Post by thread:What use is 32 bits ? [Was MPLAB Win CE]

At 10:02 AM 8/2/97 +0200, you wrote: >> What makes you think 32 bit apps are more efficient ? >> >> Windows 95 is 95% 16 bit drivers and libraries most of which is ported >> >> over from win 3.11. >> >> there is a lot of propoganda and hype Win NT comes close to be a 32 >> bit app >> but even that uses 16 bit dlls ! >> >> rdgs >> >> mike > >So is there a TRUE 32-bit OS out there? Does it matter ? Why FORCE one instruction per 32 bits when you can get TWO or more ? WHY do you NEED an 8 bit character to occupying 32 bits - when its more efficient for it to occupy a byte or at most a 16 bit word ? >What will Win97 be? (If ever it surfaces from the rumours!) Not important - Actually it will be Win98 and WILL have lots of 16 bit instructions in it (even though the buss might fetch 2 or 4 of them at a time). Why slow down a machine forcing it to use 32 bit instructions ? Did you know that 8 bit micros have faster IO than any 486 or Pentium in a windows type GUI - apart from operating system overhead - most comms is 8 bits - so the value of a wider word size is a overdoing it, a bit;) Rgds Mike Perth, Western Australia Some say there is no magic but, all things begin with thought then it becomes academic, then some poor slob works out a practical way to implement all that theory, this is called Engineering - for most people another form of magic. Massen

>>So is there a TRUE 32-bit OS out there? > >Does it matter ? I don't really care if 95 is a *true* 32-bit OS or not. But there are some things I like about it. (And some I don't...) I like not having to worry about memory models and whether I'm going to wrap around a 64k segment and how big my pointers are and whether it's a near call or a far call. I also like that the 32 bit address space gives you room for neat tricks like memory mapped files. >Why FORCE one instruction per 32 bits when you can get TWO or more ? It's not the simple. Not every x86 protected mode instruction is 32 bits wide. Not every instruction on the 8086 was 16 bits wide, either. It's my understanding that a 32 bit OS has a 32 bit address space, and that the size of the opcodes is irrelevant. later, newell