search

edwar-vhd / SFU-Piecewise-Polynomial-Approximation

Special Function Units (SFUs) are hardware accelerators, their implementation helps improve the performance of GPUs to process some of the most complex operations. This SFU implements the Piecewise Polynomial Approximation, which provides high performance, low area costs and good accuracy for real implementation of hardware.
6 stars 1 forks source link

The difference between SFU_00 to SFU_13 #4

Closed zqcawzs closed 1 year ago

zqcawzs commented 1 year ago

Hi,

I see four different SFU designs under Description and I can't find any document to say what's the difference. Could you give some hints?

Thanks

edwar-vhd commented 1 year ago

Hi,

The firt version of the SFU is the design without redundancy un the arquitecture, the others are variations on the design to apply the TMR in the SFU, you can read the section "7.2. TMR EN LA SFU DE PPA" for details.

https://github.com/edwar-vhd/SFU-Piecewise-Polynomial-Approximation/blob/main/Implementaci%C3%B3n%20y%20evaluaci%C3%B3n%20de%20una%20unidad%20de%20funciones%20especiales%20tolerante%20a%20fallas%20basada%20en%20aproximaci%C3%B3n%20polinomial%20por%20partes.pdf

zqcawzs commented 1 year ago

Hi,

The firt version of the SFU is the design without redundancy un the arquitecture, the others are variations on the design to apply the TMR in the SFU, you can read the section "7.2. TMR EN LA SFU DE PPA" for details.

https://github.com/edwar-vhd/SFU-Piecewise-Polynomial-Approximation/blob/main/Implementaci%C3%B3n%20y%20evaluaci%C3%B3n%20de%20una%20unidad%20de%20funciones%20especiales%20tolerante%20a%20fallas%20basada%20en%20aproximaci%C3%B3n%20polinomial%20por%20partes.pdf

Any chance to provide an English version? Thanks.