<h2><a name="Node78"></a></h2><h2><a name="Node78"></a></h2><h2>Hello,<br></h2><h2><a name="Node78"> MPICH2's predefined reduce operation seems it can not support quadruple precision.</a></h2><h2><a name="Node78"></a></h2>
<h2> I used the attached program to test this. And got the following error message,<br></h2>Fatal error in PMPI_Reduce: Invalid MPI_Op, error stack:<br>PMPI_Reduce(1198).......: MPI_Reduce(sbuf=0012FF00, rbuf=0012FF10, count=1, dtyp<br>
e=0x4c00102b, MPI_SUM, root=0, MPI_COMM_WORLD) failed<br>MPIR_SUM_check_dtype(97): MPI_Op MPI_SUM operation not defined for this datatype<br><br><h2> If so, is there any plan to support this ?<br></h2><h2> Thank you.</h2>
<h2><br></h2><h2>YIN<a name="Node78"><br></a></h2>