Several bugs fixed in new LLR
- Hi All,
To-day, I replaced the three new LLR zip files after having fixed several bugs :
- In SSE2 mode, false negatives with all zeroes in the 64 bits residue sometimes occured.
- More serious bug : The Proth mode was not initialized correctly !
- This program is not still able to use correctly the IBDWT mode with a rational FFT, that is
to say, when the exponent is a multiple of the fft length, so, I force using the Proth mode in
Would you excuse me for the difficulties you came across due to these bugs...