<HTML><BODY style="word-wrap: break-word; -khtml-nbsp-mode: space; -khtml-line-break: after-white-space; "><DIV>just to share my experience:</DIV>on our (CSCS) XT3 machine IT people installed a special library (iobuf, vendor libraries from CRAY)..<DIV>this library does what you were describing in your e-mail.. is intercepting all IO operations and keeps them in memory and only when the IO buffer is full (~ 100 MB per node) everything is flushed.. in this way all the problems related to the scalability of the code due to IO operations are highly dumped..</DIV><DIV><DIV><BR><DIV><DIV>On 9 Jul 2007, at 19:31, Axel wrote:</DIV><BR class="Apple-interchange-newline"><BLOCKQUOTE type="cite"><P style="margin: 0.0px 0.0px 0.0px 0.0px"><FONT face="Helvetica" size="3" style="font: 12.0px Helvetica">this may not be a big issue for most of the current machine and</FONT></P> <P style="margin: 0.0px 0.0px 0.0px 0.0px"><FONT face="Helvetica" size="3" style="font: 12.0px Helvetica">users, but i expect that more and more machines will have to use</FONT></P> <P style="margin: 0.0px 0.0px 0.0px 0.0px"><FONT face="Helvetica" size="3" style="font: 12.0px Helvetica">parallel file systems like lustre, GPFS and alike and at the same</FONT></P> <P style="margin: 0.0px 0.0px 0.0px 0.0px"><FONT face="Helvetica" size="3" style="font: 12.0px Helvetica">time, people want to run larger jobs faster on their new fancy</FONT></P> <P style="margin: 0.0px 0.0px 0.0px 0.0px"><FONT face="Helvetica" size="3" style="font: 12.0px Helvetica">machines,</FONT></P> <P style="margin: 0.0px 0.0px 0.0px 0.0px"><FONT face="Helvetica" size="3" style="font: 12.0px Helvetica">so this may become a much more important over the next few years.</FONT></P> </BLOCKQUOTE></DIV><BR></DIV></DIV></BODY></HTML>