阅读说明：本技术 一种基于前向相容的折叠集分层压缩方法及装置 (Folding set layered compression method and device based on forward compatibility ) 是由 詹文法 陶鹏程 蔡雪原 邵志伟 彭勇 张振林 丁文祥 彭登辉 华铭 都奕 于 2019-11-28 设计创作，主要内容包括：本发明涉及一种基于前向相容的折叠集分层压缩方法,步骤如下：S1、划分测试向量；S2、从划分的测试向量中,获取待编码数据块与折叠集；S3、根据折叠集依次对待编码数据块进行编码；S4、判断待编码数据块是否为最后一个,如果是输出压缩后的测试向量,如果不是返回执行步骤S3。本发明还公开了一种基于前向相容的折叠集分层压缩装置。本发明通过将测试向量划分,利用参考向量作为折叠种子生成折叠集,依次对剩余的待编码数据块与参考向量进行相容性判断,从而实现降低参考向量的相容条件,使原本不具有相容关系的两个数据块产生相容关系,即增大了数据块间的相容性,从而在原来相容压缩的基础上进一步提高了数据的压缩率。(The invention relates to a folding set layered compression method based on forward compatibility, which comprises the following steps: s1, dividing test vectors; s2, acquiring a data block to be coded and a folding set from the divided test vectors; s3, coding the data blocks to be coded in sequence according to the folding sets; s4, judging whether the data block to be coded is the last one, if so, outputting the compressed test vector, and if not, returning to execute the step S3. The invention also discloses a folding set layered compression device based on forward compatibility. The invention divides the test vector, generates the folding set by using the reference vector as the folding seed, and sequentially judges the compatibility of the residual data blocks to be coded and the reference vector, thereby realizing the reduction of the compatibility condition of the reference vector, leading two data blocks which originally do not have the compatibility relationship to generate the compatibility relationship, namely increasing the compatibility between the data blocks, and further improving the compression ratio of the data on the basis of the original compatibility compression.)

1. A folding set layered compression method based on forward compatibility is characterized by comprising the following steps:

s1, dividing test vectors;

s2, acquiring a data block to be coded and a folding set from the divided test vectors;

s3, coding the data blocks to be coded according to the folding sets in sequence, and triggering the step S4;

s4, judging whether the data block to be coded is the last one, if so, outputting the compressed test vector, and if not, returning to execute the step S3.

2. The method for hierarchical compression of a folded set based on forward compatibility according to claim 1, wherein the step S1 includes

Sequentially dividing the test vectors in the test set into k-bit data blocks with equal length, filling and completing the last data block with no relevant bit, wherein the last data block is less than k bits;

wherein k is a positive integer; in the test set, a test vector comprises a plurality of data blocks which are Z1 and Z2 … … Zn in sequence, wherein n is any positive integer;

each data block is a set of 0,1, x, where x is an independent bit.

3. The forward compatibility-based folded set hierarchical compression method of claim 2, wherein the step S2 includes;

taking a first data block in a sequence formed by data blocks as a reference data block r, wherein the reference data block r is a folding seed s, and performing operation processing on the folding seed s by using a folding counter to generate a folding set F1(x) and storing the folding set F1(x) in an RAM 2; the rest data blocks are used as data blocks to be coded;

where, F1(x) { a0, a1, a2, …, ak-1, ak }, and a0 … ak are data blocks.

4. The forward compatibility based folded set hierarchical compression method of claim 3, wherein the folded set F1(x) is generated by;

the fold counter produces n +1 states s1 from the fold seed s: s1 ═ F (0, s), F (1, s), …, F (n, s), and the set of n +1 states is called a folded set;

where the transition from state F (n-1, s) to F (n, s) indicates that the previous code value of n-1 is kept unchanged, F (n, s) indicates the nth folded set generated by folding the seed s, and n is a positive integer.

