<div dir="ltr">Hi All,<div><br></div><div>Sanity check --- my reading of the spec. indicates that accumulate operations performed on pair types guarantee atomicity at the level of the whole pair type. Here are the relevant passages:</div><div><br></div><div><p class="MsoNormal" style="margin:0in 0in 0.0001pt 0.5in;font-size:12pt;font-family:Calibri,sans-serif"><span style="font-size:11pt">Section
11.7.1: The outcome of concurrent accumulate operations to the same location
with the same <u>predefined</u> datatype is as if the accumulates were done at that
location in some serial order.</span></p>
<p class="MsoNormal" style="margin:0in 0in 0.0001pt 0.5in;font-size:12pt;font-family:Calibri,sans-serif"><span style="font-size:11pt"> </span></p>
<p class="MsoNormal" style="margin:0in 0in 0.0001pt 0.5in;font-size:12pt;font-family:Calibri,sans-serif"><span style="font-size:11pt">Section
5.9.4: In order to use MPI_MINLOC and MPI_MAXLOC in a reduce operation, one
must provide a datatype argument that represents a pair (value and index). MPI
provides nine such <u>predefined</u> datatypes.</span></p></div><div><br></div><div>Is that right?</div><div><br></div><div>Thanks,</div><div> ~Jim.</div></div>