Virtex 5 GTP Simulation

[maxascent]

I have a Virtex 5 design that uses a GTP. I have a simple simulation tha
justs transmits some data and looks at the received data. The simulatio
works fine and I see the correct received data. The trouble is when I ru
the design in hardware and look at the data using chipscope it doesn
match. I am using 8b10b coding and the comma character is detecte
correctly but the following data is not correct. The data doesnt seem to b
rubbish as if I send two bytes that are the same (4A 4A) I get two byte
that are the same (38 38) just the wrong value. Has anybody seen anythin
like this?

Thanks

Jon

---------------------------------------
Posted through http://www.FPGARelated.com

 

[maxascent]

I have the ILA instantiated in the design. As I said the data does seem t
have some sort of pattern so I guess it could be a chipscope issue. Its
bit strange that the comma value is correct but the data isnt. I will hav
a go using the inserter and see if that makes any difference.

Cheers
Jon

---------------------------------------
Posted through http://www.FPGARelated.com

 

[Gabor]

On Oct 29, 11:08 am, "maxascent"
<maxascent@n_o_s_p_a_m.n_o_s_p_a_m.yahoo.co.uk> wrote:

I have a Virtex 5 design that uses a GTP. I have a simple simulation that
justs transmits some data and looks at the received data. The simulation
works fine and I see the correct received data. The trouble is when I run
the design in hardware and look at the data using chipscope it doesnt
match. I am using 8b10b coding and the comma character is detected
correctly but the following data is not correct. The data doesnt seem to be
rubbish as if I send two bytes that are the same (4A 4A) I get two bytes
that are the same (38 38) just the wrong value. Has anybody seen anything
like this?

Thanks

Jon        

---------------------------------------        
Posted throughhttp://www.FPGARelated.com

How did you attach the ChipScope ILA? Is it instantiated in your
project or did
you use the inserter? I have in the past tried to debug bit-ordering
issues that
ended up being ChipScope re-ordering the bits for me. See if you can
map
the values you get to the expected values by re-arranging the bits.
For example
hex 38 and hex 4A each have exactly three 1 bits and five 0 bits.

Regards,
Gabor