CN113782086B - Method and device for grading evaluation of PMEM memory - Google Patents
Method and device for grading evaluation of PMEM memory Download PDFInfo
- Publication number
- CN113782086B CN113782086B CN202110959424.0A CN202110959424A CN113782086B CN 113782086 B CN113782086 B CN 113782086B CN 202110959424 A CN202110959424 A CN 202110959424A CN 113782086 B CN113782086 B CN 113782086B
- Authority
- CN
- China
- Prior art keywords
- pmem
- memory
- test
- mode
- pmem memory
- 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.)
- Active
Links
- 230000015654 memory Effects 0.000 title claims abstract description 297
- 238000011156 evaluation Methods 0.000 title claims abstract description 104
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000012360 testing method Methods 0.000 claims abstract description 271
- 238000005315 distribution function Methods 0.000 claims abstract description 46
- 230000006870 function Effects 0.000 claims abstract description 23
- 238000007635 classification algorithm Methods 0.000 claims description 32
- 238000004422 calculation algorithm Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000012163 sequencing technique Methods 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 5
- 102220479871 Protein FAM180A_S53A_mutation Human genes 0.000 claims description 3
- 102220536518 THAP domain-containing protein 1_S51A_mutation Human genes 0.000 claims description 3
- 102220536512 THAP domain-containing protein 1_S52A_mutation Human genes 0.000 claims description 3
- 102220536494 THAP domain-containing protein 1_S55A_mutation Human genes 0.000 claims description 3
- 238000007637 random forest analysis Methods 0.000 claims description 3
- 102220313179 rs1553259785 Human genes 0.000 claims description 3
- 230000007334 memory performance Effects 0.000 claims description 2
- 238000011056 performance test Methods 0.000 description 6
- 238000013112 stability test Methods 0.000 description 6
- 238000010998 test method Methods 0.000 description 6
- 238000012545 processing Methods 0.000 description 4
- 238000001914 filtration Methods 0.000 description 3
- 238000012216 screening Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000007812 deficiency Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000012430 stability testing Methods 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 101100498818 Arabidopsis thaliana DDR4 gene Proteins 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002085 persistent effect Effects 0.000 description 1
- 238000013403 standard screening design Methods 0.000 description 1
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
- G11C29/08—Functional testing, e.g. testing during refresh, power-on self testing [POST] or distributed testing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
- G06F18/24155—Bayesian classification
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Mathematics (AREA)
- Mathematical Physics (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Algebra (AREA)
- Operations Research (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Techniques For Improving Reliability Of Storages (AREA)
Abstract
The invention provides a method and a device for carrying out grading evaluation on a PMEM memory, wherein the method comprises the following steps: s1, layering test items of a PMEM memory according to a working mode; s2, acquiring a PMEM memory sample set, testing according to layered test items, and recording layered test results; s3, calculating a probability distribution function of a test result in the PMEM memory sample set; s4, dividing level intervals on the probability distribution function, setting weight values for each level interval, counting the number of test results in each level interval, and carrying out weighted summation on each PMEM memory to obtain an evaluation index; s5, grading evaluation is carried out on each PMEM memory according to the evaluation index. The invention realizes the grading evaluation of the PMEM memory so as to carry out the quality sorting, function recommendation and risk early warning of the PMEM memory.
Description
Technical Field
The invention belongs to the technical field of memory testing, and particularly relates to a method and a device for grading evaluation of a PMEM memory.
Background
With the development of memory technology, new memory types have been developed. PMEM is a new type of memory today, also known as DPCM or DCPMM. Sub-products of PMEM memory include AEP and BPS, and CPS about to be released in 2021, and PMEM of each generation corresponds to each generation of platform.
Compared with the common memory, the PMEM memory has the following characteristics: 1. as a large capacity memory, data and process loading time can be saved after the server is restarted. 2. The total cost is reduced, the quantity of devices is intelligently increased for the CPU, the situation that the memory capacity is insufficient is avoided, the software authorization is increased, the cabinet is increased, the energy consumption is increased, and meanwhile, the original scattered service can be combined. 3. PMEM memory can provide relatively large capacity while providing read-write speeds and delays approaching those of DRAM, which are not comparable to conventional SSDs. 4. The single available capacity is large and compatible with DDR4 slots.
Various modes of the PMEM memory can be set to meet the use of various functions, and the memory modes of the PMEM mainly comprise the following modes: 1. the memory mode is suitable for the memory with large capacity. 2. Direct access mode is applied in which the PMEM memory is persistent as if it were stored. 3. Hybrid mode, i.e., a mode in which a portion is charged with memory and a portion is in an application direct access mode, in which applications can use high performance storage to avoid transferring data back and forth on the I/O bus.
The test method of the PMEM memory is very immature, the existing test method can only simply evaluate according to certain test items, and can not only singly evaluate the single PMEM memory, but also evaluate the excellent grade of a batch of PMEM memory products.
This is a deficiency of the prior art, and therefore, it is desirable to provide a method and apparatus for hierarchical evaluation of PMEM memory that addresses the above-described deficiencies of the prior art.
