WO2022041198A1 - 一种三维堆叠存储芯片的温度变化计算方法 - Google Patents

一种三维堆叠存储芯片的温度变化计算方法 Download PDF

Info

Publication number
WO2022041198A1
WO2022041198A1 PCT/CN2020/112484 CN2020112484W WO2022041198A1 WO 2022041198 A1 WO2022041198 A1 WO 2022041198A1 CN 2020112484 W CN2020112484 W CN 2020112484W WO 2022041198 A1 WO2022041198 A1 WO 2022041198A1
Authority
WO
WIPO (PCT)
Prior art keywords
nodes
node
temperature
memory chip
chip
Prior art date
Application number
PCT/CN2020/112484
Other languages
English (en)
French (fr)
Inventor
王毅
王先华
廖好
周池
毛睿
Original Assignee
深圳大学
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 深圳大学 filed Critical 深圳大学
Priority to PCT/CN2020/112484 priority Critical patent/WO2022041198A1/zh
Publication of WO2022041198A1 publication Critical patent/WO2022041198A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11C—STATIC STORES
    • G11C29/00—Checking stores for correct operation ; Subsequent repair; Testing stores during standby or offline operation
    • G11C29/04—Detection or location of defective memory elements, e.g. cell constructio details, timing of test signals
    • G11C29/50—Marginal testing, e.g. race, voltage or current testing

