I think the meaning of this is related to compiler flags and optimisation. While compiling the binaries, you have the option to pass to the compiler the type of CPU and optimization levels. For example, you can pass the cpu type as athlon XP and enable certain CPU extensions. Same with P4. These are not enabled so that slower systems can also run the OS.
The people who build the binaries have to make sure that the binaries run on a 486 system also and that the OS should run on the minimum RAM mentioned. If you obtain the source code for windows and compile the OS with custom compiler flags, the system should run faster. But that doesnt mean that the OS wouldnt do any enhancements if it found a faster processor and lots of RAM. Runtime detection handles a lot of stuff. But you can avoid a runtime CPU detection by passing the CPU type to the compiler.
I run an LFS 5 system built on my Athlon XP and I can tell you I havent seen any operating system running faster than this. The programs run much faster than on other linux OSes. To fully optimize your LFS system, you should read more on compiling the base system with optimization flags. glibc and gcc can also be optimized to obtain faster performance.
|