Disclosure of Invention
Aiming at the defects that the prior memory test method in the prior art cannot perform single evaluation on single PMEM memories and cannot perform excellent middle-level and poor-level evaluation on a batch of PMEM memory products, the invention provides a method and a device for performing grading evaluation on PMEM memories, and aims to solve the technical problems.
In a first aspect, the present invention provides a method for hierarchical evaluation of PMEM memory, comprising the steps of:
s1, layering test items of a PMEM memory according to a working mode;
s2, acquiring a PMEM memory sample set, testing according to layered test items, and recording layered test results;
s3, calculating a probability distribution function of a test result in the PMEM memory sample set;
s4, dividing level intervals on the probability distribution function, setting weight values for each level interval, counting the number of test results in each level interval, and carrying out weighted summation on each PMEM memory to obtain an evaluation index;
s5, grading evaluation is carried out on each PMEM memory according to the evaluation index.
Further, the specific steps of step S1 are as follows:
s11, acquiring a major class of the PMEM memory;
s12, acquiring a test mode of each major class of the PMEM memory;
s13, acquiring test sub-items of each test mode under each major class of the PMEM memory;
s14, layering all test sub-items of the PMEM memory according to a large class and a test mode. And layering the PMEM memory according to the major categories, the test modes and the test sub-items under each test mode, testing comprehensively, and facilitating subsequent layer-by-layer statistics of test results.
Further, the specific steps of S2 are as follows:
s21, acquiring a PMEM memory sample set;
s22, testing each test subitem of each PMEM in the PMEM memory sample set for preset times, and recording each test result;
s23, judging the test of each test sub-item, and judging whether the Bernoulli distribution function is met;
if yes, go to step S24;
if not, eliminating the PMEM memory corresponding to the test subitem;
s24, removing the maximum value and the minimum value of all the test results of each test sub-item, and taking the average value as the test result of the corresponding test sub-item. And (5) performing primary screening on the test result through the Bernoulli distribution function, and filtering coarse errors.
Further, the specific steps of step S3 are as follows:
s31, counting test results of test sub-items of each PMEM memory in the PMEM memory sample set, and setting the test results as test values;
s32, calculating the mean value and variance of all test values in the PMEM memory sample set;
s33, taking the mean value of all the test values as offset, taking the variance of all the test values as amplitude, and generating a Gaussian distribution function taking the test values as variables and taking the probability as an output value. The Gaussian distribution function reflects probability distribution conditions of all test results in the PMEM memory sample set.
Further, the specific steps of step S4 are as follows:
s41, dividing the Gaussian distribution function into four sections according to a curve symmetry axis and turning points;
s42, dividing the four intervals into four grades according to the variable size, setting weights for each grade, and setting interval weights with large variable values larger than interval weights with small variable values;
s43, counting the number of test values in each level interval in each PMEM, and carrying out weighted summation according to set weights to obtain an evaluation index of each PMEM memory. And grading the test result according to a probability distribution function, wherein the interval with a large variable value has a large influence on the level, increasing the weight of the interval with a large variable value, and the interval with a small variable value has a small influence on the level, and reducing the weight of the interval with a small variable value.
Further, the specific steps of step S5 are as follows:
s51, sequencing the PMEM memories in the PMEM memory sample set according to the evaluation index;
s52, the PMEM memory performance with large evaluation index value is higher than that of the PMEM memory with low evaluation index value. The evaluation index sequencing realizes the simplest hierarchical evaluation of the PMEM memory.
Further, the specific steps of step S5 are as follows:
S51A, creating a PMEM memory classification algorithm model;
S52A, acquiring evaluation indexes and actual index performances of each PMEM memory in a PMEM memory sample set;
S53A, training a PMEM memory classification algorithm model through evaluation indexes and actual index performances of each PMEM memory;
S54A, predicting, grading and evaluating the PMEM memory in the new PMEM memory sample set by using a PMEM memory classification algorithm model to obtain a classification result;
S55A, obtaining time index performance of the PMEM memory in the new PMEM memory sample set, comparing the time index performance with a classification result of prediction grading evaluation, and correcting a PMEM memory classification algorithm model. The classification evaluation prediction of the PMEM memory can be realized by creating a PMEM memory classification algorithm model and continuously training the PMEM memory classification algorithm model.
Further, the PMEM memory classification algorithm model adopts one or more of a Bayesian algorithm, a random forest algorithm or a mean algorithm.
Further, setting classification result parameters in the PMEM memory classification algorithm model;
when the number of samples in the PMEM memory sample set is smaller than a set threshold, setting the classification result parameter weight to 0;
when the sum of the number of samples in the PMEM memory sample set and the number of samples in the new PMEM memory sample set reaches a set threshold value, increasing the weight of the classification result parameters;
when the test requirement is to evaluate a certain function or a certain mode of the PMEM memory in a grading manner, setting the weight of the parameter corresponding to the function or mode of the non-to-be-tested item in the PMEM memory classification algorithm model to 0. When the number of samples in the PMEM memory sample set is smaller, the parameter weight of the classification result can be set to 0, and the response weight is improved after the PMEM memory classification algorithm model is trained accurately; when performing a hierarchical evaluation for a certain function or a certain mode of the PMEM memory, for example, only considering the level in the memory mode, the parameter weight of the other modes may be set to 0.