Definitions

  • the invention relates to the technical field of temperature measurement, in particular to a temperature change calculation method for three-dimensional stacked memory chips.
  • the prior art proposes various stability optimization solutions for the three-dimensional stacked memory chips based on temperature information. These optimization schemes are all monitoring the temperature information of the three-dimensional stacked memory chips through temperature sensors or thermal cameras. This method of obtaining temperature information has the following shortcomings:
  • the temperature sensor or thermal camera must keep running and collect temperature information when the memory chip is working, thus generating more power consumption;
  • the chip temperature information obtained by using the temperature sensor or thermal camera needs to be continuously transmitted to the memory chip controller, which occupies a large amount of memory chip bandwidth;
  • the technical problem to be solved by the present invention is to overcome the defects of high power consumption, occupy a large amount of memory chip bandwidth, and low temperature measurement accuracy when measuring the temperature of three-dimensional stacked memory chips in the prior art, thereby providing a temperature measurement of three-dimensional stacked memory chips. Change calculation method.
  • the present invention provides the following technical solutions:
  • an embodiment of the present invention provides a temperature change calculation method for a three-dimensional stacked memory chip, including the following steps:
  • the surrounding nodes are uniformly selected for each node in the three-dimensional physical model, and the thermal conductivity differential equation is established based on the chip performance parameters and the ambient temperature, and the temperature information of each node at the next moment is obtained by solving it as the program simulation temperature; Repeat the above steps during the operation of the memory chip, And periodically start the temperature measuring device to measure the real-time temperature of the memory chip to replace the program simulation temperature, so as to correct the temperature field.
  • the physical structure parameters include the number of physical blocks, the number of physical pages in each physical block, and the specifications of the physical pages;
  • the chip performance parameters include: chip thermal conductivity, chip specific heat capacity, chip density, and chip-air convection heat transfer coefficient.
  • the constructing a three-dimensional physical model according to the physical structure parameters includes:
  • a three-dimensional Cartesian coordinate system is established and the physical page is used as the basic unit to divide the memory chip into multiple nodes, and assign unique three-dimensional coordinates according to the spatial position of each node.
  • each node in the three-dimensional physical model evenly selects its surrounding nodes, and establishes a thermal conductivity differential equation based on chip performance parameters and ambient temperature, and solves to obtain the temperature information of each node at the next moment as the program simulation temperature;
  • the above steps are repeated, and the temperature measurement device is periodically started to measure the real-time temperature of the memory chip to replace the simulated temperature of the program for correcting the temperature field, including:
  • Step S01 obtain the initial temperature information of the memory chip by the temperature measuring device, and initialize the corrected counter Ct to zero;
  • Step S02 judge whether the operation of reading/writing/erasing the memory chip occurs at this time
  • Step S03 if an operation occurs, update the temperature information according to the type of operation that occurs in the chip and its target physical page node, and continue to step S04; if no operation occurs, continue to step S04;
  • Step S04 for the target physical page node to calculate the temperature value at the next moment, select some nodes from its surrounding nodes through the first preset rule;
  • Step S05 according to the second preset rule, the number of selected nodes is uniform in each area divided by the target physical page node by adding nodes and deleting nodes;
  • Step S06 Analyze the thermal conductivity relationship between all selected nodes and the target node, and obtain the corresponding thermal conductivity differential equation, solve the thermal conductivity differential equation by substituting the chip performance parameters and the current ambient temperature, and obtain the target physical page node temperature at the next moment. value as program simulation temperature;
  • Step S07 make the counter Ct accumulate random integer values in the range of 1 to 3, and judge whether Ct is greater than the correction threshold N;
  • Step S08 if the counter Ct is greater than N, then start the temperature measuring device to measure the real-time temperature information of the memory chip, use the real-time temperature substitute program simulation temperature to correct the temperature field, and set Ct to zero, then continue step S09; if the counter Ct is less than N, then continue to step S09;
  • Step S09 determine whether the memory chip stops working; if the memory chip stops working, end the operation; if the memory chip does not stop working, return to step S02.
  • the physical page nodes around the target physical page node are classified into three types of nodes according to their Euclidean distances from the coordinates of the target physical page node:
  • the nodes whose Euclidean distance between the coordinates and the coordinates of the target node is less than or equal to the first threshold are classified as the first type of nodes;
  • the nodes whose Euclidean distance between the coordinates and the coordinates of the target node is greater than the first threshold and less than or equal to the second threshold are classified as the second type of nodes;
  • the nodes whose Euclidean distance between the coordinates and the coordinates of the target node is greater than the second threshold are classified as the third type of nodes;
  • A% of the nodes in the first type of nodes, B% of the nodes in the second type of nodes, and C% of the nodes in the third type of nodes where A, B, and C satisfy A>aB, B>bC , A+B+C ⁇ S, where a, b, and s are all positive integers, and a ⁇ b ⁇ s.
  • the process of selecting some nodes from the surrounding nodes by the first preset rule includes:
  • Step S11 Allocating a set of serial numbers increasing from 0 to all nodes in each type of nodes according to the coordinate sequence;
  • Step S12 Assuming that there are M nodes in total, randomly generate an integer random number in the range of 0 to M-1, and find the node whose serial number corresponds to the random number;
  • Step S13 judging whether the node is already in the selected node set
  • Step S14 if the node is already in the selected node set, return to step S12;
  • Step S15 if the node is not in the selected set, add the node and the symmetrical node centered on the target node to the selected node set;
  • Step S16 judging whether the number of this type of nodes in the currently selected node set accounts for the percentage of this type of nodes that reaches the required percentage of selection;
  • Step S17 if the proportion of this type of nodes in the selected node set to this type of nodes does not meet the requirement, then return to step S12;
  • Step S18 If the proportion of the number of nodes of this type in the selected node set to the nodes of this type meets the requirement, the selection of this type of nodes is completed.
  • the process of making the number of selected nodes uniform in each area divided with the target physical page node as the center by adding nodes and deleting nodes to the selected nodes according to the second preset rule including:
  • the process of adding/deleting node operations to the selected node set, so that the number of nodes in the selected node set in all areas is t includes:
  • Step S21 judge whether the selected node number r is equal to t in the area
  • Step S22 if the number of selected nodes in the area is equal to t, then end the operation;
  • Step S23 if the selected node in the area is not equal to t, then judge whether r is greater than t;
  • Step S24 if r is less than t, then the t-r unselected nodes with the minimum Euclidean distance from the target node in the area are added to the selected node set, and the operation is ended;
  • Step S25 if r is greater than t, then assign the selected nodes in the area according to the size of the coordinates and distribute the sequence numbers increasing from 0, then generate random numbers within the range of the rt sequence number size and delete the sequence numbers corresponding to the random numbers in the selected node set. Node, end operation.
  • an embodiment of the present invention provides a computer-readable storage medium, where the computer-readable storage medium stores computer instructions, and the computer instructions are used to cause the computer to execute the three-dimensional stacking of the first aspect of the embodiment of the present invention Calculation method of temperature change of memory chip.
  • an embodiment of the present invention provides a computer device, including: a memory and a processor, the memory and the processor are connected in communication with each other, the memory stores computer instructions, and the processor executes the The computer instructions are used to execute the temperature change calculation method of the three-dimensional stacked memory chip according to the first aspect of the embodiment of the present invention.
  • the invention discloses a temperature change calculation method of a three-dimensional stacked memory chip, which combines heat transfer with the structure of the three-dimensional stacked memory chip, constructs a three-dimensional physical model according to the physical structure parameters of the memory chip, and uniformly measures each node in the three-dimensional physical model. Select its surrounding nodes, establish thermal conductivity differential equations based on chip performance parameters and ambient temperature, obtain the temperature information of each node at the next moment as the program simulation temperature, and regularly start temperature measurement devices such as temperature sensors/thermal cameras to obtain accurate memory chips The temperature information corrects the simulated temperature of the program. While obtaining more accurate temperature information, it also greatly reduces the working time of the temperature measurement device when the chip is running, thus reducing the power consumption of the entire storage system and the bandwidth occupation of the storage chip. , while extending the service life of the temperature measurement device.
  • FIG. 1 is a working flowchart of a specific example of a method for calculating temperature changes of three-dimensional stacked memory chips in an embodiment of the present invention
  • FIG. 2 is a schematic diagram of an overall flow of temperature field correction in an embodiment of the present invention.
  • FIG. 3 is a flowchart of selecting a node in an embodiment of the present invention.
  • FIG. 4 is a flowchart of selecting nodes from various types of nodes in an embodiment of the present invention.
  • FIG. 6 is a flowchart of an operation of adding/deleting nodes to a selected node set in an embodiment of the present invention
  • FIG. 7 is a composition diagram of a specific example of a computer device provided by an embodiment of the present invention.
  • Embodiments of the present invention provide a temperature change calculation method for three-dimensional stacked memory chips, which can be applied to large-capacity three-dimensional stacked memory chips (such as high-bandwidth memory chips, Intel Optane SSDs, etc.) in storage clusters, data servers, personal business storage, and the like.
  • the temperature change calculation of the chip as shown in Figure 1, specifically includes the following steps:
  • Step S10 Obtain physical structure parameters, chip performance parameters, and ambient temperature of the three-dimensional stacked memory chip to be calculated.
