Patent Application
20230195945
The present application claims the priority of Chinese Patent Application No. 202111571211.7, filed on Dec. 21, 2021, with the title of “INTEGRATED CIRCUIT CHIP VERIFICATION METHOD AND APPARATUS, ELECTRONIC DEVICE, AND STORAGE MEDIUM.” The disclosure of the above application is incorporated herein by reference in its entirety.
The present disclosure relates to the field of artificial intelligence technologies, and in particular, to an integrated circuit (IC) chip verification method and apparatus, an electronic device, and a storage medium in fields such as artificial intelligence chips and cloud computing.
Randomized data excitation is very important for verification of IC chips, which may be voice chips or the like.
At present, mainstream IC verification platforms are generally built based on system verilog. The system verilog refers to an SV language, which is a hardware description and verification language. However, a system verilog platform only provides randomization for fixed-point (integer) data, not for floating-point data.
The present disclosure provides an IC chip verification method and apparatus, an electronic device, and a storage medium.
A method for IC chip verification, including acquiring a first parameter and a second parameter, the first parameter being an upper limit of a required floating-point number, and the second parameter being a lower limit of the required floating-point number; generating a first sign, a first exponent, and a first fraction of a randomized floating-point number respectively according to the first parameter and the second parameter; generating the floating-point number according to the first sign, the first exponent, and the first fraction; and performing IC chip verification by using the floating-point number.
An electronic device, including at least one processor; and a memory communicatively connected with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform a method for integrated circuit (IC) chip verification, wherein the method includes acquiring a first parameter and a second parameter, the first parameter being an upper limit of a required floating-point number, and the second parameter being a lower limit of the required floating-point number; generating a first sign, a first exponent, and a first fraction of a randomized floating-point number respectively according to the first parameter and the second parameter; generating the floating-point number according to the first sign, the first exponent, and the first fraction; and performing IC chip verification by using the floating-point number.
A non-transitory computer readable storage medium with computer instructions stored thereon, wherein the computer instructions are used for causing a method for integrated circuit (IC) chip verification, wherein the method includes acquiring a first parameter and a second parameter, the first parameter being an upper limit of a required floating-point number, and the second parameter being a lower limit of the required floating-point number; generating a first sign, a first exponent, and a first fraction of a randomized floating-point number respectively according to the first parameter and the second parameter; generating the floating-point number according to the first sign, the first exponent, and the first fraction; and performing IC chip verification by using the floating-point number.
It should be understood that the content described in this part is neither intended to identify key or significant features of the embodiments of the present disclosure, nor intended to limit the scope of the present disclosure. Other features of the present disclosure will be made easier to understand through the following description.
The accompanying drawings are intended to provide a better understanding of the solutions and do not constitute a limitation on the present disclosure. In the drawings,
Exemplary embodiments of the present disclosure are illustrated below with reference to the accompanying drawings, which include various details of the present disclosure to facilitate understanding and should be considered only as exemplary. Therefore, those of ordinary skill in the art should be aware that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for clarity and simplicity, descriptions of well-known functions and structures are omitted in the following description.
In addition, it should be understood that the term “and/or” herein is merely an association relationship describing associated objects, indicating that three relationships may exist. For example, A and/or B indicates that there are three cases of A alone, A and B together, and B alone. In addition, the character “/” herein generally means that associated objects before and after it are in an “or” relationship.
In step 101, a first parameter and a second parameter are acquired, the first parameter being an upper limit of a required floating-point number, and the second parameter being a lower limit of the required floating-point number.
In step 102, a first sign, a first exponent, and a first fraction of a randomized floating-point number are generated respectively according to the first parameter and the second parameter.
In step 103, the floating-point number is generated according to the first sign, the first exponent, and the first fraction.
In step 104, IC chip verification is performed by using the floating-point number.
As can be seen, by use of the solution of the above method embodiment, randomization of floating-point data can be realized, thereby providing required floating-point data excitation for the verification of relevant operation units in IC chip verification and improving a verification effect.
The floating-point data may include a floating-point number and a floating-point number vector. The floating-point number may consist of a sign, an exponent, and a fraction.
In order to be distinguished from other signs, exponents, and fractions that appear later, a sign, an exponent, and a fraction of a randomized floating-point number are referred to as a first sign, a first exponent, and a first fraction respectively. Subsequent similar situations are not described in detail.
In the solution of the present disclosure, an obtained floating-point number is required to be in a data range constrained. That is, a user may set that there is a need to constrain a value range of the floating-point number generated by randomization, and finally return a floating-point number that meets a constraint condition.
Accordingly, a first parameter and a second parameter may be set. The first parameter is an upper limit of a required floating-point number, and the second parameter is a lower limit of the required floating-point number.
In the manner of the present disclosure, a first sign, a first exponent, and a first fraction of a randomized floating-point number may be generated respectively according to the first parameter and the second parameter. In one embodiment of the present disclosure, a second sign, a second exponent, and a second fraction of the first parameter may be acquired respectively, a third sign, a third exponent, and a third fraction of the second parameter may be acquired respectively, and then the first sign, the first exponent, and the first fraction of the floating-point number may be generated according to the second sign, the second exponent, the second fraction, the third sign, the third exponent, and the third fraction acquired.