In a second aspect, the present invention provides an apparatus for hierarchical evaluation of PMEM memory, comprising:
the test item layering module is used for layering the test items of the PMEM memory according to the working mode;
the layering test module is used for acquiring a PMEM memory sample set, testing according to layering test items and recording layering test results;
the probability distribution function calculation module is used for calculating a probability distribution function of the test result in the PMEM memory sample set;
the evaluation index calculation module is used for dividing the level intervals on the probability distribution function, setting a weight for each level interval, counting the number of test results in each level interval, and carrying out weighted summation for each PMEM memory to obtain an evaluation index;
and the PMEM memory grading evaluation module is used for grading and evaluating each PMEM memory according to the evaluation index.
The invention has the advantages that,
the method and the device for grading and evaluating the PMEM memories realize grading and evaluating the PMEM memories, can sort the quality of a batch of PMEM memories according to the evaluation index, can recommend the most suitable pieces in a batch of PMEM memories according to a certain functional requirement, and can also perform risk early warning on the PMEM memories according to the evaluation index.
In addition, the invention has reliable design principle, simple structure and very wide application prospect.
It can be seen that the present invention has outstanding substantial features and significant advances over the prior art, as well as the benefits of its implementation.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the description of the embodiments or the prior art will be briefly described below, and it will be obvious to those skilled in the art that other drawings can be obtained from these drawings without inventive effort.
FIG. 1 is a flow chart of a method for hierarchical evaluation of PMEM memory according to the present invention.
FIG. 2 is a flow chart of a method for hierarchical evaluation of PMEM memory according to the present invention.
FIG. 3 is a flowchart of a method for hierarchical evaluation of PMEM memory according to the present invention.
FIG. 4 is a schematic diagram of an apparatus for hierarchical evaluation of PMEM memory according to the present invention.
FIG. 5 is a graph showing the Gaussian distribution function in example 4 of the present invention.
In the figure, 1-test item layering module; 2-layering test module; 3-a probability distribution function calculation module; 4-an evaluation index calculation module; and a 5-PMEM memory grading evaluation module.
Detailed Description
In order to make the technical solution of the present invention better understood by those skilled in the art, the technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
Example 1:
as shown in fig. 1, the present invention provides a method for grading evaluation of PMEM memory, comprising the following steps:
s1, layering test items of a PMEM memory according to a working mode;
s2, acquiring a PMEM memory sample set, testing according to layered test items, and recording layered test results;
s3, calculating a probability distribution function of a test result in the PMEM memory sample set;
s4, dividing level intervals on the probability distribution function, setting weight values for each level interval, counting the number of test results in each level interval, and carrying out weighted summation on each PMEM memory to obtain an evaluation index;
s5, grading evaluation is carried out on each PMEM memory according to the evaluation index.
The method for grading and evaluating the PMEM memories realizes grading and evaluating the PMEM memories, can sort the quality of a batch of the PMEM memories according to the evaluation index, can recommend the most suitable pieces in the batch of the PMEM memories according to a certain functional requirement, and can also perform risk early warning on the PMEM memories according to the evaluation index.
Example 2:
as shown in fig. 2, the present invention provides a method for grading evaluation of PMEM memory, including the following steps:
s1, layering test items of a PMEM memory according to a working mode; the method comprises the following specific steps:
s11, acquiring a major class of the PMEM memory;
s12, acquiring a test mode of each major class of the PMEM memory;
s13, acquiring test sub-items of each test mode under each major class of the PMEM memory;
s14, layering all test sub-items of the PMEM memory according to a major class and a test mode; layering the PMEM memory according to the major categories, the test modes and the test sub items under each test mode, comprehensively testing, and facilitating subsequent layer-by-layer statistics of test results;
s2, acquiring a PMEM memory sample set, testing according to layered test items, and recording layered test results; the method comprises the following specific steps:
s21, acquiring a PMEM memory sample set;
s22, testing each test subitem of each PMEM in the PMEM memory sample set for preset times, and recording each test result;
s23, judging the test of each test sub-item, and judging whether the Bernoulli distribution function is met;
if yes, go to step S24;
if not, eliminating the PMEM memory corresponding to the test subitem;
s24, removing the maximum value and the minimum value of all the test results of each test sub-item, and taking the average value as the test result of the corresponding test sub-item; performing primary screening on the test result through a Bernoulli distribution function, and filtering coarse errors;
s3, calculating a probability distribution function of a test result in the PMEM memory sample set; the method comprises the following specific steps:
s31, counting test results of test sub-items of each PMEM memory in the PMEM memory sample set, and setting the test results as test values;
s32, calculating the mean value and variance of all test values in the PMEM memory sample set;
s33, taking the mean value of all the test values as offset, taking the variance of all the test values as amplitude, and generating a Gaussian distribution function taking the test values as variables and the probability as an output value; the Gaussian distribution function reflects probability distribution conditions of all test results in the PMEM memory sample set;
s4, dividing level intervals on the probability distribution function, setting weight values for each level interval, counting the number of test results in each level interval, and carrying out weighted summation on each PMEM memory to obtain an evaluation index; the method comprises the following specific steps:
s41, dividing the Gaussian distribution function into four sections according to a curve symmetry axis and turning points;
s42, dividing the four intervals into four grades according to the variable size, setting weights for each grade, and setting interval weights with large variable values larger than interval weights with small variable values;
s43, counting the number of test values in each level interval in each PMEM, and carrying out weighted summation according to set weights to obtain an evaluation index of each PMEM memory; grading the test result according to a probability distribution function, wherein the interval with a large variable value has a large influence on the level, increasing the weight of the interval with a large variable value, and the interval with a small variable value has a small influence on the level, and reducing the weight of the interval with a small variable value;
s5, grading evaluation is carried out on each PMEM memory according to the evaluation index;
the method comprises the following specific steps:
s51, sequencing the PMEM memories in the PMEM memory sample set according to the evaluation index;
s52, the PMEM memory with large evaluation index value has higher performance than the PMEM memory with low evaluation index value; the evaluation index sequencing realizes the simplest hierarchical evaluation of the PMEM memory.
