual line request and 2kColumn request, output includes 2jIndividual line answering, 2kThe system comprises an individual column response and 2 event information external output channels, an event output channel 1 and an event output channel 2; the circuit groups the column requests of the circuit array, increases the path of the external output of the protocol circuit, and doubles the circuit speed; the large-scale column request arbiter is divided into two smaller-scale column arbiters, so that the delay of a circuit in column arbitration is reduced, the time for queuing and waiting required by the output units in the same row is reduced, and the accuracy of information contained in output data is improved.
Description
Technical Field
The present invention relates to the field of integrated circuits, and more particularly to a circuit implementation of an address-event representation transfer protocol for a dual-sided read mode.
Background
Address-event Representation (AER) transport protocol is oneThe novel asynchronous high-speed serial data transmission protocol is suitable for high-speed serial output of a large-scale evened circuit array, and the circuit structure of the novel asynchronous high-speed serial data transmission protocol is shown in figure 1 and comprises a row arbiter, a row address encoder, a column arbiter and a column address encoder. The circuit array has 2jStripe line request bus and 2jA bar line reply bus, and 2kRowbar request bus and 2kThe columnar response bus is used for interacting with the AER protocol circuit. The operating principle of the AER protocol circuit is as follows: the circuit array generates requests on at least one row request bus; after the AER protocol circuit receives the row requests, the row arbiter sorts the row requests according to the principle of first-in first-out, and sequentially responds signals on the corresponding row response buses according to the sequence, and meanwhile, the row address encoder encodes the row number of the response signal; the circuit array generates a request on at least one column request bus after receiving the row response signal; after the AER protocol circuit receives the column requests, the column arbiter also sequences the row requests according to the principle of first-in first-out, and sequentially responds signals on the corresponding column response buses according to the sequence, and simultaneously, the column address encoder encodes the column number where the response signal is located; after one row response and one column response are completed, the protocol circuit outputs two groups of codes to the outside, the signal transmission process of one unit in the circuit array is completed, and the signal of the next unit is processed. According to the working principle, the AER protocol circuit serially completes the signal transmission process of each unit needing to be output in the circuit array.
The core of the AER protocol circuit is an arbiter, and the arbiter units of 2 to 1 are usually connected into a tree structure to complete the response sorting function. FIG. 2 shows a general formula of (2)n-1) an n-level binary tree structure composed of arbitration units, the bottom level (i.e. the n-th level) having from-req 0 to-req (2)n-1) 2 in allnA request signal input and from-ack 0 to-ack (2)n-1) 2 in allnAnd outputting the response signal. Each arbitration unit can complete the arbitration work of two simultaneously generated requests of the layer, generates a request signal of the previous layer, and returns a response signal to the next layer according to the arbitration sequence after receiving the response signal returned by the previous layerIn the cell. The top layer (i.e. layer 1) unit of the arbiter transmits the generated request signal to the external circuit, and receives the response signal returned by the external circuit to transmit to the lower layer. If the requests received by the arbitration unit are not generated simultaneously, the arbitration result is generated according to the sequence of the requests.
The existing AER protocol circuit has the defects that each external output unit of the circuit array can only be sequenced through one channel and output in series in the output process, and when a plurality of units in the circuit array need to output externally at the same time, the unselected units can only wait for the output of the unit sequenced before the unit to output externally. Since the accuracy of the event information is determined by the accuracy of the time stamp at the time of the event generation, this latency will cause errors to the output information. Furthermore, the time required for the arbitration process will increase as the number of requests increases, which means that circuit array scaling will result in worse event information accuracy.
When the number of requests is 1, the response delay of the arbiter to the request is
Whereint reqIndicating the delay time for a request signal to travel one level from the lower level of the arbitration tree to the upper level,t ackindicating the delay time for an acknowledgment signal to travel one level from the upper level to the lower level of the arbitration tree,t ack-MEthe delay time of the acknowledgement signal of the arbitration unit at the topmost layer of the arbitration tree is transmitted to the lower layer by one level. And for a full request, the response delay of the arbiter to the last request is
Whereint swThe delay time of one switching of the acknowledgement signal by each alternative arbiter is referred to,T prefers to the time each time the arbiter waits for external processing.
It can be seen that as the size of the arbiter continues to grow, the arbitration delay will grow linearly, further degrading event information accuracy.
