some things i need to say:
about dual-cpu:
me and AntiPasta think that the best use for the second core is a translation core & cache so the overhead of decoding instructions will be in the second core,the problem is that we can`t access the internal cache of the second core to use it,as it should be , a cache.
in the Dynamic recompiler area there is alot of use for the second cpu if we get more control on it.
about the diffrence between cpus/arch(s)/whatever:
the Pentium core(most correct the whole X86) is cisc while the arm is a risc proccessor,this allow for more optimizaion at asm level for example few x86 vanilla instructions can be replaced by one MMX or SSE command.
also the arm doesn`t have a FPU so all fpu-based calculations are made in software emulation/softfpu which is slow.
to use better the arm cpu we need to things:
1.use more asm,arm asm is one of the most powerfullls as it support condional executin ,that means that a cmp(compare command can be described as a if in high-level language) can be include into the normal instruction,this means that two instructions become one and that means that the speed of that operation is now two times faster !!!
2.get rid of fpu calculation as this means a huge speed ups
the pentium also got better compilers(gcc is alot better on x86 than in arm).
something that you can try is to use Intel compiler which is pretty much compatible with gcc and it will work faster(on most intel machines it run faster coz it is optimized for intel)
Uriel