CN1859282A - Storage controller and control method - Google Patents
Storage controller and control method Download PDFInfo
- Publication number
- CN1859282A CN1859282A CNA2005101006612A CN200510100661A CN1859282A CN 1859282 A CN1859282 A CN 1859282A CN A2005101006612 A CNA2005101006612 A CN A2005101006612A CN 200510100661 A CN200510100661 A CN 200510100661A CN 1859282 A CN1859282 A CN 1859282A
- Authority
- CN
- China
- Prior art keywords
- memory
- module
- time
- data
- read
- 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
Images
Landscapes
- Dram (AREA)
- Time-Division Multiplex Systems (AREA)
Abstract
The present invention discloses memory controller and control method, referring to storage technology in IP network equipment. It contains command resolving and initializing module, address and control interface module, read-write data path module and read-write data interface module, which are respectively interconnected between user application layer and memory; time division multiplexing control module connected with command resolving and initializing module; first in first out data processing module connected with read-write data path module. The present invention adopts time division multiplexing principle, dividing each time slice into number of time slot, distributing each memory block read write operation to relevant time slot, and ensuring twice consecutive access time interval to same memory block greater than or equal to tRC time, raising memory read-write data efficiency. The present invention processes data in memory according to first in first out mode, raising memory interface data access efficiency.
Description
[technical field]
The present invention relates to the technical field of memory in a kind of IP network equipment, relate in particular to a kind of controller and control method of memory.
[background technology]
Along with the development of IP network, IP network data flow transmitted amount is increasing.Because the IP network data traffic is sudden, more and more much capacity, the memory of high data traffic is used in the IP network equipment, for example QDR SSRAM (Quad Data Rate Synchronous Static Random Access Memory quad data rate synchronized SRAM), RLDRAM (the low dynamic random access memory that postpones of Reduced Latency Dynamic Random Access Memory), DDR SDRAM (Double Data Rate Synchronous Dynamic Random AccessMemory Double Data Rate synchronous DRAM), SDRAM (Synchronous Dynamic RandomAccess Memory synchronous DRAM) and FCRAM (Fast Cycle Random Access Memory Rapid Cycle dynamic memory).Because QDR SSRAM Capacity Ratio is less, price is more expensive, and DDR SDRAM data rate is lower, and RLDRAM is applied in a lot of occasions.
All there is a tRC problem in these memories such as RLDRAM, DDR SDRAM, SDRAM and FCRAM, and tRC refers to the minimum interval to twice burst operation of same bank (memory block) in the memory.So when the controller of these memories of design, fully take into account the tRC problem of memory.
Shown in 1 figure, the controller of memory generally comprises following components: command analysis and initialization module, the channel module that reads and writes data, address and control interface module and the interface module that reads and writes data.Command analysis and initialization module are finished the power-up initializing of memory and the order that the user application layer face is sent here are resolved, and address and control interface module are delivered in the order that will resolve, address and control interface module are sequential and the waveform that memory can be discerned with the command conversion of resolving, thereby realize the control to memory.The channel module that reads and writes data comprises: read data path and write data path, these two paths are separate.Read data path is delivered to the user application layer face to data from the interface module reading of data that reads and writes data; The data that the write data path is sent the user application layer face here write data interface module.The interface module that reads and writes data comprises: read data interface and write data interface, data and clock that the read data interface is sent here according to memory, the data of DDR (Double Data Rate Double Data Rate) are converted to the data (data bit width doubles) of SDR (Single Data Rate haploidy number is according to speed), at last data are delivered to read data path.The SDR data that the write data interface is sent the write data path here are converted to the data (data bit width be reduced to original half) of DDR, and data are delivered to memory.
If the controller of memory is above-mentioned controller, a bigger problem then can appear, exactly because there is the tRC problem in memory itself, if double same bank to memory carries out burst operation, behind last secondary burst EO, must could carry out burst operation to same bank once more through a tRC time, so just waste the bandwidth of memory (tRC time) access data during this period of time.For example: suppose the burst length BL=2 of memory, promptly a burst operation of memory is read the number of (writing) RAM bit wide data.With bank0 is example, user application layer face back has earlier come once, writes the burst operation and the burst operation of reading bank0 of bank0, and the controller of memory is after handling the burst operation of writing bank0, a period of time (tRC time) must be waited, the burst operation of bank0 could be begun to read.
