Documentation

Data Type Conversions

Converting HDL Data to Send to MATLAB

If your HDL application needs to send HDL data to a MATLAB® function, you may first need to convert the data to a type supported by MATLAB and the HDL Verifier™ software.

To program a MATLAB function for an HDL model, you must understand the type conversions required by your application. You may also need to handle differences between the array indexing conventions used by the HDL you are using and MATLAB (see following section).

The data types of arguments passed in to the function determine the following:

  • The types of conversions required before data is manipulated

  • The types of conversions required to return data to the HDL simulator

The following table summarizes how the HDL Verifier software converts supported VHDL® data types to MATLAB types based on whether the type is scalar or array.

VHDL-to-MATLAB Data Type Conversions

VHDL Types...As Scalar Converts to...As Array Converts to...
STD_LOGIC, STD_ULOGIC, and BITA character that matches the character literal for the desired logic state.  
STD_LOGIC_VECTOR, STD_ULOGIC_VECTOR, BIT_VECTOR, SIGNED, and UNSIGNED A column vector of characters (as defined in VHDL Conversions for the HDL Simulator) with one bit per character.
Arrays of STD_LOGIC_VECTOR, STD_ULOGIC_VECTOR, BIT_VECTOR, SIGNED, and UNSIGNED An array of characters (as defined above) with a size that is equivalent to the VHDL port size.
INTEGER and NATURALType int32.Arrays of type int32 with a size that is equivalent to the VHDL port size.
REALType double.Arrays of type double with a size that is equivalent to the VHDL port size.
TIMEType double for time values in seconds and type int64 for values representing simulator time increments (see the description of the 'time' option in hdldaemon).Arrays of type double or int64 with a size that is equivalent to the VHDL port size.
Enumerated types Character vector that contains the MATLAB representation of a VHDL label or character literal. For example, the label high converts to 'high' and the character literal 'c' converts to '''c'''.Cell array of character vectors with each element equal to a label for the defined enumerated type. Each element is the MATLAB representation of a VHDL label or character literal. For example, the vector (one, '2', three) converts to the column vector ['one'; '''2'''; 'three']. A user-defined enumerated type that contains only character literals, and then converts to a vector or array of characters as indicated for the types STD_LOGIC_VECTOR, STD_ULOGIC_VECTOR, BIT_VECTOR, SIGNED, and UNSIGNED.

The following table summarizes how the HDL Verifier software converts supported Verilog® data types to MATLAB types. The software supports only scalar data types for Verilog.

Verilog-to-MATLAB Data Type Conversions

Verilog Types... Converts to...
wire, regA character or a column vector of characters that matches the character literal for the desired logic states (bits).
integerA 32-element column vector of characters that matches the character literal for the desired logic states (bits).

Array Indexing Differences Between MATLAB and HDL

In multidimensional arrays, the same underlying OS memory buffer maps to different elements in MATLAB and the HDL simulator (this mapping only reflects different ways the different languages offer for naming the elements of the same array). When you use both the matlabtb and matlabcp functions, be careful to assign and interpret values consistently in both applications.

In HDL, a multidimensional array declared as:

type matrix_2x3x4 is array (0 to 1, 4 downto 2) of std_logic_vector(8 downto 5);

has a memory layout as follows:

bit   01 02 03 04  05 06 07 08  09 10 11 12  13 14 15 16  17 18 19 20  21 22 23 24
 -
 dim1 0  0  0  0   0  0  0  0   0  0  0  0   1  1  1  1   1  1  1  1   1  1  1  1
 dim2 4  4  4  4   3  3  3  3   2  2  2  2   4  4  4  4   3  3  3  3   2  2  2  2
 dim3 8  7  6  5   8  7  6  5   8  7  6  5   8  7  6  5   8  7  6  5   8  7  6  5

This same layout corresponds to the following MATLAB 4x3x2 matrix:

 bit  01 02 03 04  05 06 07 08  09 10 11 12  13 14 15 16  17 18 19 20  21 22 23 24
 -
 dim1 1  2  3  4   1  2  3  4   1  2  3  4   1  2  3  4   1  2  3  4   1  2  3  4 
 dim2 1  1  1  1   2  2  2  2   3  3  3  3   1  1  1  1   2  2  2  2   3  3  3  3 
 dim3 1  1  1  1   1  1  1  1   1  1  1  1   2  2  2  2   2  2  2  2   2  2  2  2

Therefore, if H is the HDL array and M is the MATLAB matrix, the following indexed values are the same:

 b1  H(0,4,8) = M(1,1,1)
 b2  H(0,4,7) = M(2,1,1)
 b3  H(0,4,6) = M(3,1,1)
 b4  H(0,4,5) = M(4,1,1)
 b5  H(0,3,8) = M(1,2,1)
 b6  H(0,3,7) = M(2,2,1)
 ...
 b19 H(1,3,6) = M(3,2,2)
 b20 H(1,3,5) = M(4,2,2)
 b21 H(1,2,8) = M(1,3,2)
 b22 H(1,2,7) = M(2,3,2)
 b23 H(1,2,6) = M(3,3,2)
 b24 H(1,2,5) = M(4,3,2)

You can extend this indexing to N-dimensions. In general, the dimensions—if numbered from left to right—are reversed. The right-most dimension in HDL corresponds to the left-most dimension in MATLAB.

