f/4, and f/5, where f is the maximum operating frequency of the array
architectures.
4.
You are required to design 2-parallel and 3-parallel direct form structures for 4tap FIR filter. Write the HDL code for the designs and synthesize them for FPGA
implementation. By running suitable test benches show that the designs are
functionally correct. Find the LUT count and maximum usable frequency and
throughput in each of these two cases and compare those. Use fractional values
of filter coefficients of your choice in the range [-1 1]. Consider the input signal
to be random fractions in the range [-1 1]. Assume the coefficients as well as the
input signals to be of 8-bit width. Perform the scaling of input, coefficients, and
output wherever you feel necessary.
5.
You are required to design an 8-tap FIR filter in its direct form structure using 8
multipliers and an adder-tree consisting of 7 adders without any pipelining. Next,
you need to explore all possible pipelining options within the adder-tree and in
between adder-tree and the multiplication stage. You are allowed to use as many
registers as you need for pipelining. Synthesize and test your design without
pipelining and the designs with all possible pipelining options. From the synthesis
result, find the maximum operating frequency along with the LUT requirement
and register count in each of the designs. Use fractional values of filter
coefficients of your choice in the range [-1 1]. Consider the input signal to be
random fractions in the range [-1 1]. Assume the coefficients as well as the input
signals to be of 8-bit width. Perform the scaling of input, coefficients, and output
wherever you feel necessary.
6. Design a structure for the implementation of 16-point radix-2 decimation-infrequency (DIF) fast Fourier transform (FFT) using only 8 butterfly units, where
different stages for FFT computation are time multiplexed. Write the HDL code
for the design and synthesize that for FPGA implementation. By running suitable
test bench show that the design is functionally correct. From the synthesis result,
find the maximum operating frequency along with the LUT requirement and
register count in each of the designs. Consider the input signal to be random
fractions in the range [-1 1]. Assume the multiplying coefficients as well as the
input signals to be of 8-bit width. Perform the scaling of input, coefficients, and
output wherever you feel necessary.
7.
8. RSA (named after the surnames of Ron Rivest, Adi Shamir, and Leonard
Adleman) is one of most widely used public-key cryptosystems. Explore
algorithm and architecture optimization of RSA encryption and decryption for
256 bits key length. Write HDL code of your design and synthesize your design.
Estimate the area, performance, and power consumption of your design and
compare that with one of the popular/state of the art implementation. Try to
reduce the energy consumption of your design through algorithm and
architecture optimization over the state of the art low power designs.
9. Design architectures to implement a direct form 32-tap FIR filter using Wallace
tree multiplier and pipeline adder tree using seamless-delay model and discretedelay model. Write the HDL code of both the designs of this FIR filter and
synthesize them for FPGA implementation. By running suitable test benches
show that the designs are functionally corrects. Compare the LUT counts and
maximum operating frequencies of both the designs. Use fractional values of
filter coefficients of your choice in the range [-1 1]. Consider the input signal to
be random fractions in the range [-1 1]. Assume the coefficients as well as the
input signals to be of 8-bit width. Perform the scaling of input, coefficients, and
output wherever you feel necessary.