The second sign, the second exponent, and the second fraction of the first parameter may be denoted by sign_max, exp_max, and frac_max respectively, the third sign, the third exponent, and the third fraction of the second parameter may be denoted by sign_min, exp_min, and frac_min respectively, and the first sign, the first exponent, and the first fraction of the floating-point number may be denoted by sign_result, exp_result, and frac_result.
In one embodiment of the present disclosure, the first sign, the first exponent, and the first fraction may be generated in a first generation manner if the second sign and the third sign are equal to 0, the first sign, the first exponent, and the first fraction may be generated in a second generation manner if the second sign is equal to 0 and the third sign are equal to 1, and the first sign, the first exponent, and the first fraction may be generated in a third generation manner if the second sign is equal to 1 and the third sign are equal to 1.
In other words, the second sign and the third sign may be compared, and a corresponding generation manner may be used to generate the first sign, the first exponent, and the first fraction according to different comparison results, so that the generation is more targeted, thereby improving accuracy of a generation result.
If sign_max==1 and sign_min==0, indicating that the first parameter has a negative sign and the second parameter has a positive sign, a floating-point number that meets the constraint condition cannot be obtained in this case. Accordingly, report an error and exit. If sign_max==0 and sign_min==0, the first sign, the first exponent, and the first fraction may be generated in the first generation manner. If sign_max==0 and sign_min==1, the first sign, the first exponent, and the first fraction may be generated in the second generation manner. If sign_max==1 and sign_min==1, the first sign, the first exponent, and the first fraction may be generated in the third generation manner.
Specification implementations of the first generation manner, the second generation manner, and the third generation manner are introduced below respectively.
In one embodiment of the present disclosure, the generating, by the first generation module, the first sign, the first exponent, and the first fraction in a first generation manner may include: taking 0 as the first sign, the second exponent or the third exponent as the first exponent, and a random value in a range of [the third fraction: the second fraction] as the first fraction if the second exponent is equal to the third exponent and the second fraction is greater than or equal to the third fraction.
In addition, in one embodiment of the present disclosure, the generating, by the first generation module, the first sign, the first exponent, and the first fraction in a first generation manner may further include: generating a random value in a range of [the third exponent: the second exponent] as a first intermediate variable if the second exponent is greater than the third exponent; taking 0 as the first sign, the third exponent as the first exponent, and a random value in a range of [the third fraction: a first predetermined value] as the first fraction if the first intermediate variable is equal to the third exponent; taking 0 as the first sign, the second exponent as the first exponent, and a random value in a range of [a second predetermined value: the second fraction] as the first fraction if the first intermediate variable is equal to the second exponent; and taking 0 as the first sign, the first intermediate variable as the first exponent, and a random value in a range of [the second predetermined value: the first predetermined value] as the first fraction if the first intermediate variable is equal to neither the third exponent nor the second exponent.
Based on the above introduction,
In step 201, exp_max is compared with exp_min, if exp_max<exp_min, step 202 is performed, if exp_max==exp_min, step 203 is performed, and if exp_max>exp_min, step 205 is performed.
In step 202, report an error and exit, and then end the process.
When sign_max==0 and sign_min==0, indicating that both the first parameter and the second parameter have a positive sign, if exp_max<exp_min, indicating that the exponent of the first parameter is less than that of the second parameter, a floating-point number that meets the constraint condition cannot be obtained in this case. Accordingly, report an error and exit.
In step 203, it is determined whether frac_max<frac_min, if yes, step 202 is performed, and otherwise, step 204 is performed.
When sign_max==0, sign_min==0, and exp_max==exp_min, if frac_max<frac_min, indicating that both the first parameter and the second parameter have a positive sign and the exponent of the first parameter is equal to that of the second parameter, but the fraction of the first parameter is less than that of the second parameter, a floating-point number that meets the constraint condition cannot be obtained in this case. Accordingly, report an error and exit.
In step 204, 0 is taken as sign_result, exp_max or exp_min is taken as exp_result, a random value in a range of [frac_min:frac_max] is taken as frac_result, and then the process is ended.
That is, sign_result=0;
exp_result=exp_max or exp_min;
frac_result=random([frac_min:frac_max]).
The random value in the range of [frac_min:frac_max] may be generated as frac_result by using a randomization method provided by a system verilog platform.
In step 205, a random value in a range of [exp_min:exp_max] is generated as a first intermediate variable tmp1.
The random value in the range of [exp_min:exp_max] may be generated as the first intermediate variable tmp1 by using the randomization method provided by the system verilog platform. That is, tmp1=random([exp_min:exp_max]).
In step 206, tmp1 is compared with exp_min and exp_max, if tmp1==exp_min, step 207 is performed, if tmp1==exp_max, step 208 is performed, and otherwise, step 209 is performed.
In step 207, 0 is taken as sign_result, exp_min is taken as exp_result, a random value in a range of [frac_min:a first predetermined value] is taken as frac_result, and then the process is ended.
A specific value of the first predetermined value may be determined according to an actual requirement, for example, ′h7F_FFFF.
Then, sign_result=0;
exp_result=exp_min;
frac_result=random([frac_min:′h7F_FFFF]).
The random value in the range of [frac_min:′h7F_FFFF] may be generated as frac_result by using the randomization method provided by the system verilog platform.
In step 208, 0 is taken as sign_result, exp_max is taken as exp_result, a random value in a range of [a second predetermined value:frac_max] is taken as frac_result, and then the process is ended.
