[go: up one dir, main page]

CN102157192A - Method for compensating memory particle data delay of synchronous dynamic random access memory - Google Patents

Method for compensating memory particle data delay of synchronous dynamic random access memory Download PDF

Info

Publication number
CN102157192A
CN102157192A CN201010598448XA CN201010598448A CN102157192A CN 102157192 A CN102157192 A CN 102157192A CN 201010598448X A CN201010598448X A CN 201010598448XA CN 201010598448 A CN201010598448 A CN 201010598448A CN 102157192 A CN102157192 A CN 102157192A
Authority
CN
China
Prior art keywords
memory
random access
dynamic random
clock
synchronous dynamic
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
Application number
CN201010598448XA
Other languages
Chinese (zh)
Inventor
李静
白宗元
张磊
纪奎
张英文
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Dawning Information Industry Co Ltd
Original Assignee
Dawning Information Industry Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Dawning Information Industry Co Ltd filed Critical Dawning Information Industry Co Ltd
Priority to CN201010598448XA priority Critical patent/CN102157192A/en
Publication of CN102157192A publication Critical patent/CN102157192A/en
Pending legal-status Critical Current

Links

Images

Landscapes

  • Dram (AREA)

Abstract

The invention provides a method for compensating memory particle data delay of an SDRAM (Synchronous Dynamic Random Access Memory), which is realized by utilizing an FPGA (Field Programmable Gate Array). Each mainboard is provided with a plurality of memory controllers; each memory controller is used for controlling read-write of four memories and measuring working clock of each memory; and the FPGA is used for determining a phase of the working clock according to the position of each memory and regulating time delay error brought by different PCB (Printed Circuit Board) wirings of the memory positions through regulating the phase of the clock. In the design of a high-capacity multi-memory controller system, the wiring error caused by the PCB wirings can be greatly eliminated, the PCB wiring difficulty is decreased, and the design risk is reduced.

