[kaffe] Re: Re: m68k/jit3 can print "Hello World!"
inaba at src.ricoh.co.jp
Fri Jul 30 21:55:41 PDT 2004
>It would be interesting to see if the difference comes from the time
>spent jitting the code, or from differences in generated code, dealing
>with registers, etc.
I also am so interested why it happens. But (as you know), I have
a business trip next week, and the analysis may be able to do after
My feeling is both parts may be slower than jit, and I am afraid it
should be same for other 'register-rich' CPUes. I know that jit3
is faster than jit for ia32, but how about arm? After my trip, I
plan to make spart-jit3 port and mips-jit port to compare the
performance. Or, shall I make jit/jit3 for arm work? ;-)
P.S. And of course, I am so interested to see the performance of
jit4 for some non-trivial CPUes (including sh3 and m68k).
More information about the kaffe