Example 3:
as shown in fig. 3, the present invention provides a method for grading evaluation of PMEM memory, comprising the steps of:
s1, layering test items of a PMEM memory according to a working mode; the method comprises the following specific steps:
s11, acquiring a major class of the PMEM memory;
s12, acquiring a test mode of each major class of the PMEM memory;
s13, acquiring test sub-items of each test mode under each major class of the PMEM memory;
s14, layering all test sub-items of the PMEM memory according to a major class and a test mode; layering the PMEM memory according to the major categories, the test modes and the test sub items under each test mode, comprehensively testing, and facilitating subsequent layer-by-layer statistics of test results;
s2, acquiring a PMEM memory sample set, testing according to layered test items, and recording layered test results; the method comprises the following specific steps:
s21, acquiring a PMEM memory sample set;
s22, testing each test subitem of each PMEM in the PMEM memory sample set for preset times, and recording each test result;
s23, judging the test of each test sub-item, and judging whether the Bernoulli distribution function is met;
if yes, go to step S24;
if not, eliminating the PMEM memory corresponding to the test subitem;
s24, removing the maximum value and the minimum value of all the test results of each test sub-item, and taking the average value as the test result of the corresponding test sub-item; performing primary screening on the test result through a Bernoulli distribution function, and filtering coarse errors;
s3, calculating a probability distribution function of a test result in the PMEM memory sample set; the method comprises the following specific steps:
s31, counting test results of test sub-items of each PMEM memory in the PMEM memory sample set, and setting the test results as test values;
s32, calculating the mean value and variance of all test values in the PMEM memory sample set;
s33, taking the mean value of all the test values as offset, taking the variance of all the test values as amplitude, and generating a Gaussian distribution function taking the test values as variables and the probability as an output value; the Gaussian distribution function reflects probability distribution conditions of all test results in the PMEM memory sample set;
s4, dividing level intervals on the probability distribution function, setting weight values for each level interval, counting the number of test results in each level interval, and carrying out weighted summation on each PMEM memory to obtain an evaluation index; the method comprises the following specific steps:
s41, dividing the Gaussian distribution function into four sections according to a curve symmetry axis and turning points;
s42, dividing the four intervals into four grades according to the variable size, setting weights for each grade, and setting interval weights with large variable values larger than interval weights with small variable values;
s43, counting the number of test values in each level interval in each PMEM, and carrying out weighted summation according to set weights to obtain an evaluation index of each PMEM memory; grading the test result according to a probability distribution function, wherein the interval with a large variable value has a large influence on the level, increasing the weight of the interval with a large variable value, and the interval with a small variable value has a small influence on the level, and reducing the weight of the interval with a small variable value;
s5, grading evaluation is carried out on each PMEM memory according to the evaluation index; the method comprises the following specific steps:
S51A, creating a PMEM memory classification algorithm model; the PMEM memory classification algorithm model adopts one or more of a Bayesian algorithm, a random forest algorithm or a mean algorithm;
S52A, acquiring evaluation indexes and actual index performances of each PMEM memory in a PMEM memory sample set;
S53A, training a PMEM memory classification algorithm model through evaluation indexes and actual index performances of each PMEM memory;
S54A, predicting, grading and evaluating the PMEM memory in the new PMEM memory sample set by using a PMEM memory classification algorithm model to obtain a classification result;
S55A, obtaining time index expression of the PMEM memory in a new PMEM memory sample set, comparing the time index expression with a classification result of predictive classification evaluation, and correcting a PMEM memory classification algorithm model; the classification evaluation prediction of the PMEM memory can be realized by creating a PMEM memory classification algorithm model and continuously training the PMEM memory classification algorithm model.