5. The method for hierarchical compression of a forward compatibility-based folded set according to claim 1, wherein the step S3 includes:

s31, sequentially judging the compatibility of the data block to be coded and the reference data block r, and coding according to the coding mode graph, including;

1) acquiring a current data block to be encoded, judging the compatibility of the current data block to be encoded and a reference vector r, if the current data block to be encoded is incompatible, executing step 32, if the current data block to be encoded is compatible, executing step 11), and if the current data block to be encoded is reversely compatible, executing step 12);

11) outputting '00', and if the determined bit exists in the data block to be encoded and the position of the reference vector r corresponding to the determined bit in the data block to be encoded is the irrelevant bit X, filling the irrelevant bit X of the reference vector r to update the reference vector r, and then executing the step S4;

wherein the determined bit is 0 or 1; the filling rule is as follows: if the reference vector is compatible with the current database to be coded, filling irrelevant bits in the reference vector with the determined bits; if the reverse compatibility exists, filling irrelevant bits in the reference vector after determining that the bits are inverted;

12) outputting '01', and if the determined bit exists in the data block to be encoded and the position of the reference vector r corresponding to the determined bit in the data block to be encoded is the irrelevant bit X, filling the irrelevant bit X of the reference vector r to update the reference vector r, and executing step S4;

s32, judging the compatibility of the current data block to be coded and the folding set F1(x), and coding according to a coding mode graph, wherein the coding mode graph comprises the following steps;

judging the current data to be codedDatabase a in Block AND F1(x)_iWherein i ═ 1, 2, … k; if the current data block to be coded is a_iIf they are compatible, step 21) is executed, if the current data block to be coded is compatible with a_iBackward compatible, step 22) is performed, if the current data block to be coded is a_iIf not, executing step S33;

21) the output "100" is added with i-1 expressed by binary, the irrelevant bit X of the reference vector r is filled to update the reference vector r, and then step S4 is executed;

wherein i represents a folding distance; where "100" is the prefix and fold set tag combination in this case;

22) outputs "101" plus i-1 in binary representation, fills the irrelevant bit X of the reference vector r to update the reference vector r, and then executes step S4;

wherein "101" is the prefix and fold set tag combination in this case;

s33, outputting 11 plus the data block to be coded; the current data block to be encoded is stored as a new reference data block r in the register RAM1, and the reference vector r is stored as a folding seed generation folding set F2(x) in the RAM2, and then step S4 is performed.

6. The method of claim 5, wherein the current data block to be encoded is consistent with a folded set by: if the current data block to be encoded is equal to F1(x) { a }₀，a₁，a₂，…，a_k-1，a_kAny one data block a in₀，a₁，a₂，…，a_k-1，a_kThere is a compatibility relationship.

7. The forward compatibility-based folded set layered compression method of claim 5, wherein the current data block to be encoded is related to a_iIncompatibility means that the data block to be encoded is not compatible with F1(x) ═ a₀，a₁，a₂，…，a_k-1，a_kAny one data block a in₀，a₁，a₂，…，a_k-1，a_kAre incompatible.

8. The method of claim 5, wherein the consistent means that the code values of corresponding positions in any two data blocks do not collide;

the backward compatibility means that code values of corresponding positions in any two data blocks are in conflict under the condition of not considering irrelevant bits;

the incompatibility refers to the condition other than compatibility and reverse compatibility.

9. The forward compatibility based folded set layered compression method of claim 1, wherein the step S4 includes;

judging the input triggering the step S4, and if the input is the step S31 or S32, executing the step 31); if the input is step S33, execute step S32);

31) judging whether the current data block to be coded is the last data block in a folding set F1(x), if so, ending coding to obtain a compressed test vector; if not, taking the next data block to be encoded of the database to be encoded as the current data block to be encoded, and returning to execute the step S3;

32) judging whether the current data block to be coded is the last data block in a folding set F1(x), if so, ending coding to obtain a compressed test vector; if not, step S3 is performed.

10. A compression apparatus using the forward compatibility-based folded set hierarchical compression method according to any one of claims 1 to 9, comprising;

a partitioning module (301) for partitioning the test vectors;

an obtaining module (302) for obtaining a block to be encoded and a folding set from the divided test vectors;

the encoding module (303) is used for sequentially encoding the data blocks to be encoded according to the folding sets;

and the judging module (304) judges whether the data block to be coded is the last one, if so, the compressed test vector is output, and if not, the compressed test vector is returned to the coding module (303).

Technical Field

The invention relates to a data coding technology, in particular to a folding set layered compression method and a folding set layered compression device based on forward compatibility.