  • the physical structure parameters include the number of physical blocks, the number of physical pages in each physical block, and the specifications of the physical pages (surface area S, volume V); chip performance parameters include: chip thermal conductivity ⁇ , chip specific heat capacity c, chip Density ⁇ and chip and air convection heat transfer coefficient h; the ambient temperature of the chip is t f .
  • Step S20 constructing a three-dimensional physical model according to the physical structure parameters.
  • a three-dimensional Cartesian coordinate system is established, and a physical page is used as a basic unit to divide the memory chip into multiple nodes, and a unique three-dimensional coordinate is allocated according to the spatial position of each node, such as (m,n,j).
  • Step S30 uniformly select the surrounding nodes for each node in the three-dimensional physical model, establish a thermal conductivity differential equation based on the chip performance parameters and the ambient temperature, and solve to obtain the temperature information of each node at the next moment as the program simulation temperature; repeat during the operation of the memory chip In the above steps, the temperature measuring device is periodically started to measure the real-time temperature of the memory chip instead of the simulated temperature of the program, so as to correct the temperature field.
  • Step S01 obtain the initial temperature information of the memory chip by the temperature measuring device, and initialize the corrected counter Ct to zero;
  • Step S02 judge whether the operation of reading/writing/erasing the memory chip occurs at this time
  • Step S03 If an operation occurs, update the temperature information according to the type of operation that occurs in the chip and its target physical page node (for example, after a 32°C physical page node has a read operation, the temperature is updated to 37°C), and continue to step S04; An operation occurs, continue to step S04;
  • Step S04 for the target physical page node to calculate the temperature value at the next moment, select some nodes from its surrounding nodes through the first preset rule;
  • Step S05 according to the second preset rule, the number of selected nodes is uniform in each area divided by the target physical page node by adding nodes and deleting nodes;
  • Step S06 analyze the thermal conductivity relationship between all the selected nodes and the target node, and obtain the corresponding thermal conductivity differential equation (assuming that the target node is p, and the selected surrounding nodes are p 1 , p 2 , p 3 , ..., p M ) , the temperature of the target node at time i is expressed as Then the thermal conductivity differential equation obtained for the target node at time i should be: where distance is the function of the Euclidean distance between the two points of the input parameters), solve the thermal conductivity differential equation by substituting the chip performance parameters and the current ambient temperature, and obtain the target physical page node temperature value at the next moment as the program simulation temperature;
  • Step S07 make the counter Ct accumulate random integer values within the scope of 1 to 3, and judge whether Ct is greater than the correction threshold N (the value of N can be determined according to empirical values or actual needs, and is not limited here);
  • Step S08 if the counter Ct is greater than N, then start the temperature measurement device to measure the real-time temperature information of the memory chip (which can be a temperature sensor or a thermal camera), substitute the real-time temperature for the program simulation temperature for correcting the temperature field, and set Ct to zero , then continue to step S09; if the counter Ct is less than N, then continue to step S09;
  • Step S09 determine whether the memory chip stops working; if the memory chip stops working, end the operation; if the memory chip does not stop working, return to step S02.
  • the flow chart of node selection is shown in FIG. 3.
  • the physical page nodes around the target physical page node are classified into three groups according to the Euclidean distance between the coordinates of the target physical page node and the target physical page node.
  • Classify the nodes whose Euclidean distance between the coordinates and the target node coordinate is less than or equal to the first threshold as the first type of node for example: classify the node whose Euclidean distance between the coordinates and the target node coordinate is less than or equal to 3 as the first type of node;
  • the nodes whose Euclidean distance between the coordinates and the target node coordinates is greater than the first threshold and less than or equal to the second threshold are classified as the second type of nodes; for example: the nodes whose Euclidean distance between the coordinates and the coordinates of the target node is greater than 3 and less than or equal to 5 Classified as the second type of node;
  • the nodes whose Euclidean distance between the coordinates and the target node coordinate is greater than the second threshold are classified as the third type of nodes; for example, the nodes whose Euclidean distance between the coordinates and the target node coordinate is greater than 5 are classified as the third type of nodes.
  • A% of the nodes in the first type of nodes, B% of the nodes in the second type of nodes, and C% of the nodes in the third type of nodes where A, B, and C satisfy A>aB, B>bC , A+B+C ⁇ S, where a, b, and s are all positive integers, and a ⁇ b ⁇ s.
  • A, B, and C satisfy A>2B, B>5C, and A+B+C ⁇ 95.
  • the values of the first threshold, the second threshold, A, B, C, a, b, and s may be the most accurate value of the test result after many tests, which are only for illustration, and are not based on this. limited.
  • Step S11 For all nodes in each type of nodes in the order of coordinates (for example, compare the x coordinates first, compare the y coordinates when the x coordinates are equal, and compare the z coordinates if the x and y coordinates are equal, such as (0, 0, 0) greater than (0, 0, 1)) assign a set of serial numbers that increase from 0;
  • Step S12 Assuming that there are M nodes in total, randomly generate an integer random number in the range of 0 to M-1, and find the node whose serial number corresponds to the random number;
  • Step S13 judging whether the node is already in the selected node set
  • Step S14 if the node is already in the selected node set, return to step S12;
  • Step S15 if the node is not in the selected set, add the node and the symmetrical node centered on the target node to the selected node set;
  • Step S16 determine whether the proportion of this type of nodes in the currently selected node set to this type of nodes reaches the required percentage of selection (the first type of nodes needs to reach A% of the total number of this type of nodes, and the second type of nodes needs to reach this type of nodes. B% of the total number, the third type of nodes need to reach C% of the total number of nodes of this type);
  • Step S17 if the proportion of this type of nodes in the selected node set to this type of nodes does not meet the requirement, then return to step S12;
  • Step S18 If the proportion of the number of nodes of this type in the selected node set to the nodes of this type meets the requirement, the selection of this type of nodes is completed.
  • Figure 5 is a flow chart of the homogenization of the selected nodes in the above step S05.
  • the purpose of this step is to make the selected nodes evenly distributed, so that the temperature value of the target node at the next moment obtained by the simulation calculation is more accurate. Specific steps include:
  • Step S21 judge whether the selected node number r is equal to t in the area
  • Step S22 if the number of selected nodes in the area is equal to t, then end the operation;
  • Step S23 if the selected node in the area is not equal to t, then judge whether r is greater than t;
  • Step S24 if r is less than t, then the t-r unselected nodes with the minimum Euclidean distance from the target node in the area are added to the selected node set, and the operation is ended;
  • Step S25 if r is greater than t, then assign the selected nodes in the area according to the size of the coordinates and distribute the sequence numbers increasing from 0, then generate random numbers within the range of the rt sequence number size and delete the sequence numbers corresponding to the random numbers in the selected node set. Node, end operation.
  • the method for calculating the temperature change of a three-dimensional stacked memory chip combines heat transfer with the structure of a three-dimensional stacked memory chip, constructs a three-dimensional physical model according to the physical structure parameters of the memory chip, and uniformly measures each node in the three-dimensional physical model. Select its surrounding nodes, establish thermal conductivity differential equations based on chip performance parameters and ambient temperature, obtain the temperature information of each node at the next moment as the program simulation temperature, and regularly start temperature measurement devices such as temperature sensors/thermal cameras to obtain accurate memory chips The temperature information corrects the simulated temperature of the program. While obtaining more accurate temperature information, it also greatly reduces the working time of the temperature measurement device when the chip is running, thus reducing the power consumption of the entire storage system and the bandwidth occupation of the storage chip. , while extending the service life of the temperature measurement device.
  • FIG. 7 An embodiment of the present invention provides a computer device. As shown in FIG. 7 , the device may include a processor 51 and a memory 52, where the processor 51 and the memory 52 may be connected through a bus or in other ways. FIG. 7 takes the connection through a bus as an example .
  • the processor 51 may be a central processing unit (Central Processing Unit, CPU).
  • the processor 51 can also be other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuit (Application Specific Integrated Circuit, ASIC), Field-Programmable Gate Array (Field-Programmable Gate Array, FPGA) or Other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components and other chips, or a combination of the above types of chips.
  • DSP Digital Signal Processor
  • ASIC Application Specific Integrated Circuit
  • FPGA Field-Programmable Gate Array
  • Other programmable logic devices discrete gate or transistor logic devices, discrete hardware components and other chips, or a combination of the above types of chips.
  • the memory 52 can be used to store non-transitory software programs, non-transitory computer-executable programs and modules, such as corresponding program instructions/modules in the embodiments of the present invention.
  • the processor 51 executes various functional applications and data processing of the processor by running the non-transitory software programs, instructions and modules stored in the memory 52, that is, to realize the temperature change of the three-dimensional stacked memory chips in the above method embodiment 1. calculation method.
  • the memory 52 may include a storage program area and a storage data area, wherein the storage program area may store an operating system and an application program required by at least one function; the storage data area may store data created by the processor 51 and the like. Additionally, memory 52 may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, memory 52 may optionally include memory located remotely from processor 51, which may be connected to processor 51 via a network. Examples of such networks include, but are not limited to, the Internet, intranets, intranets, mobile communication networks, and combinations thereof.
  • One or more modules are stored in the memory 52, and when executed by the processor 51, execute the temperature change calculation method of the three-dimensional stacked memory chips in Embodiment 1.
  • the storage medium may be a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM), a flash memory (Flash Memory), a hard disk (Hard Disk Drive) , abbreviation: HDD) or solid-state drive (Solid-State Drive, SSD), etc.; the storage medium may also include a combination of the above-mentioned types of memories.