In one embodiment, setting classification result parameters in a PMEM memory classification algorithm model;
when the number of samples in the PMEM memory sample set is smaller than a set threshold, setting the classification result parameter weight to 0;
when the sum of the number of samples in the PMEM memory sample set and the number of samples in the new PMEM memory sample set reaches a set threshold value, increasing the weight of the classification result parameters;
when the test requirement is to carry out grading evaluation on a certain function or a certain mode of the PMEM memory, setting the weight of the parameter corresponding to the function or the mode which is not to be tested in the PMEM memory classification algorithm model to 0; when the number of samples in the PMEM memory sample set is smaller, the parameter weight of the classification result can be set to 0, and the response weight is improved after the PMEM memory classification algorithm model is trained accurately; when performing a hierarchical evaluation for a certain function or a certain mode of the PMEM memory, for example, only considering the level in the memory mode, the parameter weight of the other modes may be set to 0.
The test and evaluation method of the original PMEM memory comprises the following steps: one is to follow the manufacturer's test guidelines entirely. According to the test file provided by the manufacturer or the proposal of the manufacturer personnel; and secondly, adopting a test method which is the same as that of a common memory. Such as stability testing, performance testing, RMT testing, etc.; and thirdly, adopting a test method which is the same as that of a common hard disk. Such as stability testing, performance testing, signal testing, etc.; fourth, only a certain function of the PMEM is tested, such as a memory mode. The second and third methods are all carried out by adopting a common method, and are not carried out for the new PMEM product, and the second and third methods can evaluate whether the product is qualified or not, and the evaluation method of one cut can not evaluate the product; the method completely follows the test guidance of manufacturers, so that the reliability of the test evaluation method is low, and the phenomenon that certain test items are desalted for hiding certain defects cannot be avoided; the fourth method can not comprehensively measure the condition of the PMEM memories, and too single method can cause a batch of PMEM memories to miss the PMEM memories with different characteristics, and only one characteristic of the PMEM memories is reserved; the four methods are characterized in that the whole PMEM memory cannot be classified, and the novel memory PMEM cannot be effectively subjected to secondary test.
The invention carries out grading evaluation on the PMEM memory by combining the data processing with the PMEM memory classification algorithm model through the Bernoulli distribution function, and the data processing and the algorithm model are mutually iterated and mutually verified, so that a simple mathematical characteristic processing mode is abandoned, the processing is carried out in a probability mode, and the invention not only can carry out overall evaluation, but also can carry out specific item evaluation.
Example 4:
in the above embodiment 3, step S1 performs test item layering on the PMEM memory according to the working mode, and specifically includes the following steps:
obtaining a main class of PMEM memory, including AEP memory, BPS memory and CPS memory; wherein AEP is a mature generation PMEM memory product, BPS is a memory product in the development stage, and CPS memory is a PMEM memory product planned in the future;
each major class comprises three basic modes, namely a memory mode, an application direct access mode and a mixed mode, wherein the application direct access mode comprises two sub-modes, namely an application direct access interleaving mode and an application direct access non-interleaving mode; the mixed mode also comprises two modes, namely a mixed mode of a memory mode and an application direct access interleaving mode, and a mixed mode of a memory mode and an application direct access non-interleaving mode;
taking AEP memory as an example, five seed patterns: the memory mode, the application direct reading mode, the mixed mode of the memory mode and the application direct reading interleaving mode, and the mixed mode of the memory mode and the application direct reading non-interleaving mode are respectively marked as an A mode, a B mode, a C mode, a D mode and an E mode;
the test sub-items in the A mode comprise an RMT test, a stability test, a performance test and an information checking test; the RMT test is an identification memory test;
the test sub-items in the B mode comprise a new function test, a stability test, a performance test and an information checking test;
the test sub-items in the C mode comprise a new function test, a stability test, a performance test and an information checking test;
the test sub-items in the D mode comprise a new function test, a stability test, a performance test and an information checking test;
the test sub-items in the E mode comprise a new function test, a stability test, a performance test and an information checking test;
the RMT test result in the a mode is designated as a11; the stability test comprises a restarting test, an AC power-off test and a DC power-off test, and test results are recorded as A21, A22 and A23; the performance test comprises MLC test and PTU test, and test results are marked as A31 and A32; the information checking test result is marked as A41; MLC test is intel memory delay test;
the RMT test in a mode was performed 10 times, and the results were designated as a110, a111, a112, a113, a114, a115, a116, a117, a118, a119;
evaluating the result of the RMT test with bernoulli distribution, if P (x=fail) >0; the AEP memory is directly eliminated;
if P (x=fail) =0, RMT test 10 test results, the maximum and minimum values thereof are removed. The remaining 8 values are averaged and assigned to A11;
other tests in the A mode are processed by referring to the RMT test, and are respectively assigned to A21, A22, A23, A31, A32 and A41;
B. c, D, E mode imitates A mode to process data;
the parameters finally output in the B mode are assigned to B11, B12, B21, B22, B23, B31, B32 and B41; the new function test in the B mode comprises an ADR test and a secure erase test, which are respectively assigned to B11 and B12;
the final output parameters in the C mode are assigned to C11, C12, C21, C22, C23, C31, C32 and C41; the new function test in the C mode comprises an ADR test and a secure erase test, and the ADR test and the secure erase test are respectively assigned to the B11 and the B12;
the final output parameters in the D mode are assigned to D11, D12, D21, D22, D23, D31, D32 and D41; the new function test in the D mode comprises an ADR test and a secure erase test, which are respectively assigned to D11 and D12;
the final output parameters in E mode are assigned to E11, E12, E21, E22, E23, E31, E32 and E41; the new function test in the E mode comprises an ADR test and a secure erase test, which are respectively assigned to E11 and E12;
taking 10 AEPs as an example, they are denoted as P0, P1, P2, P3, P4, P5, P6, P7, P8, P9, respectively.
