[Fpga-synth] impact failing to pick up changes?

Magnus Danielson magnus at rubidium.dyndns.org
Sun Oct 26 02:41:27 CET 2008


Mason wrote:
> Hey all,
>    So I'm using ISE 10.1.02 now, and I'm wondering if anyone else is 
> running into this bug:
>    I compile a design, flash it to the FPGA in JTAG mode, and it runs. 
> Then I make a small change (I've been playing with a basic LED test; the 
> change would be something like the size of the register that gets 
> counted through) and regenerate everything. I flash the new file to the 
> FPGA, and the old bitfile gets transferred.
>    I know I asked the list about this same problem back in the 9.1 era, 
> and even called Xilinx about it and they said the issue was selecting 
> the wrong clock (i.e. not the JTAG clock) when generating the bitstream.
>    So I've changed that, thinking maybe that bug was still there, to no 
> avail.
>    So, my fellow people programming Xilinx boards, how on earth are you 
> doing it?
> 
> (I'm using the 3AN starter kit, with libusb under linux, if that changes 
> anything.)

It may not be relevant, but the ISE has a tendency to skip over errors 
and go to next step using old data even on errors. So please check that 
you do not run into errors.

I'm on ISE 10.1.03 for your reference.

I've found the ISE simulator to be usefull even if it has a few annoying 
aspects. Some are probably due to my inexperience with it. Never the 
less, testbenching the design I am working on properly has proven highly 
meaningfull as I now know I had a bunch of bugs lurking in it.

Cheers,
Magnus


More information about the Fpga-synth mailing list