Landscapes

  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

本发明公开了一种三维堆叠存储芯片的温度变化计算方法，将传热学与三维堆叠存储芯片的结构结合，根据存储芯片的物理结构参数构建三维物理模型，对三维物理模型中每个节点均匀选取其周围节点，基于存储芯片性能参数和环境温度建立导热微分方程，求解得到各节点下一时刻温度信息作为程序模拟温度，并定期启动温度传感器/热感相机等温度测量装置，获得精确的存储芯片温度信息对程序模拟温度进行校正，在得到较为精确温度信息的同时，也大大减少了温度测量装置在芯片运行时的工作时间，减少了整个存储系统的功耗、减少了存储芯片的带宽占用，同时延长了温度测量装置的使用寿命。

Description

一种三维堆叠存储芯片的温度变化计算方法 技术领域
本发明涉及温度测量技术领域，具体涉及一种三维堆叠存储芯片的温度变化计算方法。
背景技术
随着云服务、大数据、人工智能等新技术的发展以及应用，市场对存储介质的存储可靠性、读写性能以及容量等指标有了更高的需求。随着硅通孔等技术的发展，存储芯片从二维架构跨向了三维架构。与传统存储技术不同，三维堆叠架构的存储芯片有着更高的存储密度，更大的存储容量，通过增大并行宽度或者利用串行传输提升存储带宽，不同程度简化系统存储控制设计难度，具有高集成度、高带宽、高能效等性能优势。因此大容量三维堆叠存储芯片在存储市场中占据了越来越重要的地位，然而三维堆叠存储芯片对于温度十分敏感，高温会加速电荷的放射和移动最终使得数据误码率大幅升高、存储芯片可靠性下降。
为减少高温对于三维堆叠存储芯片造成的不良影响，现有技术提出了多种基于温度信息的三维堆叠存储芯片稳定性优化方案。这些优化方案都是通过温度传感器或热感相机来监测三维堆叠存储芯片的温度信息。这种获取温度信息方法有以下几点不足：
1、温度传感器或热感相机必须在存储芯片工作时一直保持运行状态并收集温度信息，因此产生更多的功耗；
2、运用温度传感器或热感相机获取到的芯片温度信息需要持续不断地传输到存储芯片控制器中，占用了大量存储芯片带宽；
3、运用温度传感器或热感相机获取温度信息极度依赖于硬件的可靠性，而温度传感器或热感相机寿命相较存储芯片更短，在长期运行状态下容易损坏导致提供错误温度信息。
发明内容
因此，本发明要解决的技术问题在于克服现有技术对三维堆叠的存储芯片测量温度时功耗高、占用大量存储芯片带宽，温度测量精度低的缺陷，从而提供一种三维堆叠存储芯片的温度变化计算方法。
为达到上述目的，本发明提供如下技术方案：
第一方面，本发明实施例提供三维堆叠存储芯片的温度变化计算方法，包括如下步 骤：
获取待计算三维堆叠存储芯片的物理结构参数、芯片性能参数、环境温度；
根据所述物理结构参数构建三维物理模型；
对三维物理模型中每个节点均匀选取其周围节点，并基于芯片性能参数和环境温度建立导热微分方程，求解得到各节点下一时刻温度信息作为程序模拟温度；在存储芯片运行期间重复上述步骤，并周期性启动温度测量装置测量存储芯片的实时温度替代程序模拟温度，用于校正温度场。
在一实施例中，所述物理结构参数包含物理块个数、各物理块中物理页个数、物理页的规格；
所述芯片性能参数包括：芯片导热系数、芯片比热容、芯片密度以及芯片与空气对流换热系数。
在一实施例中，所述根据所述物理结构参数构建三维物理模型，包括：
根据三维堆叠存储芯片的物理结构，建立三维笛卡尔坐标系并以物理页为基本单元，将存储芯片划分成多个节点，根据各节点的空间位置分配唯一的三维坐标。
在一实施例中，所述对三维物理模型中每个节点均匀选取其周围节点，并基于芯片性能参数和环境温度建立导热微分方程，求解得到各节点下一时刻温度信息作为程序模拟温度；在存储芯片运行期间重复上述步骤，并周期性启动温度测量装置测量存储芯片的实时温度替代程序模拟温度，用于校正温度场，包括：
步骤S01:通过温度测量装置获得存储芯片的初始温度信息，并将校正的计数器Ct初始化为零；
步骤S02:判断此时是否发生对存储芯片读/写/擦除的操作；
步骤S03:若发生操作，则根据芯片发生的操作类型以及其目标物理页节点更新温度信息，继续步骤S04；若不发生操作，继续步骤S04；
步骤S04:对要计算下一时刻温度值的目标物理页节点，通过第一预设规则从其周围节点中选取部分节点；
步骤S05:对已选取节点按照第二预设规则通过增加节点和删除节点的方式使得以目标物理页节点为中心划分的各区域内已选取节点数量均匀；
步骤S06:对所有已选取节点和目标节点的导热关系作分析，并求得对应的导热微分方程，通过代入芯片性能参数和当前环境温度对导热微分方程求解，得到下一时刻目标物理页节点温度值作为程序模拟温度；
步骤S07:使计数器Ct累加1到3范围内的随机整数值，并判断Ct是否大于校正阈值N；
步骤S08:若计数器Ct大于N，则启动温度测量装置测量存储芯片的实时温度信息，将实时温度替代程序模拟温度用于校正温度场，并将Ct置零，则继续步骤S09；若计数器Ct小于N，则继续步骤S09；
步骤S09:判断存储芯片是否停止工作；若存储芯片停止工作，则结束运行；若存储芯片没有停止工作，则返回步骤S02。
在一实施例中，对目标物理页节点周围的物理页节点根据其与目标物理页节点的坐标欧氏距离分类成三类节点：
将坐标与目标节点坐标欧氏距离小于等于第一阈值的节点归为第一类节点；
将坐标与目标节点坐标欧氏距离大于第一阈值且小于等于第二阈值的节点归为第二类节点；
将坐标与目标节点坐标欧氏距离大于第二阈值的节点归为第三类节点；
在第一类节点中选取A％的节点、在第二类节点中选取B％的节点，在第三类节点中选取C％的节点，其中A、B、C满足A>aB，B>bC，A+B+C<S，其中，a、b、s均为正整数，且a<b<s。
在一实施例中，通过第一预设规则从其周围节点中选取部分节点的过程，包括：
步骤S11：对每类节点内的所有节点按照坐标顺序分配从0开始递增的一组序号；
步骤S12：假设共有M个节点，随机生成一个0至M-1范围内的整数随机数，并找到序号与该随机数对应的节点；
步骤S13：判断该节点是否已经在已选取节点集合中；
步骤S14：如果该节点已经在已选取节点集合中，则返回步骤S12；
步骤S15：如果该节点不在已选取集合中，将该节点和该节点以目标节点为中心的对称节点添加到已选取节点集合中；
步骤S16：判断当前已选取节点集合中该类节点数量占该类节点比例是否达到所要求选取的百分比；
步骤S17：若已选取节点集合中该类节点数量占该类节点比例未达到要求，则返回步骤S12；
步骤S18：若已选取节点集合中该类节点数量占该类节点比例达到要求，则该类节点选取完成。
在一实施例中，对已选取节点按照第二预设规则通过增加节点和删除节点的方式使 得以目标物理页节点为中心划分的各区域内已选取节点数量均匀的过程，包括：