The minimum tRC=20ns of present memory, we are example with tRC=20ns, as shown in Figure 2: the t1 time point that to be last time conduct interviews to some bank of memory, the time point that t2 conducted interviews to this bank for next time, the time interval of t2 and t1 minimumly should be 20ns as can be seen, no matter how many clock frequencies of reference to storage is, all must just can carry out later at 20ns.If the clock frequency of reference to storage is 200MHz, need to wait for 4 clock cycle at least.If the clock frequency of reference to storage is 100MHz, need to wait for 2 clock cycle at least.So just waste memory in tRC access data bandwidth during this period of time, thereby reduced the read-write efficiency of memory.
[summary of the invention]
The technical problem to be solved in the present invention provides a kind of controller and control method of memory, has made full use of memory in tRC access data bandwidth during this period of time, has improved the read-write efficiency of memory.
The present invention realizes by following technical scheme:
A kind of controller of memory comprises command analysis and initialization module, address and control interface module, the channel module and the interface module that reads and writes data read and write data; Described command analysis and initialization module and address and control interface module are interconnected between user application layer face and the memory; Described channel module and the interface module that reads and writes data of reading and writing data is interconnected between user application layer face and the memory; Also comprise time division multiplexing control module and first in first out data processing module, described time division multiplexing control module is connected with described command analysis and initialization module, and described first in first out data processing module is connected with the channel module that reads and writes data; Described time division multiplexing control module the memory operation time be divided at least two interval greater than or equal the time slot of the tRC time of memory, and the read/write operation to memory block is fixed in the time slot corresponding; Described first in first out data processing module carries out first-in first-out to the data that write memory block to be handled.
Further improvement of the present invention is: described memory is low dynamic random access memory, Double Data Rate synchronous DRAM, synchronous DRAM or the Rapid Cycle dynamic memory of postponing.
A kind of control method of memory may further comprise the steps:
301, the time division multiplexing control module is divided at least one timeslice to the operating time of memory, and each timeslice is divided at least two time slots again;
302, time division multiplexing control module handle is corresponding one by one with described time slot to the specific operation type of each memory block in the memory, and preserves this corresponding relation;
What 303, command analysis and initialization module resolved that the user application layer face sends carries out the command request of specific operation to a certain memory block in the memory, obtains this memory block is carried out the pairing time slot of specific operation;
304, judge whether the current time belong to this time slot corresponding,, then carry out specific operation this memory block if belong to; If do not belong to, then do not carry out specific operation to this memory block.
Step 301 further comprises: command analysis and initialization module carry out power-up initializing to memory.
The number of time slot described in the step 301 is 2 times of number of memory block in the memory.
Timeslice described in the step 301 is greater than or equal to 2 times of tRC time of memory.
The type of specific operation described in the step 302 is read operation or write operation.
In the step 302 slot time at the read operation of a certain memory block and write operation place is greater than or equal at interval the tRC time of memory.
When carrying out the write operation to this memory block in the step 304, then before data are write this memory block, by the first in first out data processing module these data are carried out first-in first-out and handle.
Because adopted above technical scheme, the present invention is divided into equal timeslice to the time by the principle of TDM (time division multiplexing), also can not divide, as a timeslice; Owing in the memory given number bank is arranged, therefore need be divided at least two time slots to each timeslice, read-write operation to the bank of given number in the memory is assigned in the middle of the corresponding time slot, and guarantee the time of twice connected reference of same bank is greater than or equal to the tRC time, thereby can reasonably in time visit same bank once more in the time, and also can conduct interviews during this period of time to other bank at tRC, make full use of the access data bandwidth of memory, improved the efficient of memory read write data; The present invention handles being stored in the memory data according to FIFO (first in first out) mode, has guaranteed the first in first out of data, has improved the data access efficiency of memory interface; The present invention realizes simply, has solved the tRC problem that exists in the memory.
[description of drawings]
Fig. 1 be controller block diagram of the prior art and with the connection diagram of memory.
Fig. 2 is the explanation schematic diagram of tRC time.
Fig. 3 be controller block diagram of the present invention and with the connection diagram of memory.
Fig. 4 is that the access type of bank is carried out corresponding schematic diagram with time slot.
Fig. 5 is the storage format schematic diagram of message in memory after the FIFO mode is handled.
Fig. 6 is to the control method flow chart of memory among the present invention.