Background

The feature size of transistors on VLSI (Very Large Scale Integration) chips is decreasing at approximately 10.5% per year, resulting in transistor densities increasing at approximately 22.1% per year, with increasing transistor densities, the number of test vectors required is increasing, a Large amount of test data not only exceeds the capacity of automated test equipment memory and I/O channels, but also extends test time, impacting test cost.

Test data compression methods can be largely divided into coding-based and scan chain-based schemes. The basic idea behind code-based schemes for compressing multidimensional test sets or test vectors is to segment the test data into modules called symbols. Each symbol is then converted to a codeword using a compression algorithm. When in test, the compression is restored through a decompression circuit, and the compression method is a lossless compression method.

Coding-based schemes can be roughly classified into four categories, which are divided according to whether the capacity of symbols and codewords is fixed or not. For example, dictionary coding belongs to fixed-fixed, dictionary indexes are in one-to-one correspondence with data to be coded, the dictionary indexes are covered to all kinds of data to be coded, dictionary data need to be stored, the compression effect is low, and the hardware overhead is high; the Huffman coding belongs to fixed-variable, the Huffman coding uses shorter code words to represent the test data with high occurrence frequency, uses longer code words to represent the test data with low occurrence frequency, the compression effect is improved, but the compression effect for the test data with high occurrence frequency but short data length is limited; the Run _ length coding belongs to variable-fixed, a test vector is divided into symbols with variable length, the symbols consist of continuous 1s or 0s Run and are code words with fixed length, and the coding mode has poor effect on compressing short runs; golomb coding belongs to variable-variable, and variable-length code words are used for representing variable-length runs, so that the adaptability is enhanced, and the compression effect is better.

The test data compression belongs to external self test, has the characteristics of not changing the IP core of the chip, reducing the test data capacity to meet the memory limitation of automatic test equipment, having simple decompression structure and the like, and is a hotspot problem researched by the integrated circuit test

When testing an integrated circuit, test vectors need to be input into ATE automated detection equipment, and whether the circuit has a fault is determined according to test responses generated by the test vectors. Because ATE equipment has limited memory and a large number of test vectors are needed by an integrated circuit, test data needs to be compressed, and compatible compression is an effective encoding compression method, and the encoding rule is simple and easy to restore. However, in the test set, there is often no compatibility between the data blocks divided by the test vector, thereby reducing the compression rate.

The invention patent with application number "CN 201811306241.3" discloses a data compression method, which comprises the following steps: 1) acquiring a vector to be compressed, and segmenting data to be compressed into a plurality of data blocks with equal length; 2) taking the vector to be compressed behind the current reference vector as the current vector to be compressed; 3) judging the compatibility between the data blocks at the corresponding positions; if the reference vector is consistent, setting the coding value of the data block as a first code value, and updating the current reference vector; if the reverse compatibility exists, setting the coding value of the current vector to be compressed as a second code value, and updating the current reference vector; if the current reference vector is incompatible, setting the coding value of the current vector to be compressed as a third code value, and updating the current reference vector; 4) and taking the next vector of the current vector to be compressed as the current vector to be compressed, and returning to execute the step 3) until all the vectors to be compressed are compressed. However, the data blocks in the patent scheme often have no compatibility relationship.

Disclosure of Invention

The technical problem to be solved by the present invention is to provide a folding set layered compression method based on forward compatibility, so as to solve the problem that there is often no compatibility relationship between data blocks divided by test vectors in the background art, thereby reducing the compression rate.

In order to solve the above problems, the present invention provides the following technical solutions:

a folding set layered compression method based on forward compatibility comprises the following steps:

s1, dividing test vectors;

s2, acquiring a data block to be coded and a folding set from the divided test vectors;

s3, coding the data blocks to be coded in sequence according to the folding sets;

s4, judging whether the data block to be coded is the last one, if so, outputting the compressed test vector, and if not, returning to execute the step S3.

The existing test vector is divided, then the data blocks in the test vector are sequentially subjected to compatibility judgment by utilizing the folding set, and the data blocks which do not have the compatibility relation originally are approximately compatible by encoding the data blocks to be encoded, so that the data compression rate is improved.

