AnsweredAssumed Answered

Data format bug in VDSP++ 5.0

Question asked by wojti on Jul 23, 2010
Latest reply on Jul 25, 2010 by Afinko

Hi,

 

while debugging my project I was watching Data Memory, which I'd like to dump to a text file. I compute my algorithm on 32-bit float numbers. The hexadecimal representation is alright but when I swiched it to floating point 32 I noticed some values are viewed wrongly, for example:

a number 0x00156D66 is viewed as 1.#IND in stead of 1.9677901841096963e-39.

 

Is there any patch or anything that fixes this situation?

 

Regards,

wojti

Outcomes