The test results of P0 may be recorded as P0a11, P0a21, P0a22, P0a23, P0a31, P0a32, P0a41, P0B11, P0B12, P0B21, P0B22, P0B23, P0B31, P0B32, P0B41, P0C11, P0C12, P0C21, P0C22, P0C23, P0C31, P0C32, P0C41, P0D11, P0D12, P0D21, P0D22, P0D23, P0D31, P0D32, P0E11, P0E12, P0E21, P0E22, P0E23, P0E31, P0E32, P0E41, respectively;
p1, P2, P3, P4, P5, P6, P7, P8, P9 refer to the recording mode of P0;
generating a Gaussian distribution function with the test value as a variable and the probability as an output value by taking the average value of all the test values as offset and the variance of all the test values as amplitudeWherein μ is the mean of all test values and σ is the variance of all test values;
each test value of the 10 AEP memories obtains probability distribution, and uses the symmetry axis O, the first turning point M and the second turning point N as grading basis, and can be divided into four grades, as shown in fig. 5, the right side of the second turning point N is the first grade, the second turning point N to the symmetry axis O is the second grade, the symmetry axis 0 to the first turning point M is the third grade, and the left side of the first turning point M is the fourth grade;
the number of the grades is compared to rank 10 AEPs, the four grades are respectively weighted, the grading comparison is further carried out, the first grade weight is 40%, the second grade weight is 30%, the third grade weight is 20%, the fourth grade weight is 10%, and the total value is the evaluation index;
for example, 10 parameters in P0 are at the first level, 9 parameters are at the second level, 0 parameters are at the third level, and 20 parameters are at the fourth level; in P1, 0 parameters are in a first level, 20 parameters are in a second level, 19 parameters are in a third level, and 0 parameters are in a fourth level;
calculate p0=10x0.4+9x0.3+0x0.2+20x0.1=8.7;
P1＝0x0.4+20x0.3+19x0.2+0x0.1＝9.8；
since P1 is greater than P0, P1 is preferred over P0.
Example 5:
as shown in fig. 4, the present invention provides an apparatus for hierarchical evaluation of PMEM memory, including:
the test item layering module 1 is used for layering the test items of the PMEM memory according to the working mode;
the layering test module 2 is used for acquiring a PMEM memory sample set, testing according to layering test items and recording layering test results;
the probability distribution function calculation module 3 is used for calculating a probability distribution function of the test result in the PMEM memory sample set;
the evaluation index calculation module 4 is used for dividing the level intervals on the probability distribution function, setting a weight for each level interval, counting the number of test results in each level interval, and carrying out weighted summation for each PMEM memory to obtain an evaluation index;
and the PMEM memory grading evaluation module 5 is used for grading and evaluating each PMEM memory according to the evaluation index.
Although the present invention has been described in detail by way of preferred embodiments with reference to the accompanying drawings, the present invention is not limited thereto. Various equivalent modifications and substitutions may be made in the embodiments of the present invention by those skilled in the art without departing from the spirit and scope of the present invention, and it is intended that all such modifications and substitutions be within the scope of the present invention/be within the scope of the present invention as defined by the appended claims. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (9)
1. A method for hierarchical evaluation of PMEM memory, comprising the steps of:
s1, layering test items of a PMEM memory according to a working mode; the specific steps of the step S1 are as follows:
s11, acquiring a major class of the PMEM memory, wherein the major class comprises an AEP memory, a BPS memory and a CPS memory;
s12, acquiring a test mode of each major class of the PMEM memory, wherein each major class comprises three basic modes, namely a memory mode, an application direct access mode and a mixed mode;
s13, acquiring test sub-items of each test mode under each major class of the PMEM memory; the application direct access mode includes a test sub-item of an application direct access interleaving mode and an application direct access non-interleaving mode; the mixed mode comprises a test subitem of a mixed mode of a memory mode and an application direct access interleaving mode, and a test subitem of a mixed mode of a memory mode and an application direct access non-interleaving mode;
s14, layering all test sub-items of the PMEM memory according to a major class and a test mode; when the test requirement is to carry out grading evaluation on a certain function or a certain mode of the PMEM memory, setting the weight of the parameter corresponding to the function or the mode which is not to be tested in the PMEM memory classification algorithm model to 0; when the number of samples in the PMEM memory sample set is smaller than a set threshold value, setting the classification result parameter weight to 0 until the PMEM memory classification algorithm model is accurately trained, and then improving the corresponding weight;
s2, acquiring a PMEM memory sample set, testing according to layered test items, and recording layered test results;
s3, calculating a probability distribution function of a test result in the PMEM memory sample set;
s4, dividing level intervals on the probability distribution function, setting weight values for each level interval, counting the number of test results in each level interval, and carrying out weighted summation on each PMEM memory to obtain an evaluation index;
s5, grading evaluation is carried out on each PMEM memory according to the evaluation index.