A specific value of the second predetermined value may be determined according to an actual requirement, for example, ′h0.
Then, sign_result=0;
exp_result=exp_max;
frac_result=random([′h0:frac_max]).
The random value in the range of [′h0:frac_max] may be generated as frac_result by using the randomization method provided by the system verilog platform.
In step 209, 0 is taken as sign_result, tmp1 is taken as exp_result, a random value in a range of [the second predetermined value:the first predetermined value] is taken as frac_result, and then the process is ended.
That is, sign_result=0;
exp_result=tmp1;
frac_result=random([′h0:′h7F_FFFF]).
The random value in the range of [′h0:′h7F_FFFF] may be generated as frac_result by using the randomization method provided by the system verilog platform.
After the first sign, the first exponent, and the first fraction are acquired, a required floating-point number may be generated according to such information. That is, a randomized floating-point number that meets a constraint condition may be generated according to the acquired sign_result, exp_result, and frac_result. The generation manner is a related art.
In one embodiment of the present disclosure, the generating, by the first generation module, the first sign, the first exponent, and the first fraction in a second generation manner may include: generating a random value in a range of [0:1] as a second intermediate variable; generating a random value in a range of [0:the second exponent] as a third intermediate variable if the second intermediate variable is equal to 0; taking 0 as the first sign, the second exponent as the first exponent, and a random value in a range of [a second predetermined value:the second fraction] as the first fraction if the third intermediate variable is equal to the second exponent; and taking 0 as the first sign, the third intermediate variable as the first exponent, and a random value in a range of [the second predetermined value:a first predetermined value] as the first fraction if the third intermediate variable is not equal to the second exponent.
In addition, in one embodiment of the present disclosure, the generating, by the first generation module, the first sign, the first exponent, and the first fraction in a second generation manner may further include: generating a random value in a range of [0:the third exponent] as a fourth intermediate variable if the second intermediate variable is not equal to 0; taking 1 as the first sign, the third exponent as the first exponent, and a random value in a range of [a second predetermined value:the third fraction] as the first fraction if the fourth intermediate variable is equal to the third exponent; and taking 1 as the first sign, the fourth intermediate variable as the first exponent, and a random value in a range of [the second predetermined value:the first predetermined value] as the first fraction if the fourth intermediate variable is not equal to the third exponent.
Based on the above introduction,
In step 301, a random value in a range of [0:1] is generated as a second intermediate variable tmp2.
When sign_max==0 and sign_min==1, indicating that the first parameter has a positive sign and the second parameter has a negative sign, firstly, the random value in the range of [0:1] may be generated as a second intermediate variable tmp2 by using the randomization method provided by the system verilog platform. That is, tmp2=random([0:1]).
In step 302, it is determined whether tmp2==0, if yes, step 303 is performed, and otherwise, step 307 is performed.
In step 303, a random value in a range of [0:exp_max] is generated as a third intermediate variable tmp3.
That is, tmp3=random([0:exp_max]).
The random value in the range of [0:exp_max] may be generated by using the randomization method provided by the system verilog platform.
In step 304, it is determined whether tmp3==exp_max, if yes, step 305 is performed, and otherwise, step 306 is performed.
In step 305, 0 is taken as sign_result, exp_max is taken as exp_result, a random value in a range of [a second predetermined value:frac_max] is taken as frac_result, and then the process is ended.
That is, sign_result=0;
exp_result=exp_max;
frac_result=random([′h0:frac_max]).
The random value in the range of [′h0:frac_max] may be generated by using the randomization method provided by the system verilog platform.
In step 306, 0 is taken as sign_result, tmp3 is taken as exp_result, a random value in a range of [the second predetermined value:the first predetermined value] is taken as frac_result, and then the process is ended.
That is, sign_result=0;
exp_result=tmp3;
frac_result=random([′h0:′h7F_FFFF]).
The random value in the range of [′h0:′h7F_FFFF] may be generated by using the randomization method provided by the system verilog platform.
In step 307, a random value in a range of [0:exp_min] is generated as a fourth intermediate variable tmp4.
The random value in the range of [0:exp_min] may be generated by using the randomization method provided by the system verilog platform. That is, exp_tmp=random([0 exp_min]).
In step 308, it is determined whether tmp4==exp_min, if yes, step 309 is performed, and otherwise, step 310 is performed.
In step 309, 1 is taken as sign_result, exp_min is taken as exp_result, a random value in a range of [the second predetermined value:frac_min] is taken as frac_result, and then the process is ended.
That is, sign_result=1;
exp_result=exp_min;
frac_result=random([′h0:frac_min]).
The random value in the range of [′h0:frac_min] may be generated by using the randomization method provided by the system verilog platform.
In step 310, 1 is taken as sign_result, tmp4 is taken as exp_result, a random value in a range of [the second predetermined value:the first predetermined value] is taken as frac_result, and then the process is ended.
That is, sign_result=1;
exp_result=tmp4;
frac_result=random([′h0:′h7F_FFFF]).
The random value in the range of [′h0:′h7F_FFFF] may be generated by using the randomization method provided by the system verilog platform.
After the first sign, the first exponent, and the first fraction are acquired, a required floating-point number may be generated according to such information. That is, a randomized floating-point number that meets a constraint condition may be generated according to the acquired sign_result, exp_result, and frac_result.
