<html><head><style type='text/css'>p { margin: 0; }</style></head><body><div style='font-family: Arial; font-size: 12pt; color: #000000'><FONT face="times new roman, new york, times, serif">
<P><BR>> Kilian CAVALOTTI wrote:<BR>>> On Wednesday 08 April 2009 19:48:30 Joe Landman wrote:<BR>>>> As an FYI, Beowulf veteran Jeff Layton wrote up a nice article on<BR>>>> memory<BR>>>> configuration issues for Nehalem (I had seen some discussion on this<BR>>>> previously).<BR>>>><BR>>>> Link is here:<BR>>>> http://www.delltechcenter.com/page/04-08-2009+-+Nehalem+and+Memory+Configur<BR>>>> ations<BR>>><BR>>> That's a great write-up, gathering all the useful information in one<BR>>> place. A must read. Thanks Jeff!<BR></P>
<P>It is a good write up, very good ... but there looks to be an error under "Memory Bandwidth</P>
<P>Performance" ... first, for socket-local-memory, I am not sure of the relevance of</P>
<P>the QPI GT rate. I would calculate the maximum rate to socket-local-memory</P>
<P>with 1 DIMM per channel at:</P>
<P> </P>
<P>1.333 GHz x 3 Channels x 8 bytes = 31.992 GBytes/sec </P>
<P> </P>
<P>not 35 GByters/sec. Maybe I am doing something wrong. Rates to socket-remote</P>
<P>memory would be lower of course, bounded by the QPI 6.4 GT/sec transfer rate, which</P>
<P>on this 2 byte duplex channel would be 25.6 GBytes/sec I believe. There would also</P>
<P>be a performance drop due to the latency increase from 60 nanos to 100 nanos.</P>
<P> </P>
<P>Perhaps I have made an error somewhere ... if so, I would love have someone tell where I</P>
<P>made my mistake.</P>
<P> </P>
<P>rbw</P></FONT></div><br />--
<br />This message has been scanned for viruses and
<br />dangerous content by
<a href="http://www.mailscanner.info/"><b>MailScanner</b></a>, and is
<br />believed to be clean.
</body></html>