| Previous | Next |
| RPC_S_FP_DIV_ZERO | RPC_S_FP_OVERFLOW |
RPC_S_FP_UNDERFLOW
A server calculation produced a magnitude below its supported range
RPC_S_FP_UNDERFLOW is a server arithmetic exception observed through RPC. A calculation in the manager routine produced a result too small for the active floating-point mode and trapped. The NDR layer can transport float and double; it does not decide the server’s numerical tolerance or exception policy.
Inspect the server stack, input scale, data normalization, and floating-point control word. Determine whether gradual underflow is acceptable or whether the algorithm should rescale values before calculation. If inputs are supplied remotely, enforce realistic minimum magnitudes and units at the application boundary rather than depending on platform-specific exception masks.
What to inspect
- Record the exact numeric inputs and units that reached the server.
- Compare optimized and debug builds for floating-point environment differences.
- Add range validation or stable rescaling where tiny values are expected.
References
- Microsoft: RPC floating-point base types
- Microsoft: server exception handling
- Microsoft: RPC status definitions
Looking for a different code? Search another status or error code.