FPGARelated.com
Forums

clock enable multicycle doesn't work with altera altshift_taps megafunction

Started by Wilhelm Klink August 14, 2004
Has anyone encountered this problem with Altera/Quartus?

I have a set of cascaded registers which can either be implemented
using cascaded lpm_ffs or an altshift_taps unit.  I have a main clock
and an enable signal with a clock enable multicycle setting of 5.  If
I implement the registers using lpm_ffs then the required clock setup
time is correctly stated in the timing analyzer as being 33.33ns (main
clock is at 150MHz -> 6.666ns period, 5*6.666ns = 33.33ns).  However
if I implement the registers using altshift_taps the required clock
setup time is stated as being 6.666ns in the timing analyzer section
of the compilation report, ie the clock enable multicycle setting of 5
has been ignored.  I first noticed this type of problem when I had the
auto shift register replacement option in the analysis and synthesis
settings turned on, which converts lpm_shiftregs to altshift_taps.  I
can provide design files if anyone wants to try and help, I have
waited 5 days for a response from Altera support.
Hi Wilhelm,

Please send me the .qar for your project, so that we can check if it is a
bug or not and help you.

Subroto Datta (sdatta@altera.com)
Altera Corp.

"Wilhelm Klink" <kommandantklink@hotmail.com> wrote in message
news:6011e208.0408140511.1eecd45a@posting.google.com...
> Has anyone encountered this problem with Altera/Quartus? > > I have a set of cascaded registers which can either be implemented > using cascaded lpm_ffs or an altshift_taps unit. I have a main clock > and an enable signal with a clock enable multicycle setting of 5. If > I implement the registers using lpm_ffs then the required clock setup > time is correctly stated in the timing analyzer as being 33.33ns (main > clock is at 150MHz -> 6.666ns period, 5*6.666ns = 33.33ns). However > if I implement the registers using altshift_taps the required clock > setup time is stated as being 6.666ns in the timing analyzer section > of the compilation report, ie the clock enable multicycle setting of 5 > has been ignored. I first noticed this type of problem when I had the > auto shift register replacement option in the analysis and synthesis > settings turned on, which converts lpm_shiftregs to altshift_taps. I > can provide design files if anyone wants to try and help, I have > waited 5 days for a response from Altera support.