As a further scheme of the invention: the step S1 includes

Sequentially dividing the test vectors in the test set into k-bit data blocks with equal length, filling and completing the last data block with no relevant bit, wherein the last data block is less than k bits;

wherein k is a positive integer; in the test set, a test vector comprises a plurality of data blocks which are Z1 and Z2 … … Zn in sequence, wherein n is any positive integer;

each data block is a set of 0,1, x, where x is an independent bit.

As a further scheme of the invention: the step S2 includes;

acquiring a first data block Z1 in a test vector, wherein the remaining Z2 … … Zn is a data block to be encoded, the first data block Z1 is recorded as a first reference data block r, the reference data block r is stored in a register RAM1, the register RAM1 directly outputs the data block as it is, the reference vector r output as it is recorded as a folding seed, the folding seed is used as an initial state of a folding counter, the folding counter performs operation on the folding seed, and a folding set F1(x) is generated and stored in the RAM 2;

where, F1(x) { a0, a1, a2, …, ak-1, ak }, and a0 … ak are data blocks.

As a further scheme of the invention: the method for generating the folding set F1(x) is as follows;

obtaining an initial state s, wherein s is a folded seed; the fold counter produces n +1 states s1 from the fold seed s: s1 ═ F (0, s), F (1, s), …, F (n, s), and the set of n +1 states is called a folded set, where the state F (n-1, s) transition to F (n, s) indicates that the code value before n-1 is kept unchanged, F (n, s) indicates the nth folded set generated by folding the seed s, and n is a positive integer.

As a further scheme of the invention: the step S3 includes:

s31, sequentially judging the compatibility of the data block to be coded and the reference data block r, and coding according to the coding mode graph, including;

1) judging the compatibility of the first data block to be coded and the reference vector r, if not compatible, executing step S4, if compatible, executing step 11), and if reversely compatible, executing step 12);

11) outputting '00', and if the determined bit exists in the data block to be encoded and the position of the reference vector r corresponding to the determined bit in the data block to be encoded is the irrelevant bit X, filling the irrelevant bit X of the reference vector r to update the reference vector r, and then executing step S4;

wherein the determined bit is 0 or 1;

wherein, the filling rule is as follows: if the reference vector is compatible with the first database to be coded, filling irrelevant bits in the reference vector with the determined bits; if the reverse compatibility exists, filling irrelevant bits in the reference vector after determining that the bits are inverted;

12) outputting '01', and if the determined bit exists in the data block to be encoded and the position of the reference vector r corresponding to the determined bit in the data block to be encoded is the irrelevant bit X, filling the irrelevant bit X of the reference vector r to update the reference vector r, and executing step S4;

wherein the determined bit is 0 or 1; if the first block of data to be encoded is not compatible with the reference vector r, then step S32 is performed;

s32, judging the compatibility of the data block to be coded and the folding set F1(x), and coding according to the coding mode graph, wherein the coding comprises the steps of;

judging the data block to be coded and a_iWherein i ═ 1, 2, … k; if the data block to be encoded is a_iIf they are compatible, step 21) is executed, if the data block to be encoded is compatible with a_iBackward compatible, step 22) is performed if the data block to be encoded is a_iIf not, executing step S33;

21) the output "100" is added with i-1 expressed by binary, the irrelevant bit X of the reference vector r is filled to update the reference vector r, and then step S4 is executed; i represents a folding distance; where "100" is the prefix and fold set tag combination in this case;

22) outputs "101" plus i-1 in binary representation, fills the irrelevant bit X of the reference vector r to update the reference vector r, and then executes step S4; i represents a folding distance; "101" is the prefix combined with the fold set tag in this case;

s33, outputting 11 plus the data block to be coded; the data block to be encoded is stored as a new reference data block r in the register RAM1, and the reference vector r is stored as a folding seed generation folding set F2(x) in the RAM2, and then step S4 is performed.

As a further scheme of the invention: the data block to be encoded is compatible with the folding set, which means that: if the first data block to be encoded is associated with F1(x) { a }₀，a₁，a₂，…，a_k-1，a_kAny one data block a in₀，a₁，a₂，…，a_k-1，a_kThere is a compatibility relationship.

As a further scheme of the invention: wherein the data block to be encoded and a_iIncompatibility means that the data block to be encoded is not compatible with F1(x) ═ a₀，a₁，a₂，…，a_k-1，a_kAny one data block a in₀，a₁，a₂，…，a_k-1，a_kAre incompatible.

As a further scheme of the invention: data block compatibility means that code values at corresponding positions in any two data blocks are not in conflict, namely 0 and 1 in the data block are in conflict and incompatible, x represents irrelevant bits, namely x and 0 are not in conflict, x and 1 are in conflict, 0 and 0,1 and 1, x and 0, and x and 1 are compatible;

incompatibility refers to the condition other than compatibility, i.e., incompatibility;

reverse compatibility means that code values at corresponding positions in any two data blocks are in conflict, and irrelevant bits are not considered;

the prefix is used for distinguishing the compatibility of the data block to be coded and the reference data block; the folding set flag is used to distinguish the compatibility of the data block to be encoded with the folding set F1(x) generated by the reference data block; suffixes are used to denote the folding distance and the content of the incompatible data block.

As a further scheme of the invention: the step S4, including;

if the input is step S31 or S32, step 31) is performed; if the input is step S33, execute step S32);

31) judging whether the first data block to be coded is the last data block in the folding set F1(x), if so, ending coding to obtain a compressed test vector; if not, continuing to read the next data block as a second data block to be encoded, and returning to execute the step S3;

32) judging whether the first data block to be coded is the last data block in the folding set F1(x), if so, ending coding to obtain a compressed test vector; if not, the folding set F1(x) is discarded, and the procedure continues to step S3 with the block to be encoded as the new reference vector r based on the folding set F2 (x).

A compression device adopting a forward compatibility-based folded set layered compression method comprises;

the dividing module is used for dividing the test vectors;

the dividing module also sequentially divides the test vectors in the test set into data blocks with equal length of k bits, and the last data block is not enough for k bits and is filled and completed by irrelevant bits;

wherein k is a positive integer; in the test set, a test vector comprises a plurality of data blocks which are Z1 and Z2 … … Zn in sequence, wherein n is any positive integer;

each data block is a set of 0,1, x, where x is an independent bit.

The acquisition module is used for acquiring a data block to be encoded and a folding set from the divided test vectors;

the obtaining module further comprises a first data block Z1 in the test vector is obtained, the remaining Z2 … … Zn is a data block to be coded, the first data block Z1 is recorded as a first reference data block r, the reference data block r is stored in a register RAM1, the register RAM1 directly outputs the data block in an original mode, the reference vector r output in the original mode is recorded as a folding seed, the folding seed is used as an initial state of a folding counter, the folding counter performs operation processing on the folding seed, and a folding set F1(x) is generated and stored in a RAM 2;

where, F1(x) { a0, a1, a2, …, ak-1, ak }, and a0 … ak are data blocks.

The method for generating the folding set F (x) is as follows;

obtaining an initial state s, wherein s is a folded seed; the fold counter produces n +1 states s1 from the fold seed s: s1 ═ F (0, s), F (1, s), …, F (n, s), and the set of n +1 states is called a folded set, where the state F (n-1, s) transition to F (n, s) indicates that the code value before n-1 is kept unchanged, F (n, s) indicates the nth folded set generated by folding the seed s, and n is a positive integer.

The encoding module is used for sequentially encoding the data blocks to be encoded according to the folding sets and triggering the step S4;

the encoding module further includes:

s31, sequentially judging the compatibility of the data block to be coded and the reference data block r, and coding according to the coding mode graph, including;

1) judging the compatibility of the first block to be coded and the reference vector r, if not compatible, executing step 11), if compatible, executing step 12), and if reversely compatible, executing step 13);

11) step S4 is executed;

12) outputting '00', and if the determined bit exists in the data block to be encoded and the position of the reference vector r corresponding to the determined bit in the data block to be encoded is the irrelevant bit X, filling the irrelevant bit X of the reference vector r to update the reference vector r, and then executing step S4;

wherein the determined bit is 0 or 1;

wherein, the filling rule is as follows: if the reference vector is compatible with the first database to be coded, filling irrelevant bits in the reference vector with the determined bits; if the reverse compatibility exists, filling irrelevant bits in the reference vector after determining that the bits are inverted;

13) outputting '01', and if the determined bit exists in the data block to be encoded and the position of the reference vector r corresponding to the determined bit in the data block to be encoded is the irrelevant bit X, filling the irrelevant bit X of the reference vector r to update the reference vector r, and executing step S4;

wherein the determined bit is 0 or 1; if the first block of data to be encoded is not compatible with the reference vector r, then step S32 is performed;

s32, judging the compatibility of the data block to be coded and the folding set F1(x), and coding according to the coding mode graph, wherein the coding comprises the steps of;

judging the data block to be coded and a_iWherein i ═ 1, 2, … k; if the data block to be encoded is the same asa_iIf they are compatible, step 21) is executed, if the data block to be encoded is compatible with a_iBackward compatible, step 22) is performed if the data block to be encoded is a_iIncompatible, execute step 23);

21) the output "100" is added with i-1 expressed by binary, the irrelevant bit X of the reference vector r is filled to update the reference vector r, and then step S4 is executed; i represents a folding distance; where "100" is the prefix and fold set tag combination in this case;

22) outputs "101" plus i-1 in binary representation, fills the irrelevant bit X of the reference vector r to update the reference vector r, and then executes step S4; i represents a folding distance; "101" is the prefix combined with the fold set tag in this case;

23) step S33 is executed.

S33, outputting 11 plus the data block to be coded; the data block to be encoded is stored as a new reference data block r in the register RAM1, and the reference vector r is stored as a folding seed generation folding set F2(x) in the RAM2, and then step S4 is performed.

The judging module is used for judging whether the data block to be coded is the last one, outputting the compressed test vector if the data block to be coded is the last one, and returning to the coding module if the data block to be coded is not the last one;

the judging module also comprises;

if the input is step S31 or S32, step 31) is performed; if the input is step S33, execute step S32);

31) judging whether the first data block to be coded is the last data block in the folding set F1(x), if so, ending coding to obtain a compressed test vector; if not, continuing to read the next data block as a second data block to be encoded, and returning to execute the step S3;

32) judging whether the first data block to be coded is the last data block in the folding set F1(x), if so, ending coding to obtain a compressed test vector; if not, the folding set F1(x) is discarded, and the procedure continues to step S3 with the block to be encoded as the new reference vector r based on the folding set F2 (x).

Compared with the prior art, the invention has the beneficial effects that:

1. the method comprises the steps of dividing test vectors, generating a folding set by using reference vectors as folding seeds, and sequentially carrying out compatibility judgment on remaining blocks to be coded and the reference vectors, so that the folding set is used for carrying out compatibility compression, when the reference vectors are determined, the reference vectors are simultaneously used as the seed vectors to generate the folding set, if the length of the reference vectors is k, the generated folding set comprises k +1 vectors, each vector in the folding set can be used as the reference vector of a block to be coded, so that the number of the reference vectors is greatly expanded, and the same block to be coded is compatible with a plurality of reference vectors; therefore, compatibility is generated between data blocks which do not have compatibility relation originally, and the compression rate is improved.

2. When a data block to be coded is incompatible with a reference vector r and a folding set F1(x), the data block to be coded is used as a new reference vector, and the reference vector is used as a folding seed to generate a folding set F2(x) and is stored in a RAM 2; on the premise of not influencing the original compression ratio, the new reference vector and the folded set F2(X) have more irrelevant bits X, meanwhile, the number of the reference vectors is expanded by using the folded set and a similar method, the compatibility of the data block to be coded and the reference data block is compared firstly, and then the hierarchical compression of the compatibility of the data block to be coded and the folded set is compared, so that the compatibility of the data block to be coded and the new reference vector and the folded set is greatly improved, and the accuracy is improved.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention.

Fig. 1 is a block diagram illustrating a forward consistent-based folded set hierarchical compression method according to embodiment 1 of the present invention.

Fig. 2 is a schematic structural diagram of a forward compatible based folded ensemble layered compression apparatus according to embodiment 1 of the present invention.

Fig. 3 is a diagram of an encoding method of a forward compatible-based folded set layered compression method according to embodiment 1 of the present invention.

Detailed Description

In order to make the technical problems, technical solutions and advantageous effects to be solved by the present invention more clearly apparent, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.