VHDL fatal error with modelsim

[Adnan86]

May any one help .
this statment Is it true ?

lu it's real
sumU it's real too and both SIGNAL .

FOR j IN 2 to n LOOP
FOR k IN 1 to j-1 LOOP
sumU <= (lu( i , k ))*(lu( k , j )) + sumU ;
END LOOP ;
END LOOP ;

 

[mrflibble]

What do you think the LOOP statement does in VHDL?

 

[TrickyDicky]

whats the error?
and whats the real code?

btw, I dont think this code is going to behave the way you think it's going to.

 

[Adnan86]

YES it's VHDL ... and first value of sumu it's 0.0

- - - Updated - - -

just say fatal error in this line ...

- - - Updated - - -

I use modelsim .. ihaven't syntax eeror but as i simulate i have this error

 

[TrickyDicky]

What error - where?

 

[Adnan86]

I use variable istead SIGNAL too but i have same error ...

- - - Updated - - -

# Fatal error in ForLoop loop at D:/VHDL/Modelsim test/marix_inversion_nxn/Matrix_Inversion_NxN.vhd line 58

line 58 it's sumU := (lu( i , k ))*(lu( k , j )) + sumU ;

 

[mrflibble]

btw, I dont think this code is going to behave the way you think it's going to.

Heh, neither do I. Hence my question.

@Adnan86: before going further with more posts along the line of "help! error! it has error! help!" ... I suggest you spend 10 minutes to look through a tutorial / book / whatever is fastest for you to read what the LOOP statement actually does. Hint: probably not what you think.

edit: okay, matrix multiplication. Well now I just KNOW it does not do what you think it does. So read up on that LOOP statement.

 

[TrickyDicky]

and how about posting the whole code. The snippet you have posted means nothing out of context - post the whole code.

 

[Adnan86]

@mrflibble : I looked to 1 or 2 book my code just same that they said but i really don't now where is my problem



this my whole code

 

[TrickyDicky]

PS. Please use the code or syntax tags to post the code

(awaits post of pure code without tags)

 

[TrickyDicky]

why not post the real error, and things might have happened sooner?

Plus, your code snippet does not match the real code!

SumU running to infinity or NAN? then thats your problem to fix.

- - - Updated - - -

PS. with your code, if n> 4, you'll get other errors.

And what do you expect to do with this code - you cannot synthesise it because real is not synthesisable. I assume this is just a behavioral model or something.

 

[Adnan86]

no not infinity : generic ( n : integer := 4 ); so not infinity .
so what can i do for solving .

- - - Updated - - -

I now this is just a model ...
for n > 4 i didn't try but if i have answer for 3 or 4 for now it's good for more i'd try later but now just for n=4 ..


- - - Updated - - -

this is an error :

** Fatal: (vsim-3421) Value 1.#INF for sumu is out of range -1e+308 to 1e+308.

Any way thanks for help ...

 

[TrickyDicky]

Yup, Sumu is going to infinity for your input matrix. So you either need to find a matrix that doesnt go to infinity or modify the code.

 

[Adnan86]

i changed it but not solved ... i give input in test bench to input matrix but after simulate for my matrix input it's infinity too ...