Disclosure of Invention
The utility model provides a bilateral address event that reads shows transmission protocol circuit, its advantage lies in changing the AER protocol circuit of current single channel mode serial output into the output of binary channels, reduces the time stamp error of event under the big array, improves the degree of accuracy of event information.
A bilateral read address event representation transmission protocol circuit is shown in figure 3, and the structure comprises a row arbiter, a row address encoder, a column arbiter 1, a column arbiter 2, a column address encoder 1, a column address encoder 2, and an input 2jIndividual line request (line request)<0>Request to bank<2j-1>) And 2kIndividual column request (column request)<0>Request to rank<2k-1>) The output comprises 2jIndividual line response (line response)<0>To bank response<2j-1>)、2kIndividual column response (column response)<0>Pair response<2k-1>) And 2 event information external output channels, an event output channel 1 and an event output channel 2; the connection mode of each module is as follows: line requests<0>Request to bank<2j-1>All access the row arbiter, the row arbiter outputs the acknowledge signal to the bus row acknowledge<0>To bank response<2j-1>Each row response bus is connected to a row address encoder and a return circuit array; column request<0>Request to rank<2k-1-1>A column arbiter 1 positioned at the upper left of the array is accessed, and the column arbiter 1 outputs a response signal to a bus column response<0>Pair response<2k-1-1>Each row response bus is connected to the row address encoder 1 and the return circuit array; column request<2k-1>Request to rank<2k-1>A column arbiter 2 positioned at the lower right of the array is accessed, and the column arbiter 2 outputs a response signal to the bus column response<2k-1>Pair response<2k-1>And each row reply bus is connected to the row address encoder 2 and the return circuit array.
A two-sided read address event representation transfer protocol circuit operating as follows: each unit needing to be output in the circuit array generates a row request on a row bus of a row where the unit is located and inputs the row request into a row arbiter; the row arbiter sequences all the generated row requests and generates response signals on a row response bus corresponding to the row request positioned at the head; the row address encoder encodes the row number of the row where the response signal is located; after the row response signal returns to the circuit array, the unit circuit needing to be output in the row generates a request signal on a corresponding row request bus, wherein the row number is 0 to (2)k-1-1) column requests in the range are input to the column arbiter 1 with a column number of 2k-1~(2k-1) column requests in the range are input to the column arbiter 2. The column arbiter 1 and the column arbiter 2 respectively sequence the input column requests at the same time, and generate response signals on the column response bus corresponding to the first column request, that is, generate two column response signals at the same time; the column address encoder 1 encodes the column number of the column where the response signal generated by the column arbiter 1 is located, and the column address encoder 2 encodes the column number of the column where the response signal generated by the column arbiter 2 is located; at this time, the row address coded by the row address coder and the column address coded by the column address coder 1 form event information 1 to be output from an event output channel 1, and the row address coded by the row address coder and the column address coded by a column address coder 2 form event information 2 to be output from an event output channel 2; after the column response returns to the circuit array, the event information in the corresponding unit is eliminated, and the protocol circuit starts to process the next pair of units needing to be output.
An address event representation transfer protocol circuit for bilateral read-out, which is advantageous in that: 1. by grouping the column requests of the circuit array, the path of the external output of the protocol circuit is increased, and the circuit speed is doubled; 2. the large-scale column request arbiter is divided into two smaller-scale column arbiters, so that the delay of a circuit in column arbitration is reduced, the time for queuing and waiting required by the output units in the same row is reduced, and the accuracy of information contained in output data is improved.
Drawings
FIG. 1 is a basic block diagram of a conventional AER transport protocol circuit;
FIG. 2 is a block diagram of an arbiter;
fig. 3 is a diagram of a dual edge read AER transport protocol circuit architecture.
