Hi Danny I think I have managed to fix this bu would appreciate confirmation.
It seems that large distributed RAM buffers can cause exceptions under certain conditions. In this case the lookup buffers in the reorder blocks in the unscrambler block, if implemented in distributed RAM and bigger than a certain size cause this exception.
I have modified the scripts to enable the use of BRAMs for the lookups for 'large' lookups. I can now compile a 2^13 fft_biplex block with no hassles.
Fix is commit 3b0fc52 in the ska-sa repo. Regards Andrew On 29/01/2014 13:16, Price, Daniel wrote:
Hi all I am trying to black-box a 2^13 fft_biplex, but it is failing with a non-specific error (reported by: unspecified, error 0001: caught standard exception). It is simulating / updating OK, but the system generator step fails after a few hours. It compiles fine for 2^10, and 2^12 points, with otherwise identical settings. This is happening in our stable library (forked ~6 months ago) and the current ska-sa library. Does anyone have any ideas what could be causing this? Thanks Danny