以目标节点为中心，对所有节点沿着x、y、z轴三个方向进行切割，将其他物理页节点划分成8个区域，统计已选取节点总数记为T，令t＝T/8，对各个区域内的节点根据其已选取节点数量，对已选取节点集合进行增加/删除节点操作，使得所有区域内的已选取节点集合种节点数量都为t。
在一实施例中，所述对已选取节点集合进行增加/删除节点操作，使得所有区域内的已选取节点集合种节点数量都为t的过程，包括：
步骤S21:判断区域内已选取节点数量r是否等于t；
步骤S22:若区域内已选取节点数量等于t，则结束运行；
步骤S23:若区域内已选取节点不等于t，则判断r是否大于t；
步骤S24:若r小于t，则将区域中与目标节点欧式距离最小的t-r个未被选取节点加入到已选取节点集合中，结束运行；
步骤S25:若r大于t，则将区域中已选取节点按照坐标大小分配从0开始递增的序号，随后生成r-t序号大小范围内的随机数并在已选取节点集合中删除序号与随机数对应的节点，结束运行。
第二方面，本发明实施例提供一种计算机可读存储介质，所述计算机可读存储介质存储有计算机指令，所述计算机指令用于使所述计算机执行本发明实施例第一方面的三维堆叠存储芯片的温度变化计算方法。
第三方面，本发明实施例提供一种计算机设备，包括：存储器和处理器，所述存储器和所述处理器之间互相通信连接，所述存储器存储有计算机指令，所述处理器通过执行所述计算机指令，从而执行本发明实施例第一方面的三维堆叠存储芯片的温度变化计算方法。
本发明技术方案，具有如下优点：
本发明公开了一种三维堆叠存储芯片的温度变化计算方法，将传热学与三维堆叠存储芯片的结构结合，根据存储芯片的物理结构参数构建三维物理模型，对三维物理模型中每个节点均匀选取其周围节点，基于芯片性能参数和环境温度建立导热微分方程，求解得到各节点下一时刻温度信息作为程序模拟温度，并定期启动温度传感器/热感相机等温度测量装置，获得精确的存储芯片温度信息对程序模拟温度进行校正，在得到较为精确温度信息的同时，也大大减少了温度测量装置在芯片运行时的工作时间，因此减少了整个存储系统的功耗、减少了存储芯片的带宽占用，同时延长了温度测量装置的使用寿 命。
附图说明
为了更清楚地说明本发明具体实施方式或现有技术中的技术方案，下面将对具体实施方式或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图是本发明的一些实施方式，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1为本发明实施例中三维堆叠存储芯片的温度变化计算方法的一个具体示例的工作流程图；
图2为本发明实施例中对温度场校正的整体流程示意图；
图3为本发明实施例中选取节点的流程图；
图4为本发明实施例中在各类节点选取节点的流程图；
图5为本发明实施例中对已选取节点均匀化流程图；
图6为本发明实施例中对已选取节点集合进行增加/删除节点操作的流程图；
图7为本发明实施例提供的计算机设备一个具体示例的组成图。
具体实施方式
下面将结合附图对本发明的技术方案进行清楚、完整地描述，显然，所描述的实施例是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
此外，下面所描述的本发明不同实施方式中所涉及的技术特征只要彼此之间未构成冲突就可以相互结合。
实施例1
本发明实施例提供一种三维堆叠存储芯片的温度变化计算方法，可以应用于存储集群、数据服务器、个人商务存储等方面的大容量三维堆叠存储芯片(如高带宽存储芯片、Intel傲腾SSD等芯片)的温度变化计算，如图1所示，具体包括如下步骤：
步骤S10：获取待计算三维堆叠存储芯片的物理结构参数、芯片性能参数、环境温度。
本发明实施例中物理结构参数包含物理块个数、各物理块中物理页个数、物理页的规格(表面积S、体积V)；芯片性能参数包括：芯片导热系数λ、芯片比热容c、芯片密度ρ以及芯片与空气对流换热系数h；芯片所处环境温度为t f。
步骤S20：根据所述物理结构参数构建三维物理模型。
本发明实施例中，根据三维堆叠存储芯片的物理结构，建立三维笛卡尔坐标系并以物理页为基本单元，将存储芯片划分成多个节点，根据各节点的空间位置分配唯一的三维坐标如(m,n,j)。
步骤S30：对三维物理模型中每个节点均匀选取其周围节点，并基于芯片性能参数和环境温度建立导热微分方程，求解得到各节点下一时刻温度信息作为程序模拟温度；在存储芯片运行期间重复上述步骤，并周期性启动温度测量装置测量存储芯片的实时温度替代程序模拟温度，用于校正温度场。
如图2所示的为对存储芯片进行温度场校正的整个流程，具体步骤如下：
步骤S01:通过温度测量装置获得存储芯片的初始温度信息，并将校正的计数器Ct初始化为零；
步骤S02:判断此时是否发生对存储芯片读/写/擦除的操作；
步骤S03:若发生操作，则根据芯片发生的操作类型以及其目标物理页节点更新温度信息(如一个32℃的物理页节点发生读操作后，温度更新为37℃)，继续步骤S04；若不发生操作，继续步骤S04；
步骤S04:对要计算下一时刻温度值的目标物理页节点，通过第一预设规则从其周围节点中选取部分节点；
步骤S05:对已选取节点按照第二预设规则通过增加节点和删除节点的方式使得以目标物理页节点为中心划分的各区域内已选取节点数量均匀；
步骤S06:对所有已选取节点和目标节点的导热关系作分析，并求得对应的导热微分方程(假设目标节点为p，选取的周围节点为p 1、p 2、p 3、…、p M，i时刻目标节点的温度表示为
Figure PCTCN2020112484-appb-000001
则i时刻对目标节点求得导热微分方程应为
Figure PCTCN2020112484-appb-000002
Figure PCTCN2020112484-appb-000003
其中distance是计算传入参数的两个点之间欧式距离的函数)，通过代入芯片性能参数和当前环境温度对导热微分方程求解，得到下一时刻目标物理页节点温度值作为程序模拟温度；
步骤S07:使计数器Ct累加1到3范围内的随机整数值，并判断Ct是否大于校正阈值N(N的取值可以根据经验值或者实际需求确定，在此不做限制)；
步骤S08:若计数器Ct大于N，则启动温度测量装置测量(可以为温度传感器或热感相机)存储芯片的实时温度信息，将实时温度替代程序模拟温度用于校正温度场，并将 Ct置零，则继续步骤S09；若计数器Ct小于N，则继续步骤S09；
步骤S09:判断存储芯片是否停止工作；若存储芯片停止工作，则结束运行；若存储芯片没有停止工作，则返回步骤S02。
本发明实施例中，对于上述步骤S04中，选取节点的流程图如图3所示，首先，对目标物理页节点周围的物理页节点根据其与目标物理页节点的坐标欧氏距离分类成三类节点：