In one embodiment of the present disclosure, the generating, by the first generation module, the first sign, the first exponent, and the first fraction in a third generation manner may include: taking 1 as the first sign, the second exponent or the third exponent as the first exponent, and a random value in a range of [the second fraction:the third fraction] as the first fraction if the second exponent is equal to the third exponent and the second fraction is less than or equal to the third fraction.
In addition, in one embodiment of the present disclosure, the generating, by the first generation module, the first sign, the first exponent, and the first fraction in a third generation manner may further include: generating a random value in a range of [the second exponent:the third exponent] as a fifth intermediate variable if the second exponent is greater than the third exponent; taking 1 as the first sign, the third exponent as the first exponent, and a random value in a range of [a second predetermined value:the third fraction] as the first fraction if the fifth intermediate variable is equal to the third exponent; taking 1 as the first sign, the second exponent as the first exponent, and a random value in a range of [the second fraction:a first predetermined value] as the first fraction if the fifth intermediate variable is equal to the second exponent; and taking 1 as the first sign, the fifth intermediate variable as the first exponent, and a random value in a range of [the second predetermined value:the first predetermined value] as the first fraction if the fifth intermediate variable is equal to neither the third exponent nor the second exponent.
Based on the above introduction,
In step 401, exp_max is compared with exp_min, if exp_max>exp_min, step 402 is performed, if exp_max==exp_min, step 403 is performed, and if exp_max<exp_min, step 405 is performed.
In step 402, report an error and exit, and then end the process.
When sign_max==1 and sign_min==1, indicating that both the first parameter and the second parameter have a negative sign, if exp_max>exp_min, indicating that the exponent of the first parameter is greater than that of the second parameter, a floating-point number that meets the constraint condition cannot be obtained in this case. Accordingly, report an error and exit.
In step 403, it is determined whether frac_max>frac_min, if yes, step 402 is performed, and otherwise, step 404 is performed.
When sign_max==1, sign_min==1, and exp_max==exp_min, if frac_max>frac_min, indicating that both the first parameter and the second parameter have a negative sign and the exponent of the first parameter is equal to that of the second parameter, but the fraction of the first parameter is greater than that of the second parameter, a floating-point number that meets the constraint condition cannot be obtained in this case. Accordingly, report an error and exit.
In step 404, 1 is taken as sign_result, exp_max or exp_min is taken as exp_result, a random value in a range of [frac_max:frac_min] is taken as frac_result, and then the process is ended.
That is, sign_result=1;
exp_result=exp_max or exp_min;
frac_result=random([frac_max:frac_min]).
The random value in the range of [frac_max:frac_min] may be generated by using the randomization method provided by the system verilog platform.
In step 405, a random value in a range of [exp_max:exp_min] is generated as a fifth intermediate variable tmp5.
The random value in the range of [exp_max:exp_min] may be generated as the fifth intermediate variable tmp5 by using the randomization method provided by the system verilog platform. That is, tmp5=random([exp_max:exp_min]).
In step 406, tmp5 is compared with exp_min and exp_max, if tmp5==exp_min, step 407 is performed, if tmp5==exp_max, step 408 is performed, and otherwise, step 409 is performed.
In step 407, 1 is taken as sign_result, exp_min is taken as exp_result, a random value in a range of [a second predetermined value:frac_min] is taken as frac_result, and then the process is ended.
That is, sign_result=1;
exp_result=exp_min;
frac_result=random([′h0:frac_min]).
The random value in the range of [′h0:frac_min] may be generated by using the randomization method provided by the system verilog platform.
In step 408, 1 is taken as sign_result, exp_max is taken as exp_result, a random value in a range of [frac_max:a first predetermined value] is taken as frac_result, and then the process is ended.
That is, sign_result=1;
exp_result=exp_max;
frac_result=random([frac_max:′h7F_FFFF]).
The random value in the range of [frac_max:′h7F_FFFF] may be generated by using the randomization method provided by the system verilog platform.
In step 409, 1 is taken as sign_result, tmp5 is taken as exp_result, a random value in a range of [the second predetermined value: the first predetermined value] is taken as frac_result, and then the process is ended.
That is, sign_result=1;
exp_result=tmp5;
frac_result=random([′h0:′h7F_FFFF]).
The random value in the range of [′h0:′h7F_FFFF] may be generated as frac_result by using the randomization method provided by the system verilog platform.
After the first sign, the first exponent, and the first fraction are acquired, a required floating-point number may be generated according to such information. That is, a randomized floating-point number that meets a constraint condition may be generated according to the acquired sign_result, exp_result, and frac_result.
In either manner, randomization of floating-point data can be realized, thereby providing required floating-point data excitation for the verification of relevant operation units in IC chip verification and improving a verification effect.
In the solutions of the present disclosure, in addition to the floating-point number, a floating-point number vector can also be generated. Accordingly, the floating-point number vector can also be used for IC chip verification. In one embodiment of the present disclosure, a randomized floating-point number vector of a specified length may be generated according to the floating-point number, thereby enriching generated floating-point data types and further improving the verification effect.
The user may set a length of a floating-point number vector generated by constrained randomization, and additionally, may also specify a type of a linear operation for randomization constraint. The linear operation has to be an operation on an array element supported by the system verilog platform. Moreover, after the linear operation on a data unit in the floating-point number vector, an operation result is required to be within a specified data range. That is, the operation result is required to be between an upper limit and a lower limit that are specified. A floating-point number vector that meets a constraint condition is finally returned. The number of the data unit is len, where len represents the specified length.
