Hello Mani,
Re: Intel's long awaited 64-bit architecture
I must be getting old <hehe>. Back in the days of 16-bit operating systems, it was a real challenge to do things: unsigned integers size up to 65535, 64K of code space, 64K of data space. I remember doing some pretty radical things to work around such constraints. The previous world of 8-bit machines was even worse with their ability to count all the way from 0 to 255.
During the last days of DOS, I was amazed to be able to get to the "land" of flat 32-bit memory by using DOS extenders. Today we have the 32-bit luxury of unsigned ints which can exceed 4 billion in value, if we want more memory from the OS, we just ask for it and it's available in huge quantities.
Now the buzz is 64-bit processors and OS's. As an application programmer, I for one am not drooling in anticipation. The removal of integer value limitations of 65535 and replacing with limitations over 4 billion in value allowed me to loop anything I've ever needed to loop. It allows me to "count" everything except the national debt. The removal of code and data segments and their replacement with hassle free flat memory was a great and worthy achievement.
I'm sure there's need for 64-bit machines and applications, but I can't see the 64-bit world revolutionizing things like the 32-bit world did. 32-bits just seems to fit mankind's needs.
When Intel moved from the 8-bit world to the 16-bit world, their 16-bit chips (the 8088???) could execute the old 8-bit programs that existed at that time. That eased the transition to the 16-bit world and let 16-bit applications be written in due time. When the movement from 16-bit to 32-bit came along (80286 to 80386???), a similar approach was taken and once again, that approached worked.
I'm surprised that Intel chose to try the quantum leap to 64-bit OS's and 64-bit applications for their Itanium(applications written with highly optimized compilers that don't now exist). Could it be they forgot their past? What happens to the Itanium if the 64-bit applications don't reach critical mass? Will there be enough server only demand to justify this chip?
I seems to me the only explanation for the quantum leap approach to 64-bit is to once again reinvent computing, i.e. by changing the landscape so fast that the competition will be left in the dust. Intel's competitors spent most of the past in Intel's dust for similar reasons.
As an AMD long I really like AMD's SledgeHammer approach to 64-bits. To me it makes as much sense as Intel's approaches from 8 to 16-bit and from 16 to 32-bit have the past. I'm glad we got it!!! I'm even surprised we got it!!! But the really good new is we got it!!!
Kindest regard,
CrazyMan |
