Benchmarks

The following circuits are the main core of the floating point applications:

Name Description Circuit Diagram VHDL source (double precision) VHDL source (single precision)
dscg a digital sine-cosine generator dscg_circuit.pdf dscg_double.vhd dscg_single.vhd
bfly a Fast Fourier Transform circuit (z = x * w + y) with complex number inputs and outputs butterfly_circuit.pdf butterfly_double.vhd butterfly_single.vhd
fir4 a 4-tap finite impulse response filter fir_circuit.pdf fir_double.vhd fir_single.vhd
ode an ordinary differential equation solver ode_circuit.pdf ode_double.vhd ode_single.vhd
mm3 a 3x3 matrix multiplier mm3_circuit.pdf mm3_double.vhd mm3_single.vhd
bgm a Monte Carlo simulations of interest rate model derivatives circuit please refer to this paper Not available Not available
syn2 a synthetic floating point benchmark circuit generated by a synthetic
benchmark generator, which contains 5 floating point adders and 4 floating point multipliers
Not available syn2_double.vhd syn2_single.vhd
syn7 a synthetic floating point benchmark circuit generated by a synthetic
benchmark generator, which contains 25 floating point adders and 25 floating point multipliers
Not available syn7_double.vhd syn7_single.vhd

@

Click here to download the double precision floating point components used in the applications

@

Click here to download the single precision floating point components used in the applications

@

@