Detailed Description
A bilateral read address event representation transmission protocol circuit can adopt a structure as shown in figure 3, and comprises a row arbiter, a row address encoder, a column arbiter 1, a column arbiter 2, a column address encoder 1, a column address encoder 2, an input 2jIndividual line request (line request)<0>Request to bank<2j-1>) And 2kIndividual column request (column request)<0>Request to rank<2k-1>) The output comprises 2jIndividual line response (line response)<0>To bank response<2j-1>)、2kIndividual column response (column response)<0>Pair response<2k-1>) And 2 event information external output channels, an event output channel 1 and an event output channel 2; the connection mode of each module is as follows: line requests<0>Request to bank<2j-1>All access the row arbiter, the row arbiter outputs the acknowledge signal to the bus row acknowledge<0>To bank response<2j-1>Each row response bus is connected to a row address encoder and a return circuit array; column request<0>Request to rank<2k-1-1>A column arbiter 1 positioned at the upper left of the array is accessed, and the column arbiter 1 outputs a response signal to a bus column response<0>Pair response<2k-1-1>Each row response bus is connected to the row address encoder 1 and the return circuit array; column request<2k-1>Request to rank<2k-1>A column arbiter 2 positioned at the lower right of the array is accessed, and the column arbiter 2 outputs a response signal to the bus column response<2k-1>Pair response<2k-1>And each row reply bus is connected to the row address encoder 2 and the return circuit array.
Claims (2)
1. A two-sided read address event representation transfer protocol circuit, characterized by: the structure comprises a row arbiter, a row address encoder, a column arbiter 1, a column arbiter 2, a column address encoder 1, a column address encoder 2, an input comprising 2jIndividual line request, i.e. line request<0>Request to bank<2j-1>,2kIndividual column requests, i.e. column requests<0>Request to rank<2k-1>The output comprises 2jIndividual line response, i.e. line response<0>To bank response<2j-1>、2kIndividual column response, i.e. column response<0>Pair response<2k-1>And 2 event information external output channels, an event output channel 1 and an event output channel 2; the connection mode of each module is as follows: line requests<0>Request to bank<2j-1>All access the row arbiter, the row arbiter outputs the acknowledge signal to the bus row acknowledge<0>To bank response<2j-1>Each row response bus is connected to a row address encoder and a return circuit array; column request<0>Request to rank<2k-1-1>A column arbiter 1 positioned at the upper left of the array is accessed, and the column arbiter 1 outputs a response signal to a bus column response<0>Pair response<2k-1-1>Each row response bus is connected to the row address encoder 1 and the return circuit array; column request<2k-1>Request to rank<2k-1>A column arbiter 2 positioned at the lower right of the array is accessed, and the column arbiter 2 outputs a response signal to the bus column response<2k-1>Pair response<2k-1>And each row reply bus is connected to the row address encoder 2 and the return circuit array.
2. The dual-sided read address event representation transfer protocol circuit of claim 1, wherein: each unit needing to be output in the circuit array generates a row request on a row bus of a row where the unit is located and inputs the row request into a row arbiter; the row arbiter sequences all the generated row requests and generates response signals on a row response bus corresponding to the row request positioned at the head; the row address encoder encodes the row number of the row where the response signal is positioned(ii) a After the row response signal returns to the circuit array, the unit circuit needing to be output in the row generates a request signal on a corresponding row request bus, wherein the row number is 0 to (2)k-1-1) column requests in the range are input to the column arbiter 1 with a column number of 2k-1~(2k-1) column requests in range are input to a column arbiter 2; the column arbiter 1 and the column arbiter 2 respectively sequence the input column requests at the same time, and generate response signals on the column response bus corresponding to the first column request, that is, generate two column response signals at the same time; the column address encoder 1 encodes the column number of the column where the response signal generated by the column arbiter 1 is located, and the column address encoder 2 encodes the column number of the column where the response signal generated by the column arbiter 2 is located; at this time, the row address coded by the row address coder and the column address coded by the column address coder 1 form event information 1 to be output from an event output channel 1, and the row address coded by the row address coder and the column address coded by a column address coder 2 form event information 2 to be output from an event output channel 2; after the column response returns to the circuit array, the event information in the corresponding unit is eliminated, and the protocol circuit starts to process the next pair of units needing to be output.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911033054.7A CN112732612A (en) | 2019-10-28 | 2019-10-28 | Double-side read address event representation transmission protocol circuit |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911033054.7A CN112732612A (en) | 2019-10-28 | 2019-10-28 | Double-side read address event representation transmission protocol circuit |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112732612A true CN112732612A (en) | 2021-04-30 |