[embodiment]
Below in conjunction with drawings and Examples the present invention is further set forth:
As shown in Figure 3, controller of the present invention is the frame of broken lines part among the figure, and this controller is on the basis of prior art, has increased time division multiplexing control module and data fifo processing module (first in first out data processing module); Time division multiplexing control module and command analysis and initialization module interconnect, the data fifo processing module interconnects with the channel module that reads and writes data, described time division multiplexing control module is divided at least two time slots to the memory operation time, described slot time is greater than or equal to the tRC time of memory at interval, and the read/write operation to memory block is fixed in the time slot corresponding; Described first in first out data processing module carries out first-in first-out to the data that write memory block to be handled.Two paths are arranged between user application layer face and the memory, and a path is: from command analysis and initialization module to the address and the control interface module; Send control command information by this path user application layer towards memory, with the strategy of control storage storage data; Another path is: from the channel module that reads and writes data to the interface module that reads and writes data; Send the data of being stored or read in the memory data of being stored towards memory by this path user application layer; This path is divided into separate read data path and write data path, and data write is not disturbed mutually.
Time division multiplexing control module among the present invention is based on time-multiplexed principle, the memory operation time is divided into equal timeslice, again each timeslice is divided at least two time slots, a multiple time slot that also can be divided into bank number or bank group number, wherein, comprise at least two bank in the bank group, and the read-write operation of each bank is assigned in the middle of the corresponding time slot, and guarantee the time of twice connected reference of same bank is greater than or equal to the tRC time, thereby can reasonably in time visit same bank once more in the time, simultaneously, also can conduct interviews during this period of time at tRC to other bank.
Being example below with RLDRAM, time slot is divided into 16, just is 2 times of bank number, and principle of the present invention is specifically described:
Suppose the tRC=40ns of memory, clock frequency is 100MHz, BL=4, memory has 8 Bank, because all being the data bit width according to memory, reads each memory read data, and a Burst (burst operation) comprises BL the read-write to same bank, so when reading and writing memory data, so that (bit wide of RLDRAM * BL) is a unit.As shown in Figure 4, be the action type of bank and the corresponding relation figure of time slot.In the drawings, W represents memory is carried out write operation, and R represents memory is carried out read operation.As can be seen, when time slot time0, bank0 is carried out write operation, bank3 is carried out read operation, when time slot time1, bank0 is carried out write operation, bank4 is carried out read operation, when time slot time2, bank1 is carried out write operation, bank4 is carried out read operation, when time slot time3, bank1 is carried out write operation, bank5 is carried out read operation ... and the like, when time slot time15, bank7 is carried out write operation, bank3 is carried out read operation.For bank0, its write operation carries out at time0, and its read operation is carried out at time9, therefore must guarantee between time9 and the time0 interval greater than or equal the tRC time, so no matter what kind of the clock cycle is, can utilize tRC during this period of time in, during time1, to the read operation of bank4, bank0 is carried out write operation; During time2, bank1 is carried out write operation, bank4 is carried out read operation ..., until time9 just can carry out read operation to bank0.
Time slot is produced by the counter of a 4bit.The read and write operation always is in a definite bank.If need carry out read/write operation, then can only when belonging to the read/write time slot of this bank, carry out some bank; If will carry out the read/write operation of a certain bank but find not belong to the read/write time slot of this bank this moment, wait for that then one's own time slot arrives, just can carry out read/write operation to this bank, so just can evade the tRC problem.
Number for time slot, might not require is 2 times of bank number, it also can be other number, but be more than two at least, as long as guarantee twice burst operation of same bank (once read, once write) between interval greater than or equal the tRC time that memory itself exists, that's all.
Fig. 5 is through message the storage format schematic diagram in memory of FIFO (First In First Out first in first out) after mode is handled, in order to make the memory read write data uninterrupted, the memory interface data access efficiency is the highest, and the data that are stored in the memory are handled according to byte level FIFO mode: promptly the order of the data of reading from memory and the sequencing of the data of sending here from the user application layer face are consistent.The benefit of doing like this: have only the user application layer face that data are arranged, just can continual write memory.Simultaneously, if data are arranged in the memory, just can be continual from memory reading of data, deliver to the user application layer face.Among the figure, " * " expression invalid data, each little lattice is represented the data of a byte (byte), a newspaper of the numeral in each little lattice code, identical newspaper code represents to store the data of same message.The address that the data of first 32bits (position) of each message (in addition also have 4bits be the message control data) leave memory in is the place of 4n (n is an integer), i.e. first 32bits deposit data first row of form in the drawings.If the byte number of a message during less than 16m (m is a natural number), then uses invalid data " * " to replace, replenish enough 16m.The number of " 1 " is 9 among the figure, then needs to replenish 7 invalid datas " * "; The number of " 2 " is 10, then needs to replenish 6 invalid datas " * "; The number of supposing " 3 " is 20, then need replenish 12 invalid datas " * ".
