> I have seen that in 64-bit Linux, 32-bit processes tend to execute > a bit faster. I've seen a number of cases where 32-bit apps run > faster than their 64-bit version, but I've also seen 64-bit apps > run faster than their 32-bit versions.
The 1g/3g memory split I mentioned allowed the kernel to "share" the page tables with every process rather than using a dedicated supervisor page table, obviating the page table switch (and the associated TLB flushes, etc) that would otherwise be incurred with every crossing of the User/Kernel boundary. Since the x86_64 kernels do use a dedicated supervisor page table I have to assume that some syscall-intensive 32bit apps do run a bit slower than they would on an "equivalent" (whatever that means) x86 kernel. _______________________________________________ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