Family
ID=75588912
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201911033054.7A Pending CN112732612A (en) | 2019-10-28 | 2019-10-28 | Double-side read address event representation transmission protocol circuit |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112732612A (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103139495A (en) * | 2013-02-27 | 2013-06-05 | 天津大学 | Asynchronous pixel array with arbitration time error correction function |
CN104143181A (en) * | 2014-07-31 | 2014-11-12 | 天津大学 | Method for optimizing output of two-column TAE image sensor |
CN105611114A (en) * | 2015-11-02 | 2016-05-25 | 天津大学 | Full-digital multi-convolution core-convolution processing chip for AER (Address-Event Representation) image sensor |
CN107852379A (en) * | 2015-05-22 | 2018-03-27 | 格雷研究有限公司 | For the two-dimentional router of orientation of field programmable gate array and interference networks and the router and other circuits of network and application |
CN109547716A (en) * | 2018-10-18 | 2019-03-29 | 天津大学 | Row choosing column arbitration AER imaging sensor event transmitting device and method |
CN210781138U (en) * | 2018-06-01 | 2020-06-16 | 索尼半导体解决方案公司 | Solid-state imaging element and imaging device |
-
2019
- 2019-10-28 CN CN201911033054.7A patent/CN112732612A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103139495A (en) * | 2013-02-27 | 2013-06-05 | 天津大学 | Asynchronous pixel array with arbitration time error correction function |
CN104143181A (en) * | 2014-07-31 | 2014-11-12 | 天津大学 | Method for optimizing output of two-column TAE image sensor |
CN107852379A (en) * | 2015-05-22 | 2018-03-27 | 格雷研究有限公司 | For the two-dimentional router of orientation of field programmable gate array and interference networks and the router and other circuits of network and application |
CN105611114A (en) * | 2015-11-02 | 2016-05-25 | 天津大学 | Full-digital multi-convolution core-convolution processing chip for AER (Address-Event Representation) image sensor |
CN210781138U (en) * | 2018-06-01 | 2020-06-16 | 索尼半导体解决方案公司 | Solid-state imaging element and imaging device |
CN109547716A (en) * | 2018-10-18 | 2019-03-29 | 天津大学 | Row choosing column arbitration AER imaging sensor event transmitting device and method |
Non-Patent Citations (1)
Title |
---|
张梦醒: ""一种双列Timed-AER图像传感器后端处理系统设计"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN1104778C (en) | Convolutional interieaver and method for generating address | |
US6134624A (en) | High bandwidth cache system | |
US20170200482A1 (en) | Methods and systems for pin-efficient memory controller interface using vector signaling codes for chip-to-chip communication | |
CN1130028C (en) | Viterbi decoding apparatus and viterbi decoding method | |
CN1225492A (en) | high speed semiconductor memory device | |
US20050125590A1 (en) | PCI express switch | |
CN1965285A (en) | Apparatus and method for direct memory access in a hub-based memory system | |
CN1825479A (en) | Improved data path of double data rate type Ⅱ dynamic random access memory | |
CN102622192B (en) | A Weak Correlation Multiport Parallel Memory Controller | |
CN1766862A (en) | The microprocessor system that comprises the memory device of memory access controller and bus | |
CN1608255A (en) | Communicating transaction types between agents in a computer system using packet headers including an extended type/extended length field | |
CN1836414A (en) | Auto realignment of multiple serial byte-lanes | |
CN101038574A (en) | Bus arbitration device | |
WO2023064055A1 (en) | Internal and external data transfer for stacked memory dies | |
CN1934554A (en) | Two channel bus structure to support address information, data, and transfer qualifiers | |
CN1324499C (en) | Method for handling unexpected completion packets and completion packets with a non-successful completion status | |
CN112732612A (en) | Double-side read address event representation transmission protocol circuit | |
Du et al. | Research on XDMA high-speed data transmission architecture based on PCIe | |
WO1981002798A1 (en) | Computer system and interface therefor | |
CN101110259A (en) | FIFO memory | |
CN118069578A (en) | Data processing method, data transmission system, electronic device and storage medium | |
CN1320471C (en) | Half duplex series communication bus external device interface | |
CN201918981U (en) | Dual-phase harvard code bus signal coding-decoding circuit | |
CN1191530C (en) | PCI bridge with improved structure | |
CN101075221A (en) | Method and system for managing data stream on separation bus between bus agents |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210430 |