> Quoting Michael S. Tsirkin <[EMAIL PROTECTED]>: > Subject: Re: [PATCH] libmthca: optimize calls to htonl with constant parameter > > > So what is different in your setup that causes this patch to make a > > difference for you? > > Hmm. I agree it is somewhat strange. > > Below is a simple test that attempts to compare htonl, CONSTANT_HTONL, > and an array-driven implementation. The code line is taken directly from > htonl. > Could you compile and run it please?
OK, this was stupid, the test was missing #include <netinet/in.h> so htonl was expanded by a gcc intrinsic which seems to work worse than the macro tricks present in netinet/in.h. I guess this include got killed on the test system somehow, and this explains why I saw a difference in libmthca. -- MST _______________________________________________ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general