<div dir="ltr">
<div class="page" title="Page 493">
<div class="layoutArea">
<div class="column">
<p>
</p><div class="page" title="Page 493">
<div class="layoutArea">
<div class="column">
<p><span style="font-size:12.000000pt;font-family:'CMSS12'">"11.7.2 Ordering" says: </span></p>
</div>
</div>
</div><p><span style="font-size:11.000000pt;font-family:'CMSS10'">"MPI </span><span style="font-size:11.000000pt;font-family:'CMR10'">specifies ordering between accumulate operations from one process to the same (or
overlapping) memory locations at another process on a per-datatype granularity."</span></p>
</div>
</div>
</div><div>This seems to say that ordering is required between overlapping windows, which is a horrible semantic. We need to add a "per-window" caveat here.</div><div><br></div><div>Later, we talk about overlapping memory regions, rather than locations, which is confusing:</div><div>
<div class="page" title="Page 493">
<div class="layoutArea">
<div class="column">
<p><span style="font-size:11.000000pt;font-family:'CMSS10'">"MPI </span><span style="font-size:11.000000pt;font-family:'CMR10'">does not provide any guarantees
for accesses or updates from different origin processes to overlapping target memory regions."</span></p>
</div>
</div>
</div></div><div>What is a memory region anyways? Is it a window? Is it a set of memory locations?</div><div><br></div><div>And to be clear, I think I know what we mean in the RMA chapter, but others may not.</div><div><br></div><div>Jeff</div><div><br></div>-- <br><div class="gmail_signature">Jeff Hammond<br><a href="mailto:jeff.science@gmail.com" target="_blank">jeff.science@gmail.com</a><br><a href="http://jeffhammond.github.io/" target="_blank">http://jeffhammond.github.io/</a></div>
</div>