| Previous | Next |
| ippStsZeroOcc | ippStsSingularity |
ippStsUnderflow
Underflow occurs when a nonzero mathematical result is too small to be represented normally in the destination floating-point type or fixed-point scale. The operation may continue with zero, subnormal or rounded output.
This is usually a numerical-scale issue. It can be expected in probability, spectral or filter-tail calculations and still be important for later normalization.
Check these details
- Input scale, coefficient scale and destination precision.
- Whether denormal or flush-to-zero behavior changes the result.
- Later stages that divide by a value that may have underflowed to zero.
Intel IPP Error Reporting · IPP special-case handling
Looking for a different code? Search another status or error code.