We recently found a problem where the wrong target architecture was being used on newer servers. The code used by the driver to auto-select the architecture, get_auto_cpu_name(), used the processor name read from /proc/cpuinfo first and only fell back on the family number when the name was unknown. The problem is that the names are used for branding purposes: an "Opteron(tm) Processor 6174" (part of the "Magny-Cours" family) is more closely matched by the Open64 "barcelona" target architecture than by "opteron".
This patch changes get_auto_cpu_name() to prefer the family number over than the processor name for AMD processors. Could a gatekeeper please review the change? I have attached a copy of the patched function for convenience. -David Coakley / AMD Open Source Compiler Engineering
cpuinfo.diff
Description: Binary data
get_auto_cpu_name-patched.c
Description: Binary data
------------------------------------------------------------------------------ Start uncovering the many advantages of virtual appliances and start using them to simplify application deployment and accelerate your shift to cloud computing. http://p.sf.net/sfu/novell-sfdev2dev
_______________________________________________ Open64-devel mailing list Open64-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/open64-devel
