This is the mail archive of the gdb-patches@sources.redhat.com mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [rfa/mips] Second go at vr5500 hilo hazard fix


At Thu, 25 Mar 2004 17:14:18 -0500, Andrew Cagney wrote:

Look at it this way, if the igen mechanism is used, gcc is able to
eliminate everything :-)


In the case of a single-architecture sim (i.e., not 'multi', then
yes), GCC will automatically eliminate the test.

In the case of a multiple-architecture sim, it won't.

It still does (or at least can), trust me :-)


 But (a) at
least in my mind, i'm not worried about absolute maximal performance
from that configuration (if you want max performance, build a single
arch sim), and (b) until somebody quantifies the difference, I'll
claim that i don't believe it'll be a measurable difference.  8-)

But yes, you won't measure the difference.


Andrew



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]