1、将坐标与目标节点坐标欧氏距离小于等于第一阈值的节点归为第一类节点，例如是：将坐标与目标节点坐标欧氏距离小于等于3的节点归为第一类节点；
2、将坐标与目标节点坐标欧氏距离大于第一阈值且小于等于第二阈值的节点归为第二类节点；例如是：将坐标与目标节点坐标欧氏距离大于3且小于等于5的节点归为第二类节点；
3、将坐标与目标节点坐标欧氏距离大于第二阈值的节点归为第三类节点；例如是：将坐标与目标节点坐标欧氏距离大于5的节点归为第三类节点。
在第一类节点中选取A％的节点、在第二类节点中选取B％的节点，在第三类节点中选取C％的节点，其中A、B、C满足A>aB，B>bC，A+B+C<S，其中，a、b、s均为正整数，且a<b<s。在一具体实施例汇中，A、B、C满足A>2B，B>5C，A+B+C<95。
需要说明的是，第一阈值、第二阈值、A、B、C、a、b、s的取值可以是经过多次试验，取试验结果最准确的值，仅为举例说明，不以此为限。
本发明实施例中，在各类节点选取节点的流程如图4所示，具体步骤包括：
步骤S11：对每类节点内的所有节点按照坐标顺序(例如是：先比较x坐标，当x坐标相等时比较y坐标，若x与y坐标都相等则比较z坐标，如(0，0，0)大于(0，0，1))分配从0开始递增的一组序号；
步骤S12：假设共有M个节点，随机生成一个0至M-1范围内的整数随机数，并找到序号与该随机数对应的节点；
步骤S13：判断该节点是否已经在已选取节点集合中；
步骤S14：如果该节点已经在已选取节点集合中，则返回步骤S12；
步骤S15：如果该节点不在已选取集合中，将该节点和该节点以目标节点为中心的对称节点添加到已选取节点集合中；
步骤S16：判断当前已选取节点集合中该类节点数量占该类节点比例是否达到所要求选取的百分比(第一类节点需达到该类节点总数的A％，第二类节点需达到该类节点 总数的B％，第三类节点需达到该类节点总数的C％)；
步骤S17：若已选取节点集合中该类节点数量占该类节点比例未达到要求，则返回步骤S12；
步骤S18：若已选取节点集合中该类节点数量占该类节点比例达到要求，则该类节点选取完成。
如图5所示的为上述步骤S05中对已选取节点均匀化流程图，该步骤的目的是使得选取的节点分布均匀，从而使得模拟计算得到的下一时刻目标节点温度值更精确。具体步骤包括：
以目标节点为中心，对所有节点沿着x、y、z轴三个方向进行切割，将其他物理页节点划分成8个区域，统计已选取节点总数记为T，令t＝T/8，t为每个区域通过增删节点后应包含的被选取物理页节点数量，对各个区域内的节点根据其已选取节点数量，对已选取节点集合进行增加/删除节点操作，使得所有区域内的已选取节点集合种节点数量都为t，对已选取节点集合进行增加/删除节点操作的具体流程如图6所示，包括：
步骤S21:判断区域内已选取节点数量r是否等于t；
步骤S22:若区域内已选取节点数量等于t，则结束运行；
步骤S23:若区域内已选取节点不等于t，则判断r是否大于t；
步骤S24:若r小于t，则将区域中与目标节点欧式距离最小的t-r个未被选取节点加入到已选取节点集合中，结束运行；
步骤S25:若r大于t，则将区域中已选取节点按照坐标大小分配从0开始递增的序号，随后生成r-t序号大小范围内的随机数并在已选取节点集合中删除序号与随机数对应的节点，结束运行。
本发明实施例提供的三维堆叠存储芯片的温度变化计算方法，将传热学与三维堆叠存储芯片的结构结合，根据存储芯片的物理结构参数构建三维物理模型，对三维物理模型中每个节点均匀选取其周围节点，基于芯片性能参数和环境温度建立导热微分方程，求解得到各节点下一时刻温度信息作为程序模拟温度，并定期启动温度传感器/热感相机等温度测量装置，获得精确的存储芯片温度信息对程序模拟温度进行校正，在得到较为精确温度信息的同时，也大大减少了温度测量装置在芯片运行时的工作时间，因此减少了整个存储系统的功耗、减少了存储芯片的带宽占用，同时延长了温度测量装置的使用寿命。
实施例2
本发明实施例提供一种计算机设备，如图7所示，该设备可以包括处理器51和存储器52，其中处理器51和存储器52可以通过总线或者其他方式连接，图7以通过总线连接为例。
处理器51可以为中央处理器(Central Processing Unit，CPU)。处理器51还可以为其他通用处理器、数字信号处理器(Digital Signal Processor，DSP)、专用集成电路(Application Specific Integrated Circuit，ASIC)、现场可编程门阵列(Field-Programmable Gate Array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等芯片，或者上述各类芯片的组合。
存储器52作为一种非暂态计算机可读存储介质，可用于存储非暂态软件程序、非暂态计算机可执行程序以及模块，如本发明实施例中的对应的程序指令/模块。处理器51通过运行存储在存储器52中的非暂态软件程序、指令以及模块，从而执行处理器的各种功能应用以及数据处理，即实现上述方法实施例1中的三维堆叠存储芯片的温度变化计算方法。
存储器52可以包括存储程序区和存储数据区，其中，存储程序区可存储操作系统、至少一个功能所需要的应用程序；存储数据区可存储处理器51所创建的数据等。此外，存储器52可以包括高速随机存取存储器，还可以包括非暂态存储器，例如至少一个磁盘存储器件、闪存器件、或其他非暂态固态存储器件。在一些实施例中，存储器52可选包括相对于处理器51远程设置的存储器，这些远程存储器可以通过网络连接至处理器51。上述网络的实例包括但不限于互联网、企业内部网、企业内网、移动通信网及其组合。
一个或者多个模块存储在存储器52中，当被处理器51执行时，执行实施例1中的三维堆叠存储芯片的温度变化计算方法。
上述计算机设备具体细节可以对应参阅实施例1中对应的相关描述和效果进行理解，此处不再赘述。
本领域技术人员可以理解，实现上述实施例方法中的全部或部分流程，是可以通过计算机程序来指令相关的硬件来完成的程序可存储于一计算机可读取存储介质中，该程序在执行时，可包括如上述各方法的实施例的流程。其中，存储介质可为磁碟、光盘、只读存储记忆体(Read-Only Memory，ROM)、随机存储记忆体(Random Access Memory，RAM)、快闪存储器(Flash Memory)、硬盘(Hard Disk Drive，缩写：HDD)或固态硬盘(Solid-State Drive，SSD)等；存储介质还可以包括上述种类的存储器的组合。
显然，上述实施例仅仅是为清楚地说明所作的举例，而并非对实施方式的限定。对于所属领域的普通技术人员来说，在上述说明的基础上还可以做出其它不同形式的变化或变动。这里无需也无法对所有的实施方式予以穷举。而由此所引申出的显而易见的变化或变动仍处于本发明创造的保护范围之中。

