RP with Beefier Zynq chip
Posted: Tue Jul 07, 2015 6:43 am
Hey All,
I'm new to the forum, so I am sure that this has been asked before:
I am presently implementing a high-speed/high-order IIR/FIR filter-- I'm targeting ~10 degrees of phase at 1MHz (unlike the one currently available in the ecosystem, implemented in software, with 360 degrees at 1kHz). This requires a LOT of dsp48 slices, and I'm getting squeezed (need a _lot_ of dynamic range, so it's 4slices/multiply! YIKES!). I think there is going to be enough space for a 5th order polynomial in each of the numerator and denominator of z-transform, which is nice, but not _that_ nice... I need more dsp48s, I think...
So the question is this-- is there anything in the works to use even the zynq 7020 in a future version of the board? I don't even dare hope for one of the Kintex models... Of course none of the others use the CLG225 package so it would require a complete redesign of the pcb, but maybe???
Cheers,
Jon
I'm new to the forum, so I am sure that this has been asked before:
I am presently implementing a high-speed/high-order IIR/FIR filter-- I'm targeting ~10 degrees of phase at 1MHz (unlike the one currently available in the ecosystem, implemented in software, with 360 degrees at 1kHz). This requires a LOT of dsp48 slices, and I'm getting squeezed (need a _lot_ of dynamic range, so it's 4slices/multiply! YIKES!). I think there is going to be enough space for a 5th order polynomial in each of the numerator and denominator of z-transform, which is nice, but not _that_ nice... I need more dsp48s, I think...
So the question is this-- is there anything in the works to use even the zynq 7020 in a future version of the board? I don't even dare hope for one of the Kintex models... Of course none of the others use the CLG225 package so it would require a complete redesign of the pcb, but maybe???
Cheers,
Jon