hi danny,
i suggest trying jason's items first (below),
and then if you are still having trouble, talk to suraj gowda,
who got got the PFB and FFT to work at about 375 MHz
by using xilinx's "plan-ahead" and calling out more explicit
utilization of DSP48 slices for the FFT butterflies.
suraj will likely be writing this up soon.
best wishes,
dan
On 9/3/2010 7:08 AM, Jason Manley wrote:
Try adder latency 1, multiplier latency 2, bram latency 3. Be sure to optimise to use
DSP48 multipliers and tick any boxes asking to "pull adders into DSP48s".
This should allow adders following multipliers to use a single DSP48 slice,
with registered input and output (making for a shorter signal path and higher
speeds). You might want to check the physical layout of a completed design
(using floorplanner) to see what it's actually done. This should help you to
figure out why things aren't optimal.
If it comes down to it, you might have to resort to manual placement of some
bits (something I've never attempted).
Jason
On 03 Sep 2010, at 3:58 PM, Danny Price wrote:
Hi all, I'm trying to get the tute 3 spectrometer design running at 250MHz, and
am running into timing issues. The timing report suggests problems in the PFB
and FFT -- does anyone have some recommended settings for the PFB and FFT (i.e.
latencies etc), before I start adding delays haphazardly under the mask?
Also, any suggestions on a methodical strategy to meet timing?
Thanks
Danny
