Spectral Methods using posits

  • 1 Replies
  • 240 Views
Spectral Methods using posits
« on: April 17, 2019, 02:20:40 PM »
We have developed an open source sw emulation library for a new number system called posits to replace IEEE floating point. Posits are more accurate as compared to IEEE at the same size. And the posit system also provides a user controlled rounding mode. It is this user controlled rounding mode that makes it possible to create forward/inverse FFT pipelines that perfectly reconstruct the original signal. In this user controlled rounding mode, 16-bit posits can compete with 64-bit double precision IEEE floating point in terms of accuracy.

We would like to collaborate with the Scope-Fun team to enable this oscilloscope with spectral measurement capability.

*

Dejan

  • *****
  • 46
    • View Profile
Re: Spectral Methods using posits
« Reply #1 on: April 17, 2019, 06:50:32 PM »
Thanks for your offer. Please send us an email to discuss more about this.