Firstly, a randomized floating-point number may be generated in the manner according to the present disclosure, in which a sign, an exponent, and a fraction are referred to as a first sign, a first exponent, and a first fraction respectively. Then, a randomized floating-point number vector may be generated according to the floating-point number.
In one embodiment of the present disclosure, the randomized floating-point number vector may be generated according to the floating-point number in the following manner: acquiring a product of the floating-point number with a predetermined first factor to obtain a sixth intermediate variable if the first exponent is equal to 0; acquiring a first array of randomized integer data, a length of the first array being equal to the specified length, and a result obtained after a predetermined linear operation on the first array being the sixth intermediate variable; and evenly distributing the first factor among elements in the first array, and determining the floating-point number vector in combination with a type of the linear operation.
In one embodiment of the present disclosure, the elements in the first array may be divided by the first factor to obtain the floating-point number vector if the linear operation is an addition operation. M elements may be randomly selected from the first array, the selected elements may be divided by 2, the M elements divided by 2 are used to form a second array, and elements in the second array may be divided by 2N to obtain the floating-point number vector if the linear operation is a multiplication operation, wherein N is a quotient obtained by dividing a third predetermined value by the specified length, and M is a remainder obtained by dividing the third predetermined value by the specified length.
In one embodiment of the present disclosure, the first factor may be 2149, and the third predetermined value may be 149.
That is, the first factor is factor 1=(2{circumflex over ( )}126)*(2 {circumflex over ( )}23)=2{circumflex over ( )}149;
tmp6=F_tmp*factor1;
where F_tmp denotes the floating-point number, and tmp6 denotes the sixth intermediate variable.
Then, −(2{circumflex over ( )}23)<tmp6<(2{circumflex over ( )}23), and data fractions of tmp6 are all 0.
Then, a first array Int_V of a piece of integer data may be obtained by system verilog array randomization, so that randomize(Int_V) with {Length(Int_V)==len; F(Int_V)==tmp6}, that is, a length of the first array is equal to the specified length len, and a result after a predetermined linear operation F( ) on the first array is the sixth intermediate variable tmp6.
Further, according to a specific operation process of the linear operation, factor1 may be evenly distributed among the elements in the first array. The elements in the first array may be divided by factor1 to obtain the required floating-point number vector if the linear operation is an addition operation. M elements may be randomly selected from the first array, the selected elements may be divided by 2, the M elements divided by 2 are used to form a second array, and elements in the second array may be divided by 2N to obtain the required floating-point number vector if the linear operation is a multiplication operation.
The linear operation is generally an addition operation or a multiplication operation. The multiplication operation generally refers to scalar multiplication.
When the linear operation is the addition operation,
F_V=Int_V/factor1;
where F_V denotes the obtained floating-point number vector.
When the linear operation is the multiplication operation, firstly, 149 may be divided by len to obtain a quotient N and a remainder M, that is, 149=(len*N)+M. Then, with the randomization method provided by the system verilog platform, M elements may be randomly selected from the first array, and each element is divided by 2, so as to obtain a second array F_V_pre. After that, each element in F_V_pre is divided by 2{circumflex over ( )}N, so as to obtain the required floating-point number vector, that is, F_V=F_V_pre/2{circumflex over ( )}N.
In one embodiment of the present disclosure, the randomized floating-point number vector may also be generated according to the floating-point number in the following manner: determining a second factor according to the first exponent, and acquiring a product of the floating-point number with the second factor to obtain a seventh intermediate variable if the first exponent is not equal to 0; acquiring a third array of the randomized integer data, a length of the third array being equal to the specified length, and a result obtained after a predetermined linear operation on the third array being the seventh intermediate variable; and evenly distributing the second factor among elements in the third array, and determining the floating-point number vector in combination with a type of the linear operation.
In one embodiment of the present disclosure, the elements in the third array may be divided by the second factor to obtain the floating-point number vector if the linear operation is an addition operation. M′ elements may be randomly selected from the third array, the selected elements may be divided by 2, the M′ elements divided by 2 are used to form a fourth array, and elements in the fourth array may be divided by 2N′ to obtain the floating-point number vector if the linear operation is a multiplication operation; wherein N′ is a quotient obtained by dividing a fourth predetermined value by the specified length, and M′ is a remainder obtained by dividing the fourth predetermined value by the specified length.
In one embodiment of the present disclosure, the second factor may be 2P, where P is the fourth predetermined value and a difference between 149 and the first exponent.
That is, the second factor is factor 2=(2{circumflex over ( )}(127−exp_tmp))*(2{circumflex over ( )} 23)=2{circumflex over ( )}(149−exp_tmp);
tmp7=F_tmp*factor2;
where F_tmp denotes the floating-point number, and tmp7 denotes the seventh intermediate variable.
Then, −(2{circumflex over ( )}24)<Int_tmp<(2{circumflex over ( )} 24), and data fractions of tmp7 are all 0.
Then, a third array Int_V′ of a piece of integer data may be obtained by system verilog array randomization, so that randomize (Int_V′) with {Length(Int_V′)==len; F(Int_V′)==tmp7}, that is, a length of the third array is equal to the specified length len, and a result after a predetermined linear operation F( ) on the third array is the seventh intermediate variable tmp7.
