WHY usi is two amp operation in teradata?

Generally the Parsing Engine preapres a two amp plan for
USI, as in case of USI there will be a subtable which has
the hashed value of the USI and the ROW ID of the Primary
Index. So when you search using the USI first it will go to
the AMP which has the subtable and then using the ROW id
from the subtable it will go to the AMP where the actual
row resides.

In a case where the PI and USI value are same it can be a
one amp operation as the HASHED value of oth the PI and USI
will go to the same AMP.