Description

A kind of method that compensates source Synchronous Dynamic Random Access Memory internal memory particle data delay
Technical field
The present invention relates to the Memory Controller Hub design field, be specifically related to the method for a kind of compensation source Synchronous Dynamic Random Access Memory (Synchronous Dynamic Random Access Memory is called for short SDRAM) internal memory particle data delay.
Background technology
In high capacity, multi-memory system, signal integrity is the key of system design.Because Memory Controller Hub is inconsistent to the time that the inconsistent meeting of the track lengths between the internal memory causes arriving from the signal that Memory Controller Hub sends different sdram memory particles, walk SDRAM acknowledge(ment) signal evening time of line length with Memory Controller Hub, cabling short SDRAM acknowledge(ment) signal time, early the serious inconsistency of this cabling length significantly limited the frequency of operation of high capacity, multi-memory system.
In Memory Controller Hub design at present, in order to solve owing to the signal routing access delay of bringing different in size, cabling length to signal wire when wiring requires relatively stricter, the length gap must be controlled in the effective range, brought very big difficulty to PCB layout in high capacity, multi-memory system, design margin is very little.
Number of patent application is the method that 200510137024.2 patent " Memory Controller Hub and control method thereof " has been described the internal storage access delayed impact that a kind of PCB of reducing cabling brings, but adopt this method can obtain 8 clocks with system clock same frequency out of phase, each clock is compared with system clock and is postponed 1/8 clock period respectively.Memory Controller Hub can be selected different internal memory work clock phase places according to the different wire delays of walking, and can dynamically adjust according to postponing to change.But this method only is applicable under the situation of having only an internal memory that each internal memory can only be selected a phase place.And in multi-memory system (〉=4), this method is also inapplicable, in multi-memory system, many internal memory can be worked simultaneously, can not adopt the work clock of same phase place, article one internal memory and second internal memory promptly can occur, all different situation of phase place of the 3rd internal memory, can not lack dirigibility to these several memory shared clock phases like this.
Summary of the invention
The present invention utilizes field programmable logic device (FPGA) to realize the design of high capacity, many internal memories Memory Controller Hub, owing to the dirigibility of FPGA design, remedies the delayed impact of PCB cabling by the phase place of regulating every memory read-write clock.
A kind of method that compensates source Synchronous Dynamic Random Access Memory internal memory particle data delay is characterized in that: adopt FPGA to realize that process is as follows:
A, every mainboard assembling Memory Controller Hub;
B, each Memory Controller Hub are responsible for controlling the read-write of 4 internal memories;
The work clock of C, the every root memory of mensuration;
D, FPGA are according to the phase place of its work clock of determining positions of each internal memory, and its clock phase is regulated because the delay time error that the different PCB layout in memory bar position are brought by regulating.
A kind of optimal technical scheme of the present invention is: when the PCB cabling error of two or many internal memories differs very little, can regulate these memory shared identical clock and phase places.
Another kind of optimal technical scheme of the present invention is: described D step for measuring the track lengths of every root memory, is calculated conduction time according to the conduction of current velograph according to core position decision work clock phase method for position, sets clock phase then.
Another optimal technical scheme of the present invention is: the Memory Controller Hub that assembles on every mainboard is at least one.
The present invention can eliminate the wiring error of bringing owing to the PCB cabling greatly in high capacity, the system design of many internal memories Memory Controller Hub, less PCB layout difficulty reduces the design risk.Can be according to the phase place of the every memory read-write clock of length dynamic adjustments of PCB layout, and then the time-delay of each signal wire of adjusting internal memory, can accomplish each memory bar is disposed a clock phase separately, also can be configured to identical phase place, to satisfy sequential requirement to high capacity, many memory read-writes.
Description of drawings
Fig. 1 is the invention process process
Embodiment
As shown in Figure 1, each Memory Controller Hub is responsible for the read-write of 4 internal memories, every internal memory all has the clock clk0 of its work separately, clk1,, clk7 is because different in size cause of signal wire from FPGA to each internal memory, cause the time of signal arrival internal memory particle inconsistent, can cause the read-write of internal memory to make mistakes like this.FPGA is according to the phase place of its work clock of determining positions of each internal memory when design, and its clock phase is regulated because the error time-delay that the different PCB layout in memory bar position are brought by regulating.Nearest such as DIMM 0 from FPGA, the time-delay of PCB cabling is minimum, and the like, DIMM3 from FPGA farthest, PCB cabling time-delay is maximum, by FPGA when each internal memory provides work clock, the clock phase of DIMM0 can be postponed to transfer to maximum, the clock phase of DIMM3 postpones to transfer to minimum, like this by the clock phase adjustment, guarantee that signal wire can arrive each memory bar simultaneously, the influence that postpones to bring is reduced as far as possible.
Can also determine the phase place of each internal memory work clock according to the concrete condition of PCB cabling, differ very little such as PCB cabling error as if DIMM0 and DIMM1, like this when adjusting the clock phase of DIMM0 and DIMM1, can allow shared identical clock of DIMM0 and DIMM1 and phase place thereof, DIMM2 and DIMM3 and the like, best situation is that the PCB cabling of four internal memories is very nearly the same, and Memory Controller Hub can be these 4 memory shared same work clocks when design like this, needn't adjust successively separately.

Claims (4)

1. method that compensates source Synchronous Dynamic Random Access Memory internal memory particle data delay is characterized in that: adopt FPGA to realize that process is as follows:
A, every mainboard assembling Memory Controller Hub;
B, each Memory Controller Hub are responsible for controlling the read-write of 4 internal memories;
The work clock of C, the every root memory of mensuration;
D, FPGA are according to the phase place of its work clock of determining positions of each internal memory, and its clock phase is regulated because the delay time error that the different PCB layout in memory bar position are brought by regulating.
2. a kind of according to claim 1 method that compensates source Synchronous Dynamic Random Access Memory internal memory particle data delay, it is characterized in that: when the PCB cabling error of two or many internal memories differs very little, can regulate these memory shared identical clock and phase places.
3. a kind of according to claim 1 method that compensates source Synchronous Dynamic Random Access Memory internal memory particle data delay, it is characterized in that: described D step is according to the track lengths of core position decision work clock phase method for position for the every root memory of mensuration, calculate conduction time according to the conduction of current velograph, set clock phase then.
4. a kind of according to claim 1 method that compensates source Synchronous Dynamic Random Access Memory internal memory particle data delay is characterized in that: the Memory Controller Hub that assembles on every mainboard is at least one.
CN201010598448XA 2010-12-17 2010-12-17 Method for compensating memory particle data delay of synchronous dynamic random access memory Pending CN102157192A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201010598448XA CN102157192A (en) 2010-12-17 2010-12-17 Method for compensating memory particle data delay of synchronous dynamic random access memory

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201010598448XA CN102157192A (en) 2010-12-17 2010-12-17 Method for compensating memory particle data delay of synchronous dynamic random access memory

Publications (1)

Publication Number Publication Date
CN102157192A true CN102157192A (en) 2011-08-17