Further, according to a specific operation process of the linear operation, factor2 may be evenly distributed among the elements in the third array. The elements in the third array may be divided by factor2 to obtain the required floating-point number vector if the linear operation is an addition operation. M′ elements may be randomly selected from the third array, the selected elements may be divided by 2, the M′ elements divided by 2 are used to form a fourth array, and elements in the fourth array may be divided by 2N′ to obtain the required floating-point number vector if the linear operation is a multiplication operation.
When the linear operation is the addition operation,
F_V=Int_V′/factor2;
where F_V denotes the obtained floating-point number vector.
When the linear operation is the multiplication operation, firstly, (149-first exponent) may be divided by len to obtain a quotient N′ and a remainder M′, that is, (149-first exponent)=(len*N′)+M′. Then, with the randomization method provided by the system verilog platform, M′ elements may be randomly selected from the third array, and each element is divided by 2, so as to obtain a fourth array F_V_pre′. After that, each element in F_V_pre′ is divided by 2{circumflex over ( )}N′, so as to obtain the required floating-point number vector, that is, F_V=F_V_pre′/2{circumflex over ( )}N′.
It is to be noted that, to make the description brief, the foregoing method embodiments are expressed as a series of actions. However, those skilled in the art should appreciate that the present disclosure is not limited to the described action sequence, because according to the present disclosure, some steps may be performed in other sequences or performed simultaneously. Next, those skilled in the art should also appreciate that all the embodiments described in the specification are preferred embodiments, and the related actions and modules are not necessarily mandatory to the present disclosure. In addition, for a part that is not described in detail in one embodiment, refer to related descriptions in other embodiments.
In brief, by use of the solution in the method embodiment of the present disclosure, a floating-point number and a floating-point number vector randomized can be obtained, so as to make up for the deficiency that the current system Verilog platform provides randomization only for fixed-point data. Moreover, the user is only required to perform some simple settings, such as setting some constraint conditions, and other processing can be completed automatically, thereby simplifying user operations. In addition, the floating-point number and the floating-point number vector obtained can be used to verify the relevant operation units in IC chip verification, such as a floating-point operation unit and a floating-point vector operation unit, and a verification convergence process of the relevant operation units can be accelerated.
The above is the introduction to the method embodiments. The following is a further illustration of the solutions of the present disclosure through apparatus embodiments.
The parameter acquisition module 501 is configured to acquire a first parameter and a second parameter, the first parameter being an upper limit of a required floating-point number, and the second parameter being a lower limit of the required floating-point number.
The first generation module 502 is configured to generate a first sign, a first exponent, and a first fraction of a randomized floating-point number respectively according to the first parameter and the second parameter.
The second generation module 503 is configured to generate the floating-point number according to the first sign, the first exponent, and the first fraction.
The verification module 504 is configured to perform IC chip verification by using the floating-point number.
By use of the solution of the above apparatus embodiment, randomization of floating-point data can be realized, thereby providing required floating-point data excitation for the verification of relevant operation units in IC chip verification and improving a verification effect.
The floating-point data may include a floating-point number and a floating-point number vector. The floating-point number may consist of a sign, an exponent, and a fraction.
In one embodiment of the present disclosure, the first generation module 502 may acquire a second sign, a second exponent, and a second fraction of the first parameter respectively, acquire a third sign, a third exponent, and a third fraction of the second parameter respectively, and then generate the first sign, the first exponent, and the first fraction of the floating-point number according to the second sign, the second exponent, the second fraction, the third sign, the third exponent, and the third fraction acquired.
In one embodiment of the present disclosure, the first generation module 502 may generate the first sign, the first exponent, and the first fraction in a first generation manner if the second sign and the third sign are equal to 0, generate the first sign, the first exponent, and the first fraction in a second generation manner if the second sign is equal to 0 and the third sign are equal to 1, and generate the first sign, the first exponent, and the first fraction in a third generation manner if the second sign is equal to 1 and the third sign are equal to 1.
In one embodiment of the present disclosure, the generating, by the first generation module 502, the first sign, the first exponent, and the first fraction in a first generation manner may include: taking 0 as the first sign, the second exponent or the third exponent as the first exponent, and a random value in a range of [the third fraction: the second fraction] as the first fraction if the second exponent is equal to the third exponent and the second fraction is greater than or equal to the third fraction.
In one embodiment of the present disclosure, the generating, by the first generation module 502, the first sign, the first exponent, and the first fraction in a first generation manner may further include: generating a random value in a range of [the third exponent: the second exponent] as a first intermediate variable if the second exponent is greater than the third exponent; taking 0 as the first sign, the third exponent as the first exponent, and a random value in a range of [the third fraction: a first predetermined value] as the first fraction if the first intermediate variable is equal to the third exponent; taking 0 as the first sign, the second exponent as the first exponent, and a random value in a range of [a second predetermined value: the second fraction] as the first fraction if the first intermediate variable is equal to the second exponent; and taking 0 as the first sign, the first intermediate variable as the first exponent, and a random value in a range of [the second predetermined value: the first predetermined value] as the first fraction if the first intermediate variable is equal to neither the third exponent nor the second exponent.