2. The method for hierarchical evaluation of PMEM memory according to claim 1, wherein S2 comprises the steps of:
s21, acquiring a PMEM memory sample set;
s22, testing each test subitem of each PMEM in the PMEM memory sample set for preset times, and recording each test result;
s23, judging the test of each test sub-item, and judging whether the Bernoulli distribution function is met;
if yes, go to step S24;
if not, eliminating the PMEM memory corresponding to the test subitem;
s24, removing the maximum value and the minimum value of all the test results of each test sub-item, and taking the average value as the test result of the corresponding test sub-item.
3. The method for hierarchical evaluation of PMEM memory according to claim 2, wherein step S3 is specifically as follows:
s31, counting test results of test sub-items of each PMEM memory in the PMEM memory sample set, and setting the test results as test values;
s32, calculating the mean value and variance of all test values in the PMEM memory sample set;
s33, taking the mean value of all the test values as offset, taking the variance of all the test values as amplitude, and generating a Gaussian distribution function taking the test values as variables and taking the probability as an output value.
4. A method for hierarchical evaluation of PMEM memory according to claim 3, wherein step S4 is specifically as follows:
s41, dividing the Gaussian distribution function into four sections according to a curve symmetry axis and turning points;
s42, dividing the four intervals into four grades according to the variable size, setting weights for each grade, and setting interval weights with large variable values larger than interval weights with small variable values;
s43, counting the number of test values in each level interval in each PMEM, and carrying out weighted summation according to set weights to obtain an evaluation index of each PMEM memory.
5. The method for hierarchical evaluation of PMEM memory according to claim 4, wherein step S5 is specifically as follows:
s51, sequencing the PMEM memories in the PMEM memory sample set according to the evaluation index;
s52, the PMEM memory performance with large evaluation index value is higher than that of the PMEM memory with low evaluation index value.
6. The method for hierarchical evaluation of PMEM memory according to claim 5, wherein step S5 is specifically as follows:
S51A, creating a PMEM memory classification algorithm model;
S52A, acquiring evaluation indexes and actual index performances of each PMEM memory in a PMEM memory sample set;
S53A, training a PMEM memory classification algorithm model through evaluation indexes and actual index performances of each PMEM memory;
S54A, predicting, grading and evaluating the PMEM memory in the new PMEM memory sample set by using a PMEM memory classification algorithm model to obtain a classification result;
S55A, obtaining time index performance of the PMEM memory in the new PMEM memory sample set, comparing the time index performance with a classification result of prediction grading evaluation, and correcting a PMEM memory classification algorithm model.
7. The method of hierarchical evaluation of PMEM memory according to claim 6, wherein the PMEM memory classification algorithm model employs one or more of a bayesian algorithm, a random forest algorithm, or a mean algorithm.
8. The method for hierarchical evaluation of PMEM memory according to claim 6, wherein classification result parameters are set in a PMEM memory classification algorithm model;
when the number of samples in the PMEM memory sample set is smaller than a set threshold, setting the classification result parameter weight to 0;
when the sum of the number of samples in the PMEM memory sample set and the number of samples in the new PMEM memory sample set reaches a set threshold value, increasing the weight of the classification result parameters;
when the test requirement is to evaluate a certain function or a certain mode of the PMEM memory in a grading manner, setting the weight of the parameter corresponding to the function or mode of the non-to-be-tested item in the PMEM memory classification algorithm model to 0.
