>>7
You negate Intel Atom chips before the 2- series, which do not support long mode. However you do have a point, seeing that the last Atom to operate exclusively in 32bit mode was the N2- series (which came out in early June 2008, or more than three years ago). Perhaps every x86 model processor developed after 2008 supports native 64bit mode, yet many choose not to develop 64bit software. Why you ask? Because the majority of machines on the market currently, albeit having 64bit capabilities, come pre-installed with 32bit versions of Windows or other such operating systems. Many developers do not feel the need to write 64bit exclusive applications out of virtue of compatibility, much the same way that many software developers continued to write 16bit code even after 32bit processors became available. The philosophy there being, "Why go through the trouble of reimplementing previous software with 64bit capabilities when we can continue to improve our current implementation?" Companies such as Microsoft have released 64bit versions of their newer Operating Systems at higher costs to exploit the obvious advantages in speed and processing capabilities that 64bit processors offer purely out of profit-motive. It may also be possible that such companies believed that 32bit mode would become obsolete when 64bit processors hit the market, much the same way that 32bit processors made 16bit code obsolete. The truth is that 64bit implementations existed within supercomputer architectures from the late 1970s however these implementations were not suitable for mainstream home computer use. It was with the prolific release of the Intel 80386 that made 32bit registers the consensus among developers. Of course at the time that these architectures were conceived current understanding was so far beyond the possibility of more than 4GB of physical RAM that 32bit registers remained the norm until the early 90s, when obvious reductions in physical memory price and size led to memory implementations approaching 4GB. In response to this, both MIPS and DEC developed 64bit native architectures (mainly for server and high-end machines) although Intel had already begun it's own implementation of the 64bit model during the late 1980's. In reality, compatibility is the restriction amongst both system and application developers due to the present nature of both 32bit exclusive machines and software implementations. I project that 32bit code will remain precedent amongst the majority of computers for at least another three or four years. For all intensive purposes, however, I support your position on the development and proliferation of 64bit code. Whether we like it or not, keeping Moore's law in mind, 32bit code will eventually become extinct to the extent of how 16bit code became extinct so long ago.