In one embodiment of the present disclosure, the generating, by the first generation module 502, the first sign, the first exponent, and the first fraction in a second generation manner may include: generating a random value in a range of [0:1] as a second intermediate variable; generating a random value in a range of [0: the second exponent] as a third intermediate variable if the second intermediate variable is equal to 0; taking 0 as the first sign, the second exponent as the first exponent, and a random value in a range of [a second predetermined value: the second fraction] as the first fraction if the third intermediate variable is equal to the second exponent; and taking 0 as the first sign, the third intermediate variable as the first exponent, and a random value in a range of [the second predetermined value: a first predetermined value] as the first fraction if the third intermediate variable is not equal to the second exponent.
In one embodiment of the present disclosure, the generating, by the first generation module 502, the first sign, the first exponent, and the first fraction in a second generation manner may further include: generating a random value in a range of [0: the third exponent] as a fourth intermediate variable if the second intermediate variable is not equal to 0; taking 1 as the first sign, the third exponent as the first exponent, and a random value in a range of [a second predetermined value: the third fraction] as the first fraction if the fourth intermediate variable is equal to the third exponent; and taking 1 as the first sign, the fourth intermediate variable as the first exponent, and a random value in a range of [the second predetermined value: the first predetermined value] as the first fraction if the fourth intermediate variable is not equal to the third exponent.
In one embodiment of the present disclosure, the generating, by the first generation module 502, the first sign, the first exponent, and the first fraction in a third generation manner may include: taking 1 as the first sign, the second exponent or the third exponent as the first exponent, and a random value in a range of [the second fraction: the third fraction] as the first fraction if the second exponent is equal to the third exponent and the second fraction is less than or equal to the third fraction.
In one embodiment of the present disclosure, the generating, by the first generation module 502, the first sign, the first exponent, and the first fraction in a third generation manner may further include: generating a random value in a range of [the second exponent: the third exponent] as a fifth intermediate variable if the second exponent is greater than the third exponent; taking 1 as the first sign, the third exponent as the first exponent, and a random value in a range of [a second predetermined value: the third fraction] as the first fraction if the fifth intermediate variable is equal to the third exponent; taking 1 as the first sign, the second exponent as the first exponent, and a random value in a range of [the second fraction: a first predetermined value] as the first fraction if the fifth intermediate variable is equal to the second exponent; and taking 1 as the first sign, the fifth intermediate variable as the first exponent, and a random value in a range of [the second predetermined value: the first predetermined value] as the first fraction if the fifth intermediate variable is equal to neither the third exponent nor the second exponent.
The parameter acquisition module 501, the first generation module 502, and the second generation module 503 are the same as those in the embodiment shown in
The third generation module 505 is configured to generate a randomized floating-point number vector of a specified length according to the floating-point number. Correspondingly, the verification module 504 may be further configured to perform IC chip verification by using the floating-point number vector.
In one embodiment of the present disclosure, the third generation module 505 may acquire a product of the floating-point number with a predetermined first factor to obtain a sixth intermediate variable if the first exponent is equal to 0; acquire a first array of randomized integer data, a length of the first array being equal to the specified length, and a result obtained after a predetermined linear operation on the first array being the sixth intermediate variable; and evenly distribute the first factor among elements in the first array, and determine the floating-point number vector in combination with a type of the linear operation.
In one embodiment of the present disclosure, the third generation module 505 may divide the elements in the first array by the first factor to obtain the floating-point number vector if the linear operation is an addition operation; and randomly select M elements from the first array, divide the selected elements by 2, use the M elements divided by 2 to form a second array, and divide elements in the second array by 2N to obtain the floating-point number vector if the linear operation is a multiplication operation; wherein N is a quotient obtained by dividing a third predetermined value by the specified length, and M is a remainder obtained by dividing the third predetermined value by the specified length.
In one embodiment of the present disclosure, the first factor may be 2149, and the third predetermined value may be 149.
In one embodiment of the present disclosure, the third generation module 505 may determine a second factor according to the first exponent, and acquire a product of the floating-point number with the second factor to obtain a seventh intermediate variable if the first exponent is not equal to 0; acquire a third array of the randomized integer data, a length of the third array being equal to the specified length, and a result obtained after a predetermined linear operation on the third array being the seventh intermediate variable; and evenly distribute the second factor among elements in the third array, and determine the floating-point number vector in combination with a type of the linear operation.
In one embodiment of the present disclosure, the third generation module 505 may divide the elements in the third array by the second factor to obtain the floating-point number vector if the linear operation is an addition operation; and randomly select M′ elements from the third array, divide the selected elements by 2, use the M′ elements divided by 2 to form a fourth array, and divide elements in the fourth array by 2N′ to obtain the floating-point number vector if the linear operation is a multiplication operation; wherein N′ is a quotient obtained by dividing a fourth predetermined value by the specified length, and M′ is a remainder obtained by dividing the fourth predetermined value by the specified length.
In one embodiment of the present disclosure, the second factor may be 2P, where P is the fourth predetermined value and a difference between 149 and the first exponent.
Specific operation flows of the apparatus embodiments shown in
In brief, by use of the solution in the apparatus embodiment of the present disclosure, a floating-point number and a floating-point number vector randomized can be obtained, so as to make up for the deficiency that the current system Verilog platform provides randomization only for fixed-point data. Moreover, the user is only required to perform some simple settings, such as setting some constraint conditions, and other processing can be completed automatically, thereby simplifying user operations. In addition, the floating-point number and the floating-point number vector obtained can be used to verify the relevant operation units in IC chip verification, such as a floating-point operation unit and a floating-point vector operation unit, and a verification convergence process of the relevant operation units can be accelerated.