9. An apparatus for hierarchical evaluation of PMEM memory, comprising:
the test item layering module (1) is used for layering the test items of the PMEM memory according to the working mode;
acquiring a major class of the PMEM memory; acquiring a test mode of each major class of the PMEM memory; acquiring test sub-items of each test mode under each major class of the PMEM memory; layering all test sub-items of the PMEM memory according to a large class and a test mode;
obtaining a main class of PMEM memory, including AEP memory, BPS memory and CPS memory;
the method comprises the steps of obtaining a test mode of a PMEM memory under each major class, wherein each major class comprises three basic modes, a memory mode, an application direct access mode and a mixed mode;
acquiring test sub-items of each test mode under each major class of the PMEM memory; the application direct access mode includes a test sub-item of an application direct access interleaving mode and an application direct access non-interleaving mode; the mixed mode comprises a test subitem of a mixed mode of a memory mode and an application direct access interleaving mode, and a test subitem of a mixed mode of a memory mode and an application direct access non-interleaving mode;
layering all test sub-items of the PMEM memory according to a large class and a test mode; when the test requirement is to carry out grading evaluation on a certain function or a certain mode of the PMEM memory, setting the weight of the parameter corresponding to the function or the mode which is not to be tested in the PMEM memory classification algorithm model to 0; when the number of samples in the PMEM memory sample set is smaller than a set threshold value, setting the classification result parameter weight to 0 until the PMEM memory classification algorithm model is accurately trained, and then improving the corresponding weight;
the layering test module (2) is used for acquiring a PMEM memory sample set, testing according to layering test items and recording layering test results;
the probability distribution function calculation module (3) is used for calculating a probability distribution function of a test result in the PMEM memory sample set;
the evaluation index calculation module (4) is used for dividing the level intervals on the probability distribution function, setting a weight for each level interval, counting the number of test results in each level interval, and carrying out weighted summation for each PMEM memory to obtain an evaluation index;
and the PMEM memory grading evaluation module (5) is used for grading and evaluating each PMEM memory according to the evaluation index.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110959424.0A CN113782086B (en) | 2021-08-20 | 2021-08-20 | Method and device for grading evaluation of PMEM memory |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110959424.0A CN113782086B (en) | 2021-08-20 | 2021-08-20 | Method and device for grading evaluation of PMEM memory |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113782086A CN113782086A (en) | 2021-12-10 |
| CN113782086B true CN113782086B (en) | 2023-08-08 |
Family
ID=78838465
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110959424.0A Active CN113782086B (en) | 2021-08-20 | 2021-08-20 | Method and device for grading evaluation of PMEM memory |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113782086B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114691503B (en) * | 2022-03-22 | 2022-09-13 | 航天中认软件测评科技（北京）有限责任公司 | Test-oriented management method, device, equipment and medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102681940A (en) * | 2012-05-15 | 2012-09-19 | 兰雨晴 | Method for carrying out performance test on memory management subsystem of Linux operation system |
| CN110163459A (en) * | 2018-03-23 | 2019-08-23 | 河南工业大学 | A method of building multiple index evaluation model is classified wheat quality |
| CN112434931A (en) * | 2020-11-20 | 2021-03-02 | 首钢京唐钢铁联合有限责任公司 | Evaluation method for operation index of measurement management system |
| CN112700817A (en) * | 2021-01-18 | 2021-04-23 | 皇虎测试科技(深圳)有限公司 | Memory device quality evaluation method and device and computer readable storage medium |
-
2021
- 2021-08-20 CN CN202110959424.0A patent/CN113782086B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102681940A (en) * | 2012-05-15 | 2012-09-19 | 兰雨晴 | Method for carrying out performance test on memory management subsystem of Linux operation system |
| CN110163459A (en) * | 2018-03-23 | 2019-08-23 | 河南工业大学 | A method of building multiple index evaluation model is classified wheat quality |
| CN112434931A (en) * | 2020-11-20 | 2021-03-02 | 首钢京唐钢铁联合有限责任公司 | Evaluation method for operation index of measurement management system |
| CN112700817A (en) * | 2021-01-18 | 2021-04-23 | 皇虎测试科技(深圳)有限公司 | Memory device quality evaluation method and device and computer readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113782086A (en) | 2021-12-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110634080B (en) | Abnormal electricity utilization detection method, device, equipment and computer readable storage medium | |
| CN113139621B (en) | Method and apparatus for determining sufficiency of defect data for classification | |
| CN107169768B (en) | Method and device for acquiring abnormal transaction data | |
| CN113792825B (en) | Fault classification model training method and device for electricity information acquisition equipment | |
| CN111914090B (en) | Method and device for enterprise industry classification identification and characteristic pollutant identification | |
| CN108090678B (en) | Data model monitoring method, system, equipment and computer storage medium | |
| TW201407154A (en) | Integration of automatic and manual defect classification | |
| CN110111113B (en) | Abnormal transaction node detection method and device | |
| US20040002929A1 (en) | System and method for mining model accuracy display | |
| CN113782086B (en) | Method and device for grading evaluation of PMEM memory | |
| CN110191159A (en) | A kind of load regulation method and system, the equipment of Resource Server | |
| WO2023123869A1 (en) | Visibility value measurement method and apparatus, device, and storage medium | |
| CN111860698A (en) | Method and device for determining stability of learning model | |
| Angenent et al. | Large-scale machine learning for business sector prediction | |
| CN111626855A (en) | Bond credit interest difference prediction method and system | |
| CN111783883A (en) | Abnormal data detection method and device | |
| CN114881343B (en) | Short-term load prediction method and device for power system based on feature selection | |
| CN114926261A (en) | Method and medium for predicting fraud probability of automobile financial user application | |
| CN115081950A (en) | Enterprise growth assessment modeling method, system, computer and readable storage medium | |
| CN110852443B (en) | Feature stability detection method, device and computer readable medium | |
| CN110992043B (en) | Method and device for mining risk entity | |
| CN110472801B (en) | Electromagnetic environment assessment method and system for direct-current transmission line | |
| CN113420772A (en) | Defect detection method and device based on multi-classifier and SVDD (singular value decomposition and direct decomposition) cooperative algorithm | |
| CN114820003A (en) | Pricing information abnormity identification method and device, electronic equipment and storage medium | |
| CN114663102A (en) | Method, equipment and storage medium for predicting debt subject default based on semi-supervised model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |