How to convert to integer or double a std_logicc_vector of 40 bits in VHDL

[kalyansrinivas]

How to convert a std_logic_vector of 40 bits to integer or double in VHDL

Dear Friends

In VHDL we have conv_integer to convert a std_logic_vector of less than 32 bits to integer

But i have a output result of more than 32 bits (40-bits- std_logic_vector ) and i want it to be as an integer or as a double to be printed out to a text file

Please resolve my issue How to convert a standard logic vector of more than 32 bits to be printed as integer or double to a output text file

Regards
M Kalyansrinivas

 

[TrickyDicky]

first of all, conv_integer is not part of any IEEE standard code libraries. Std_logic_vectors should NOT be used as numbers. You should use the signed and unsigned types from the numeric_std library.

Secondly, to answer your question - you cannot use integers more than 32 bits. But you can output std_logic_vectors directly using the std_logic_textio package in binary (read/write), octal (oread/owrite) and hex (hread/hwrite).

If you have to output it as a base 10 number, you'll have to write a clever conversion function.

 

[kalyansrinivas]

Re: How to convert to integer or double a std_logic_vector of 40 bits in VHDL

Thank you very much

If i write a logic for converting the 40 bit vector to a base 10 value in vhdl then the output result should be of certain data type (as data type integer will hold only -2^32 to + 2^32) .Can you tell me what data type should i declare for holding this value

Regarding signed and unsigned should be used only in testbenches or also in the synthesizable code as we have std_logic_vector takes care of every thing(2's complement format)

Regards
M Kalyansrinivas

 

[TrickyDicky]

There is no data type to hold it. You'll have to write some clever function that converters the 40 bit number into two integers.

---------- Post added at 08:24 ---------- Previous post was at 08:23 ----------

Signed and unsigned are perfectly synthesisable. The weakness with std_logic_vector is you cannot do signed and unsigned in the same file.

 

[FvM]

I'm not sure what converting to a base 10 datatype means. Internally, all data are binary. The problem is output outputting a decimal number. You can use binary coded decimal as an intermediate result when converting the binary bit vector. It's simply a bit vector of 4*number of digits length, that can in fact hold the intended decimal representation of the number. Each 4 bit subvector represents one decimal digit.

For a conversion routine that doesn't rely on constrained integer, you can refer to double dabble algorithm Double dabble - Wikipedia, the free encyclopedia

 

[TrickyDicky]

The problem is writing the data to a text file in base 10. For string types (std_logic_vector etc) there only exist functions to write to a text file in binary, octal or hex. The only thing that will print out in base 10 are integers (the default write function) or a string. So you either have to get clever with integers, or write a base 10 std_logic_vector to string converter. neither will be trivial.