The solutions of the present disclosure may be applied to the field of artificial intelligence, and in particular, to fields such as artificial intelligence chips and cloud computing. Artificial intelligence is a discipline that studies how to make computers simulate certain thinking processes and intelligent behaviors (such as learning, reasoning, thinking and planning) of human beings, which includes hardware technologies and software technologies. The artificial intelligence hardware technologies generally include sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing and other technologies. The artificial intelligence software technologies mainly include a computer vision technology, a speech recognition technology, a natural language processing technology, machine learning/deep learning, a big data processing technology, a knowledge graph technology and other major directions.
In addition, acquisition, storage, use, processing, transmission, provision, and disclosure of users' personal information involved in the technical solutions of the present disclosure comply with relevant laws and regulations, and do not violate public order and moral.
According to embodiments of the present disclosure, the present disclosure further provides an electronic device, a readable storage medium, and a computer program product.
As shown in
A plurality of components in the device 700 are connected to the I/O interface 705, including an input unit 706, such as a keyboard and a mouse; an output unit 707, such as various displays and speakers; a storage unit 708, such as disks and discs; and a communication unit 709, such as a network card, a modem and a wireless communication transceiver. The communication unit 709 allows the device 700 to exchange information/data with other devices over computer networks such as the Internet and/or various telecommunications networks.
The computing unit 701 may be a variety of general-purpose and/or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller or microcontroller, etc. The computing unit 701 performs the methods and processing described above, such as the method described in the present disclosure. For example, in some embodiments, the method described in the present disclosure may be implemented as a computer software program that is tangibly embodied in a machine-readable medium, such as the storage unit 708. In some embodiments, part or all of a computer program may be loaded and/or installed on the device 700 via the ROM 702 and/or the communication unit 709. One or more steps of the method described in the present disclosure may be performed when the computer program is loaded into the RAM 703 and executed by the computing unit 701. Alternatively, in other embodiments, the computing unit 701 may be configured to perform the method described in the present disclosure by any other appropriate means (for example, by means of firmware).
Various implementations of the systems and technologies disclosed herein can be realized in a digital electronic circuit system, an integrated circuit system, a field programmable gate array (FPGA), an application-specific integrated circuit (ASIC), an application-specific standard product (ASSP), a system on chip (SOC), a load programmable logic device (CPLD), computer hardware, firmware, software, and/or combinations thereof. Such implementations may include implementation in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, configured to receive data and instructions from a storage system, at least one input apparatus, and at least one output apparatus, and to transmit data and instructions to the storage system, the at least one input apparatus, and the at least one output apparatus.
Program codes configured to implement the methods in the present disclosure may be written in any combination of one or more programming languages. Such program codes may be supplied to a processor or controller of a general-purpose computer, a special-purpose computer, or another programmable data processing apparatus to enable the function/operation specified in the flowchart and/or block diagram to be implemented when the program codes are executed by the processor or controller. The program codes may be executed entirely on a machine, partially on a machine, partially on a machine and partially on a remote machine as a stand-alone package, or entirely on a remote machine or a server.
In the context of the present disclosure, machine-readable media may be tangible media which may include or store programs for use by or in conjunction with an instruction execution system, apparatus or device. The machine-readable media may be machine-readable signal media or machine-readable storage media. The machine-readable media may include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses or devices, or any suitable combinations thereof. More specific examples of machine-readable storage media may include electrical connections based on one or more wires, a portable computer disk, a hard disk, a RAM, a ROM, an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.
To provide interaction with a user, the systems and technologies described here can be implemented on a computer. The computer has: a display apparatus (e.g., a cathode-ray tube (CRT) or a liquid crystal display (LCD) monitor) for displaying information to the user; and a keyboard and a pointing apparatus (e.g., a mouse or trackball) through which the user may provide input for the computer. Other kinds of apparatuses may also be configured to provide interaction with the user. For example, a feedback provided for the user may be any form of sensory feedback (e.g., visual, auditory, or tactile feedback); and input from the user may be received in any form (including sound input, speech input, or tactile input).
The systems and technologies described herein can be implemented in a computing system including background components (e.g., as a data server), or a computing system including middleware components (e.g., an application server), or a computing system including front-end components (e.g., a user computer with a graphical user interface or web browser through which the user can interact with the implementation mode of the systems and technologies described here), or a computing system including any combination of such background components, middleware components or front-end components. The components of the system can be connected to each other through any form or medium of digital data communication (e.g., a communication network). Examples of the communication network include: a local area network (LAN), a wide area network (WAN), and the Internet.
The computer system may include a client and a server. The client and the server are generally far away from each other and generally interact via the communication network. A relationship between the client and the server is generated through computer programs that run on a corresponding computer and have a client-server relationship with each other. The server may be a cloud server, a distributed system server, or a server combined with blockchain.
It should be understood that the steps can be reordered, added, or deleted using the various forms of processes shown above. For example, the steps described in the present disclosure may be executed in parallel or sequentially or in different sequences, provided that desired results of the technical solutions disclosed in the present disclosure are achieved, which is not limited herein.
The above specific implementations do not limit the scope of protection of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and replacements can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present disclosure all should be included in the scope of protection of the present disclosure.
| Number | Date | Country | Kind |
|---|---|---|---|
| 202111571211.7 | Dec 2021 | CN | national |