Family

ID=44438628

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201010598448XA Pending CN102157192A (en) 2010-12-17 2010-12-17 Method for compensating memory particle data delay of synchronous dynamic random access memory

Country Status (1)

Country Link
CN (1) CN102157192A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102521142A (en) * 2011-12-13 2012-06-27 曙光信息产业(北京)有限公司 Method for improving access efficiency of high-capacity and multiple-memory device
CN106409332A (en) * 2015-07-29 2017-02-15 瑞萨电子株式会社 Electronic device
CN107180000A (en) * 2016-03-10 2017-09-19 华为技术有限公司 storage device and data access method
CN112397109A (en) * 2020-11-19 2021-02-23 成都海光集成电路设计有限公司 Time delay compensation method for memory signal line
CN116722946A (en) * 2023-08-08 2023-09-08 浙江宜通华盛科技有限公司 Scalable synchronous clock tree system and phased array radar

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101458329A (en) * 2009-01-06 2009-06-17 清华大学 Time-domain paralleling sampling rate transformation method

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101458329A (en) * 2009-01-06 2009-06-17 清华大学 Time-domain paralleling sampling rate transformation method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
吕喜在,黄芝平,苏绍璟: "基于FPGA的宽带数据采集时钟相位校正方法", 《计算机测量与控制》 *

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102521142A (en) * 2011-12-13 2012-06-27 曙光信息产业(北京)有限公司 Method for improving access efficiency of high-capacity and multiple-memory device
CN102521142B (en) * 2011-12-13 2015-05-13 曙光信息产业(北京)有限公司 Method for improving access efficiency of high-capacity and multiple-memory device
CN106409332A (en) * 2015-07-29 2017-02-15 瑞萨电子株式会社 Electronic device
CN106409332B (en) * 2015-07-29 2021-06-08 瑞萨电子株式会社 electronic device
CN107180000A (en) * 2016-03-10 2017-09-19 华为技术有限公司 storage device and data access method
CN107180000B (en) * 2016-03-10 2020-04-14 华为技术有限公司 Storage device and data access method
CN112397109A (en) * 2020-11-19 2021-02-23 成都海光集成电路设计有限公司 Time delay compensation method for memory signal line
CN112397109B (en) * 2020-11-19 2023-05-09 成都海光集成电路设计有限公司 Method for compensating time delay of memory signal line
CN116722946A (en) * 2023-08-08 2023-09-08 浙江宜通华盛科技有限公司 Scalable synchronous clock tree system and phased array radar
CN116722946B (en) * 2023-08-08 2023-12-05 浙江宜通华盛科技有限公司 Scalable synchronous clock tree system and phased array radar

Similar Documents

Publication Publication Date Title
CN102157192A (en) Method for compensating memory particle data delay of synchronous dynamic random access memory
EP2778943B1 (en) Integrated delayed clock for high speed isolated spi communication
US5896347A (en) Semiconductor memory system using a clock-synchronous semiconductor device and semiconductor memory device for use in the same
US8760901B2 (en) Semiconductor device having a control chip stacked with a controlled chip
CN100437833C (en) Calibrating method and memory system
US6397312B1 (en) Memory subsystem operated in synchronism with a clock
CN102637155B (en) Method for configuring data strobe signal delays in DDR3 (double data rate) through training and correcting
US9143140B2 (en) Multi-function delay locked loop
JP4416580B2 (en) Delay control device
KR100273082B1 (en) Clock generator having dll and semiconductor device having clock generator
US8762657B2 (en) Method and system for synchronizing address and control signals in threaded memory modules
US7542371B2 (en) Memory controller and memory system
US20130038380A1 (en) Iimplementing chip to chip calibration within a tsv stack
EP3215950B1 (en) Calibration in a control device receiving from a source synchronous interface
EP3210124B1 (en) Dynamic selection of output delay in a memory control device
CN101859286A (en) The active training of memory command timing
US9236111B2 (en) Semiconductor device
CN1484793A (en) Digital system of adjusting delays on circuit boards
US9349488B2 (en) Semiconductor memory apparatus
JP2008028345A (en) Semiconductor integrated circuit including malfunction detection circuit and design method thereof
EP2053487A2 (en) Electric circuit and method for designing electric circuit
WO2023129399A1 (en) Multi-chiplet clock delay compensation
US9158330B1 (en) Apparatus and method to compensate for data skew for multiple memory devices and adjust delay for individual data lines based on an optimized critical window
US8705296B2 (en) Memory controller and control method
JP2000048569A (en) Clock synchronizing memory

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C12 Rejection of a patent application after its publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20110817