<br><br><div class="gmail_quote">On Wed, Jul 15, 2009 at 6:18 PM, JEFFREY S GALL <span dir="ltr"><<a href="mailto:jsg229@psu.edu">jsg229@psu.edu</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<span><span>When I compile a FORTRAN program using ifort, I need to add
the <br>
command 'OPTIONS byteswapped' </span></span></blockquote><div><br></div><div>Do not use "OPTIONS byteswapped", this construct is deprecated. Instead, explicitly state the endianess of of your data</div>
<div><br></div><div>OPTIONS little_endian</div><div><br></div><div>(if you are using ifort, good chance you have a little endian machine.) Although you can tell the compiler to automatically convert your data to big endian as Ryo suggests,</div>
<div><font class="Apple-style-span" face="arial, sans-serif" size="4"><span class="Apple-style-span" style="border-collapse: collapse; font-size: 14px;"><span class="Apple-style-span" style="border-collapse: separate; font-size: small;"><br>
</span></span></font></div><div><span class="Apple-style-span" style="font-family: arial, sans-serif; font-size: 14px; border-collapse: collapse; "> $ ifort -convert big_endian yourprog.f90</span></div><div><br></div><div>
this is not recommended unless you expect to use your data on a big-endian machine (these are becoming more of a rarity these days.) If you do this big-endian conversion, every time you use your data on a little endian machine (such as i686, x86_64 platforms, including Linux, Windows, Mac OS X/Intel) GrADS will be doing a byte swap for each byte it reads - kind of inefficient.</div>
<div><br></div><div> Arlindo</div><div><br></div><div><br></div><div> </div></div><br>-- <br>Arlindo da Silva<br><a href="mailto:dasilva@alum.mit.edu">dasilva@alum.mit.edu</a><br>