Claims (10)

  1. 一种三维堆叠存储芯片的温度变化计算方法，其特征在于，包括如下步骤：
    获取待计算三维堆叠存储芯片的物理结构参数、芯片性能参数、环境温度；
    根据所述物理结构参数构建三维物理模型；
    对三维物理模型中每个节点均匀选取其周围节点，并基于芯片性能参数和环境温度建立导热微分方程，求解得到各节点下一时刻温度信息作为程序模拟温度；在存储芯片运行期间重复上述步骤，并周期性启动温度测量装置测量存储芯片的实时温度替代程序模拟温度，用于校正温度场。
  2. 根据权利要求1所述的三维堆叠存储芯片的温度变化计算方法，其特征在于，
    所述物理结构参数包含物理块个数、各物理块中物理页个数、物理页的规格；
    所述芯片性能参数包括：芯片导热系数、芯片比热容、芯片密度以及芯片与空气对流换热系数。
  3. 根据权利要求2所述的三维堆叠存储芯片的温度变化计算方法，其特征在于，所述根据所述物理结构参数构建三维物理模型，包括：
    根据三维堆叠存储芯片的物理结构，建立三维笛卡尔坐标系并以物理页为基本单元，将存储芯片划分成多个节点，根据各节点的空间位置分配唯一的三维坐标。
  4. 根据权利要求3所述的三维堆叠存储芯片的温度变化计算方法，其特征在于，所述对三维物理模型中每个节点均匀选取其周围节点，并基于芯片性能参数和环境温度建立导热微分方程，求解得到各节点下一时刻温度信息作为程序模拟温度；在存储芯片运行期间重复上述步骤，并周期性启动温度测量装置测量存储芯片的实时温度替代程序模拟温度，用于校正温度场，包括：
    步骤S01:通过温度测量装置获得存储芯片的初始温度信息，并将校正的计数器Ct初始化为零；
    步骤S02:判断此时是否发生对存储芯片读/写/擦除的操作；
    步骤S03:若发生操作，则根据芯片发生的操作类型以及其目标物理页节点更新温度信息，继续步骤S04；若不发生操作，继续步骤S04；
    步骤S04:对要计算下一时刻温度值的目标物理页节点，通过第一预设规则从其周围节点中选取部分节点；
    步骤S05:对已选取节点按照第二预设规则通过增加节点和删除节点的方式使得以目标物理页节点为中心划分的各区域内已选取节点数量均匀；
    步骤S06:对所有已选取节点和目标节点的导热关系作分析，并求得对应的导热微分方程，通过代入芯片性能参数和当前环境温度对导热微分方程求解，得到下一时刻目标物理页节点温度值作为程序模拟温度；
    步骤S07:使计数器Ct累加1到3范围内的随机整数值，并判断Ct是否大于校正阈值N；
    步骤S08:若计数器Ct大于N，则启动温度测量装置测量存储芯片的实时温度信息，将实时温度替代程序模拟温度用于校正温度场，并将Ct置零，则继续步骤S09；若计数器Ct小于N，则继续步骤S09；
    步骤S09:判断存储芯片是否停止工作；若存储芯片停止工作，则结束运行；若存储芯片没有停止工作，则返回步骤S02。
  5. 根据权利要求4所述的三维堆叠存储芯片的温度变化计算方法，其特征在于，对目标物理页节点周围的物理页节点根据其与目标物理页节点的坐标欧氏距离分类成三类节点：
    将坐标与目标节点坐标欧氏距离小于等于第一阈值的节点归为第一类节点；
    将坐标与目标节点坐标欧氏距离大于第一阈值且小于等于第二阈值的节点归为第二类节点；
    将坐标与目标节点坐标欧氏距离大于第二阈值的节点归为第三类节点；
    在第一类节点中选取A％的节点、在第二类节点中选取B％的节点，在第三类节点中选取C％的节点，其中A、B、C满足A>aB，B>bC，A+B+C<S，其中，a、b、s均为正整数，且a<b<s。
  6. 根据权利要求5所述的三维堆叠存储芯片的温度变化计算方法，其特征在于，通过第一预设规则从其周围节点中选取部分节点的过程，包括：
    步骤S11：对每类节点内的所有节点按照坐标顺序分配从0开始递增的一组序号；
    步骤S12：假设共有M个节点，随机生成一个0至M-1范围内的整数随机数，并找到序号与该随机数对应的节点；
    步骤S13：判断该节点是否已经在已选取节点集合中；
    步骤S14：如果该节点已经在已选取节点集合中，则返回步骤S12；
    步骤S15：如果该节点不在已选取集合中，将该节点和该节点以目标节点为中心的对称节点添加到已选取节点集合中；
    步骤S16：判断当前已选取节点集合中该类节点数量占该类节点比例是否达到所要求选取的百分比；
    步骤S17：若已选取节点集合中该类节点数量占该类节点比例未达到要求，则返回步骤S12；
    步骤S18：若已选取节点集合中该类节点数量占该类节点比例达到要求，则该类节点选取完成。
  7. 根据权利要求4所述的三维堆叠存储芯片的温度变化计算方法，其特征在于，对已选取节点按照第二预设规则通过增加节点和删除节点的方式使得以目标物理页节点为中心划分的各区域内已选取节点数量均匀的过程，包括：
    以目标节点为中心，对所有节点沿着x、y、z轴三个方向进行切割，将其他物理页节点划分成8个区域，统计已选取节点总数记为T，令t＝T/8，对各个区域内的节点根据其已选取节点数量，对已选取节点集合进行增加/删除节点操作，使得所有区域内的已选取节点集合种节点数量都为t。
  8. 根据权利要求7所述的三维堆叠存储芯片的温度变化计算方法，其特征在于，所述对已选取节点集合进行增加/删除节点操作，使得所有区域内的已选取节点集合种节点数量都为t的过程，包括：
    步骤S21:判断区域内已选取节点数量r是否等于t；
    步骤S22:若区域内已选取节点数量等于t，则结束运行；
    步骤S23:若区域内已选取节点不等于t，则判断r是否大于t；
    步骤S24:若r小于t，则将区域中与目标节点欧式距离最小的t-r个未被选取节点加入到已选取节点集合中，结束运行；
    步骤S25:若r大于t，则将区域中已选取节点按照坐标大小分配从0开始递增的序号，随后生成r-t序号大小范围内的随机数并在已选取节点集合中删除序号与随机数对应的节点，结束运行。
  9. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介质存储有计算机指令，所述计算机指令用于使所述计算机执行如权利要求1-8任一项所述的三维堆叠存储芯片的温度变化计算方法。
  10. 一种计算机设备，其特征在于，包括：存储器和处理器，所述存储器和所述处理器之间互相通信连接，所述存储器存储有计算机指令，所述处理器通过执行所述计算机指令，从而执行如权利要求1-8任一项所述的三维堆叠存储芯片的温度变化计算方法。
PCT/CN2020/112484 2020-08-31 2020-08-31 一种三维堆叠存储芯片的温度变化计算方法 WO2022041198A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2020/112484 WO2022041198A1 (zh) 2020-08-31 2020-08-31 一种三维堆叠存储芯片的温度变化计算方法

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2020/112484 WO2022041198A1 (zh) 2020-08-31 2020-08-31 一种三维堆叠存储芯片的温度变化计算方法

Publications (1)

Publication Number Publication Date
WO2022041198A1 true WO2022041198A1 (zh) 2022-03-03

Family

ID=80354312

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/112484 WO2022041198A1 (zh) 2020-08-31 2020-08-31 一种三维堆叠存储芯片的温度变化计算方法

Country Status (1)

