 |
 |
Edward A. Berry wrote: > Kay Diederichs wrote: >> >> the CPU (AMD versus Intel) does not play any role for crystallographic >> computing, but you'll have to decide whether you want to install the >> 64bit or the 32bit version of RHEL4. >> 32bit programs run a bit faster on a 64bit operating system, and with >> a 64bit OS you can run programs which require arrays of more than 3.5 >> GB (seldomly needed in xtallography), but you will have to install in >> parallel many 32- and 64bit libraries. This is no problem by itself, >> but quite confusing at first ... >> So I recommend the 32bit version. >> > > Any reccommendation for compilers ( c and fortran)? > > Probably a dumb question, but is there any advantage in using the > 64-bit-aware intel compilers on an AMD processor? > > Ed Ed, the Intel compilers work well and give good speed on AMD processors (32 or 64 bit). Apart from the Gnu compilers, there is one alternative: Sun gives out its Sun Studio compilers (version 11 at http://developers.sun.com/sunstudio/downloads/index.jsp, version 12 beta at http://developers.sun.com/sunstudio/downloads/express.jsp) for free - those work (also) under Linux and are more optimized towards AMD processors than the Intel compilers (understandably). I tested them with CNS and other programs and found that, for AMD processors, they produce slightly faster code than the Intel compilers. Furthermore they come with an IDE, and a bunch of analysis and checking tools. Concerning OpenMP (which I'm particularly interested in), they appear to be as good as the Intel compilers. HTH, Kay -- Kay Diederichs http://strucbio.biologie.uni-konstanz.de email: [log in to unmask] Tel +49 7531 88 4049 Fax 3183 Fachbereich Biologie, Universität Konstanz, Box M647, D-78457 Konstanz