As shown in Figure 6, be to the control method flow chart of memory among the present invention.Controller is as follows to the detailed process of the right control method of memory:
Finish power-up initializing by command analysis and initialization module to memory, after finishing initialization, by the time division multiplexing module time is divided into equal timeslice, again each timeslice is divided into 2 times of time slots to memory bank or bank group number, guarantee the tRC time of each timeslice simultaneously more than or equal to this memory of 2 times, action type to each bank of memory is all corresponded in each time slot one by one, and preserve this corresponding relation.Because except read operation, is exactly write operation to the action type of each bank, therefore, the specific operation type of each bank can both be carried out corresponding with a time slot.
Command analysis and initialization module detect the user application layer face and whether send the request of memory being carried out burst operation, if do not have, then continue to detect; If have, then the request command that the user application layer face is sent is resolved, and draws the purpose of this request:
If the purpose of this request is that bank1 is carried out write operation, then find: must just can carry out when time slot time2 and the time3 the write operation of bank1 by the time division multiplexing module, if the current time does not also arrive time slot time2, then do not carry out write operation to bank1, if current time is up time slot time2, then carry out write operation to bank1, simultaneously, be about to write the data of bank1 when process reads and writes data path, by the data fifo processing module these data are handled, when reading these data, can be read out after making, reach the purpose of data first in first out according to the sequencing that these data deposit memory in;
If the purpose of this request is that bank1 is carried out read operation, then find: must just can carry out when time slot time11 and the time12 the read operation of bank1 by the time division multiplexing module, if the current time does not also arrive time slot time11, then do not carry out read operation to bank1, if current time is up time slot time11, then carry out the read operation to bank1, the data of reading are delivered to the user application layer face through the interface module and the channel module that reads and writes data of reading and writing data.
Through said process, to have finished memory has been carried out the control of read-write operation to a certain bank, the control procedure and the said process that other bank are carried out read-write operation are similar.
Key of the present invention is: guarantee twice burst operation of same bank (once read, once write) between interval greater than or equal the tRC time that memory itself exists, and in this section tRC time, can also carry out corresponding operating to other bank, made full use of the access data bandwidth of memory, improve the efficient of memory read write data, thus the problem of the tRC time of avoiding running in the prior art; And on the form of storage, adopt the storage format of FIFO mode, improved the data access efficiency of memory interface.
Claims (9)
1, a kind of controller of memory comprises command analysis and initialization module, address and control interface module, the channel module and the interface module that reads and writes data read and write data; Described command analysis and initialization module and address and control interface module are interconnected between user application layer face and the memory; Described channel module and the interface module that reads and writes data of reading and writing data is interconnected between user application layer face and the memory; It is characterized in that: also comprise time division multiplexing control module and first in first out data processing module, described time division multiplexing control module is connected with described command analysis and initialization module, and described first in first out data processing module is connected with the channel module that reads and writes data; Described time division multiplexing control module the memory operation time be divided at least two interval greater than or equal the time slot of the tRC time of memory, and the read/write operation to memory block is fixed in the time slot corresponding; Described first in first out data processing module carries out first-in first-out to the data that write memory block to be handled.
2, the controller of a kind of memory according to claim 1 is characterized in that: described memory is low dynamic random access memory, Double Data Rate synchronous DRAM, synchronous DRAM or the Rapid Cycle dynamic memory of postponing.
3, a kind of control method of memory is characterized in that, may further comprise the steps:
301, the time division multiplexing control module is divided at least one timeslice to the operating time of memory, and each timeslice is divided at least two time slots again;
302, time division multiplexing control module handle is corresponding one by one with described time slot to the specific operation type of each memory block in the memory, and preserves this corresponding relation;
What 303, command analysis and initialization module resolved that the user application layer face sends carries out the command request of specific operation to a certain memory block in the memory, obtains this memory block is carried out the pairing time slot of specific operation;
304, judge whether the current time belong to this time slot corresponding,, then carry out specific operation this memory block if belong to; If do not belong to, then do not carry out specific operation to this memory block.
4, the control method of a kind of memory according to claim 3 is characterized in that, step 301 further comprises: command analysis and initialization module carry out power-up initializing to memory.
5, the control method of a kind of memory according to claim 3 is characterized in that: the number of time slot described in the step 301 is 2 times of number of memory block in the memory.