Converting Data for Manipulation

Depending on how your simulation MATLAB function uses the data it receives from the HDL simulator, you may need to code the function to convert data to a different type before manipulating it. The following table lists circumstances under which you would require such conversions.

Required Data Conversions

If You Need the Function to...Then...
Compute numeric data that is received as a type other than doubleUse the double function to convert the data to type double before performing the computation. For example:
datas(inc+1) = double(idata);
Convert a standard logic or bit vector to an unsigned integer or positive decimalUse the mvl2dec function to convert the data to an unsigned decimal value. For example:
uval = mvl2dec(oport.val)

This example assumes the standard logic or bit vector is composed of the character literals '1' and '0' only. These are the only two values that can be converted to an integer equivalent.

The mvl2dec function converts the binary data that the MATLAB function receives from the entity's osc_in port to unsigned decimal values that MATLAB can compute.

See mvl2dec for more information on this function.

Convert a standard logic or bit vector to a negative decimalUse the following application of the mvl2dec function to convert the data to a signed decimal value. For example:
suval = mvl2dec(oport.val, true);

This example assumes the standard logic or bit vector is composed of the character literals '1' and '0' only. These are the only two values that can be converted to an integer equivalent.

Examples

The following code excerpt illustrates data type conversion of data passed in to a callback:

InDelayLine(1) = InputScale * mvl2dec(iport.osc_in',true);

This example tests port values of VHDL type STD_LOGIC and STD_LOGIC_VECTOR by using the all function as follows:

all(oport.val == '1' | oport.val 
== '0')

This example returns True if all elements are '1' or '0'.

Converting Data for Return to the HDL Simulator

If your simulation MATLAB function needs to return data to the HDL simulator, you may first need to convert the data to a type supported by the HDL Verifier software. The following tables list circumstances under which such conversions are required for VHDL and Verilog.

    Note:   When data values are returned to the HDL simulator, the char array size must match the HDL type, including leading zeroes, if applicable. For example:

    oport.signal = dec2mvl(2)

    will only work if signal is a 2-bit type in HDL. If the HDL type is anything else, you must specify the second argument:

    oport.signal = dec2mvl(2, N)

    where N is the number of bits in the HDL data type.

VHDL Conversions for the HDL Simulator

To Return Data to an IN Port of Type...Then...
STD_LOGIC, STD_ULOGIC, or BITDeclare the data as a character that matches the character literal for the desired logic state. For STD_LOGIC and STD_ULOGIC, the character can be 'U', 'X', '0', '1', 'Z', 'W', 'L', 'H', or '-'. For BIT, the character can be '0' or '1'. For example:
iport.s1 = 'X'; %STD_LOGIC
iport.bit = '1'; %BIT
STD_LOGIC_VECTOR, STD_ULOGIC_VECTOR, BIT_VECTOR, SIGNED, or UNSIGNEDDeclare the data as a column vector or row vector of characters (as defined above) with one bit per character. For example:
iport.s1v = 'X10ZZ'; %STD_LOGIC_VECTOR
iport.bitv = '10100'; %BIT_VECTOR
iport.uns = dec2mvl(10,8); %UNSIGNED, 8 bits
Array of STD_LOGIC_VECTOR, STD_ULOGIC_VECTOR, BIT_VECTOR, SIGNED, or UNSIGNEDDeclare the data as an array of type character with a size that is equivalent to the VHDL port size. See Array Indexing Differences Between MATLAB and HDL.
INTEGER or NATURAL

Declare the data as an array of type int32 with a size that is equivalent to the VHDL array size. Alternatively, convert the data to an array of type int32 with the MATLAB int32 function before returning it. Be sure to limit the data to values with the range of the VHDL type. If you want to, check the right and left fields of the portinfo structure. For example:

iport.int = int32(1:10)';

REALDeclare the data as an array of type double with a size that is equivalent to the VHDL port size. For example:
iport.dbl = ones(2,2); 
TIME

Declare a VHDL TIME value as time in seconds, using type double, or as an integer of simulator time increments, using type int64. You can use the two formats interchangeably and what you specify does not depend on the hdldaemon 'time' option (see hdldaemon), which applies to IN ports only. Declare an array of TIME values by using a MATLAB array of identical size and shape. All elements of a given port are restricted to time in seconds (type double) or simulator increments (type int64), but otherwise you can mix the formats. For example:

iport.t1 = int64(1:10)'; %Simulator time
                         %increments
iport.t2 = 1e-9; %1 nsec
Enumerated typesDeclare the data as a character vector for scalar ports or a cell array of character vectors for array ports with each element equal to a label for the defined enumerated type. The 'label' field of the portinfo structure lists all valid labels (see Gaining Access to and Applying Port Information). Except for character literals, labels are not case sensitive. In general, you should specify character literals completely, including the single quotes, as in the first example shown here. .
iport.char = {'''A''', '''B'''}; %Character
                                 %literal
iport.udef = 'mylabel'; %User-defined label
Character array for standard logic or bit representationUse the dec2mvl function to convert the integer. For example:
oport.slva =dec2mvl([23 99],8)';
This example converts two integers to a 2-element array of standard logic vectors consisting of 8 bits.

Verilog Conversions for the HDL Simulator

To Return Data to an input Port of Type...Then...
reg, wireDeclare the data as a character or a column vector of characters that matches the character literal for the desired logic state. For example:
iport.bit = '1'; 
integerDeclare the data as a 32-element column vector of characters (as defined above) with one bit per character.

Was this topic helpful?