Country Link
WO (1) WO2022041198A1 (zh)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114896920A (zh) * 2022-07-12 2022-08-12 中南民族大学 一种三维堆叠芯片热仿真模型建立及热点温度预测方法
CN116026512A (zh) * 2023-03-30 2023-04-28 江西飞尚科技有限公司 索力计算温度拟合方法、系统、计算机及可读存储介质
CN116304956A (zh) * 2023-05-15 2023-06-23 济宁市质量计量检验检测研究院(济宁半导体及显示产品质量监督检验中心、济宁市纤维质量监测中心) 一种芯片温度异常在线检测方法
CN116544153A (zh) * 2023-07-06 2023-08-04 成都汉芯国科集成技术有限公司 一种具有散热互联功能的3d封装系统及其封装方法
CN117074925A (zh) * 2023-10-16 2023-11-17 中诚华隆计算机技术有限公司 一种3d芯片测试分析方法和系统
CN117548928A (zh) * 2024-01-12 2024-02-13 杭州峰景科技有限公司 一种焊机物联设备的芯片调度方法及装置

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103714189A (zh) * 2012-09-28 2014-04-09 台湾积体电路制造股份有限公司 用于堆叠ic设计中的跨芯片热和功率管理的系统和方法
CN105784138A (zh) * 2016-05-09 2016-07-20 南京理工大学 一种非接触式在线温度测量方法及系统
CN106919723A (zh) * 2015-12-24 2017-07-04 中国科学院微电子研究所 三维集成阻变存储器的热效应评估及降低热串扰的方法
US20170372916A1 (en) * 2016-06-23 2017-12-28 Tokyo Electron Limited Etching process method
US20200212106A1 (en) * 2018-08-22 2020-07-02 International Business Machines Corporation Multilayer back end of line (beol)-stackable cross-point memory array with complementary pass transistor selectors
CN111415003A (zh) * 2020-02-20 2020-07-14 清华大学 面向神经网络加速芯片的三维堆叠存储优化方法及装置

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103714189A (zh) * 2012-09-28 2014-04-09 台湾积体电路制造股份有限公司 用于堆叠ic设计中的跨芯片热和功率管理的系统和方法
CN106919723A (zh) * 2015-12-24 2017-07-04 中国科学院微电子研究所 三维集成阻变存储器的热效应评估及降低热串扰的方法
CN105784138A (zh) * 2016-05-09 2016-07-20 南京理工大学 一种非接触式在线温度测量方法及系统
US20170372916A1 (en) * 2016-06-23 2017-12-28 Tokyo Electron Limited Etching process method
US20200212106A1 (en) * 2018-08-22 2020-07-02 International Business Machines Corporation Multilayer back end of line (beol)-stackable cross-point memory array with complementary pass transistor selectors
CN111415003A (zh) * 2020-02-20 2020-07-14 清华大学 面向神经网络加速芯片的三维堆叠存储优化方法及装置

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114896920A (zh) * 2022-07-12 2022-08-12 中南民族大学 一种三维堆叠芯片热仿真模型建立及热点温度预测方法
CN114896920B (zh) * 2022-07-12 2022-09-30 中南民族大学 一种三维堆叠芯片热仿真模型建立及热点温度预测方法
CN116026512A (zh) * 2023-03-30 2023-04-28 江西飞尚科技有限公司 索力计算温度拟合方法、系统、计算机及可读存储介质
CN116026512B (zh) * 2023-03-30 2023-07-04 江西飞尚科技有限公司 索力计算温度拟合方法、系统、计算机及可读存储介质
CN116304956A (zh) * 2023-05-15 2023-06-23 济宁市质量计量检验检测研究院(济宁半导体及显示产品质量监督检验中心、济宁市纤维质量监测中心) 一种芯片温度异常在线检测方法
CN116304956B (zh) * 2023-05-15 2023-08-15 济宁市质量计量检验检测研究院(济宁半导体及显示产品质量监督检验中心、济宁市纤维质量监测中心) 一种芯片温度异常在线检测方法
CN116544153A (zh) * 2023-07-06 2023-08-04 成都汉芯国科集成技术有限公司 一种具有散热互联功能的3d封装系统及其封装方法
CN116544153B (zh) * 2023-07-06 2023-09-01 成都汉芯国科集成技术有限公司 一种具有散热互联功能的3d封装系统及其封装方法
CN117074925A (zh) * 2023-10-16 2023-11-17 中诚华隆计算机技术有限公司 一种3d芯片测试分析方法和系统
CN117074925B (zh) * 2023-10-16 2023-12-29 中诚华隆计算机技术有限公司 一种3d芯片测试分析方法和系统
CN117548928A (zh) * 2024-01-12 2024-02-13 杭州峰景科技有限公司 一种焊机物联设备的芯片调度方法及装置
CN117548928B (zh) * 2024-01-12 2024-04-12 杭州峰景科技有限公司 一种焊机物联设备的芯片调度方法及装置

Similar Documents

Publication Publication Date Title
WO2022041198A1 (zh) 一种三维堆叠存储芯片的温度变化计算方法
US10031671B2 (en) Method, apparatus, and system for calculating identification threshold to distinguish cold data and hot data
TWI712900B (zh) 一種分布式集群訓練方法和裝置
WO2018149345A1 (zh) 一种数据处理方法及装置
CN108874803B (zh) 数据存储方法、装置及存储介质
CN109299190B (zh) 分布式存储系统中处理对象的元数据的方法及装置
US20210158353A1 (en) Methods, systems, apparatuses, and devices for processing request in consortium blockchain
US9454441B2 (en) Data layout for recovery and durability
WO2022227373A1 (zh) 一种硬盘健康评估方法和存储设备
CN110875938A (zh) 一种区块链中信息的发送方法及装置
CN102821122A (zh) 节点分布的方法与装置以及计算机程序产品
CN113054651A (zh) 一种网络拓扑优化方法、装置以及系统
TWI758223B (zh) 具有動態最小批次尺寸之運算方法，以及用於執行該方法之運算系統及電腦可讀儲存媒體
CN112181274B (zh) 提升存储设备性能稳定性的大块的组织方法及其存储设备
CN112037845B (zh) 一种三维堆叠存储芯片的温度变化计算方法
US20140289307A1 (en) Method for transmitting data between electronic devices
CN115617279B (zh) 分布式云数据的处理方法、装置及存储介质
CN108536931B (zh) 基于贪婪算法的三维片上网络中温度传感器的部署方法
CN113779742B (zh) 热阻网络简化模型建模方法、装置及芯片结温预测方法
CN109547236A (zh) 电子数据上传存储方法、可读存储介质和终端
EP3846037B1 (en) Storage device configured to support multi-streams and operation method thereof
EP4200768A1 (en) Handling training of a machine learning model
CN115361284B (zh) 一种基于sdn的虚拟网络功能的部署调整方法
CN108156011B (zh) 一种进行无线接入点聚类的方法与设备
TWI822162B (zh) 資訊處理裝置以及控制資訊處理裝置的方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20950867

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 22.06.2023)

122 Ep: pct application non-entry in european phase

Ref document number: 20950867

Country of ref document: EP

Kind code of ref document: A1