6, the control method of a kind of memory according to claim 3 is characterized in that: timeslice described in the step 301 is more than or equal to 2 times of tRC time of memory.
7, the control method of a kind of memory according to claim 3 is characterized in that: the type of specific operation described in the step 302 is read operation or write operation.
8, the control method of a kind of memory according to claim 6 is characterized in that: the tRC time that in the step 302 slot time at the read operation of a certain memory block and write operation place is greater than or equal at interval memory.
9, the control method of a kind of memory according to claim 7, it is characterized in that: when carrying out the write operation to this memory block in the step 304, then before data are write this memory block, by the first in first out data processing module these data are carried out first-in first-out and handle.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CNB2005101006612A CN100426793C (en) | 2005-10-22 | 2005-10-22 | Storage controller and control method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CNB2005101006612A CN100426793C (en) | 2005-10-22 | 2005-10-22 | Storage controller and control method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN1859282A true CN1859282A (en) | 2006-11-08 |
CN100426793C CN100426793C (en) | 2008-10-15 |
Family
ID=37298146
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CNB2005101006612A Active CN100426793C (en) | 2005-10-22 | 2005-10-22 | Storage controller and control method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN100426793C (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101621474B (en) * | 2009-08-11 | 2011-11-16 | 杭州华三通信技术有限公司 | Method and apparatus for realizing high linear speed of packet access to memory |
CN102411982A (en) * | 2010-09-25 | 2012-04-11 | 杭州华三通信技术有限公司 | Memory controller and method for controlling commands |
CN101324863B (en) * | 2007-06-12 | 2012-07-04 | 中兴通讯股份有限公司 | Device and method for controlling synchronous static memory |
CN102984088A (en) * | 2012-12-13 | 2013-03-20 | 中国航空无线电电子研究所 | Method applied to AFDX exchanger for ensuring uniformity of frame forwarding sequence |
WO2013170412A1 (en) * | 2012-05-14 | 2013-11-21 | 华为技术有限公司 | Memory scheduling method and memory controller |
CN103414658A (en) * | 2013-08-05 | 2013-11-27 | 北京华为数字技术有限公司 | System and method for lookup tables and packet buffer to share dynamic random access memories |
CN104021095A (en) * | 2013-02-28 | 2014-09-03 | 佳能株式会社 | Data processing apparatus and control method |
CN105701042A (en) * | 2016-03-02 | 2016-06-22 | 珠海全志科技股份有限公司 | Optimizing method and system for signal quality of memory control interface |
CN109712665A (en) * | 2018-02-27 | 2019-05-03 | 上海安路信息科技有限公司 | The function test method of memory and memory |
CN110415163A (en) * | 2019-06-28 | 2019-11-05 | 中国科学院电子学研究所 | Data matrix transposition method and device for SAR imaging |
CN111052099A (en) * | 2017-08-31 | 2020-04-21 | 美光科技公司 | In-memory processing |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101729421B (en) * | 2009-12-11 | 2013-03-20 | 杭州华三通信技术有限公司 | Storage method and device based on time division multiplex |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
SE508876C2 (en) * | 1992-06-24 | 1998-11-09 | Ericsson Telefon Ab L M | Digital selector intended for use in a node in a circuit-switched telecommunications network with distributed selector function |
US6453370B1 (en) * | 1998-11-16 | 2002-09-17 | Infineion Technologies Ag | Using of bank tag registers to avoid a background operation collision in memory systems |
CN1112789C (en) * | 1999-04-05 | 2003-06-25 | 深圳市中兴通讯股份有限公司 | Digital switch unit for sequential writing on and controlled reading from multiple page |
-
2005
- 2005-10-22 CN CNB2005101006612A patent/CN100426793C/en active Active
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101324863B (en) * | 2007-06-12 | 2012-07-04 | 中兴通讯股份有限公司 | Device and method for controlling synchronous static memory |
CN101621474B (en) * | 2009-08-11 | 2011-11-16 | 杭州华三通信技术有限公司 | Method and apparatus for realizing high linear speed of packet access to memory |
CN102411982B (en) * | 2010-09-25 | 2014-12-10 | 杭州华三通信技术有限公司 | Memory controller and method for controlling commands |
CN102411982A (en) * | 2010-09-25 | 2012-04-11 | 杭州华三通信技术有限公司 | Memory controller and method for controlling commands |
CN103703448B (en) * | 2012-05-14 | 2017-07-14 | 华为技术有限公司 | A kind of scheduling memory method and Memory Controller Hub |
WO2013170412A1 (en) * | 2012-05-14 | 2013-11-21 | 华为技术有限公司 | Memory scheduling method and memory controller |
US9514799B2 (en) | 2012-05-14 | 2016-12-06 | Huawei Technologies Co., Ltd. | Memory scheduling method and memory controller |
CN103703448A (en) * | 2012-05-14 | 2014-04-02 | 华为技术有限公司 | Memory scheduling method and memory controller |
CN102984088A (en) * | 2012-12-13 | 2013-03-20 | 中国航空无线电电子研究所 | Method applied to AFDX exchanger for ensuring uniformity of frame forwarding sequence |
CN102984088B (en) * | 2012-12-13 | 2015-10-07 | 中国航空无线电电子研究所 | Be applied to AFDX switch and guarantee the conforming method of frame transfer sequence |
CN104021095A (en) * | 2013-02-28 | 2014-09-03 | 佳能株式会社 | Data processing apparatus and control method |
CN103414658B (en) * | 2013-08-05 | 2016-08-10 | 北京华为数字技术有限公司 | Look-up table and packet buffer share the system and method for dynamic RAM |
CN103414658A (en) * | 2013-08-05 | 2013-11-27 | 北京华为数字技术有限公司 | System and method for lookup tables and packet buffer to share dynamic random access memories |
CN105701042A (en) * | 2016-03-02 | 2016-06-22 | 珠海全志科技股份有限公司 | Optimizing method and system for signal quality of memory control interface |
CN105701042B (en) * | 2016-03-02 | 2018-08-14 | 珠海全志科技股份有限公司 | The optimization method and system of Memory control interface signal quality |
CN111052099A (en) * | 2017-08-31 | 2020-04-21 | 美光科技公司 | In-memory processing |
CN111052099B (en) * | 2017-08-31 | 2021-04-27 | 美光科技公司 | In-memory processing |
CN109712665A (en) * | 2018-02-27 | 2019-05-03 | 上海安路信息科技有限公司 | The function test method of memory and memory |
CN109712665B (en) * | 2018-02-27 | 2020-09-15 | 上海安路信息科技有限公司 | Memory and function test method of memory |
CN110415163A (en) * | 2019-06-28 | 2019-11-05 | 中国科学院电子学研究所 | Data matrix transposition method and device for SAR imaging |
Also Published As
Publication number | Publication date |
---|---|
CN100426793C (en) | 2008-10-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN1859282A (en) | Storage controller and control method | |
KR101467623B1 (en) | Independently controlled virtual memory devices in memory modules | |
CN101916227B (en) | RLDRAM SIO storage access control method and device | |
CN1977336A (en) | System and method for improving performance in computer memory systems supporting multiple memory access latencies | |
TW200601742A (en) | Compact packet switching node storage architecture employing double data rate synchronous dynamic ram | |
CN1519853A (en) | On-die termination circuit and method for reducing on-chip DC current, and memory system including memory device having same | |
CN103780506A (en) | Data caching system and data caching method | |
CN101324869B (en) | Multiplexor based on AXI bus | |
JP3678412B2 (en) | DRAM data storage and migration for network processors | |
CN1815625A (en) | Synchronous dynamic storage controller designing method | |
CN1993687A (en) | Message administrator and method for controlling access to data of the message memory of a communications component | |
CN100550189C (en) | The device of data storing method, mapping affairs and the method for write data line | |
CN106951182A (en) | A kind of block device caching method and device | |
CN106599301A (en) | Multi-client concurrent data read-write accelerating method and device | |
US10067868B2 (en) | Memory architecture determining the number of replicas stored in memory banks or devices according to a packet size | |
CN100359492C (en) | System, apparatus, and method for a flexible DRAM architecture | |
CN101656586A (en) | Method and device for improving virtual concatenation delay compensation caching efficiency in synchronous digital hierarchy | |
CN1851669A (en) | Method for improving storage access efficiency and storage coutroller | |
CN1409856A (en) | Usage of SDRAM as storage for correction and track buffering in frontend ICS of optical recording or reproducing devices | |
US20130064022A1 (en) | Method and Apparatus for Memory Access | |
CN100561588C (en) | A kind of method that improves the RAM read-write efficiency | |
CN1232909C (en) | Method and structure for realizing router flow management chip buffer-storage management | |
CN101520750B (en) | Method for storing a plurality of data in SDRAM | |
CN1242659A (en) | Single chip and large capacity digital time division switching network | |
CN117055817A (en) | Data packet memory device and storage method based on single-port RAM |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |