Well it did catch on - just by the time it did it, it was the version in the 386 that was being used...
Memory management was not that different between them, although the 386 had the massive step forward of the "granularity bit" in the descriptor table gate entries that allowed you to indicate that the base and limit P words were to be interpreted as multiples of 4K bytes instead of 16 bytes. That opened up the possibility of defining a 4GB flat memory segment.
the CPU could act on the page fault, but not restart the instruction that caused it, because when the fault occurred it couldn't store enough state to do so. Because of the potential complexity of the instruction, the CPU could have done quite a bit of work before the fault occurred.
I don't know who organised the presentations, but there were two others I went to. One by Intel showing off the 8086, and one by Zilog showing off the Z8000.
In my case, several years to understand.
I do everything in JavaScript and PHP.
IBM used the 8086 in some PS/2 models and in the Displaywriter word processor. I still have my old Amstrad 1512 which also used it. The 8088 which was used in the various IBM PC models until the AT came out is basically the same chip but with only an 8 bit external data bus instead of the full 16 bit bus the 8086 was designed with so it could be used with the more common 8 bit peripherals then available.
Umpteen years ago when I ran Nortons Computer Performance Index which was benchmarked at 1.0 for an XT my Amstrad scored 1.7 which was a healthy increase in performance. However this came mainly or perhaps entirely from the higher clock speed, 8mhz vs 4.8mhz for the XT. I'm not sure if Norton took account of the peripherals or only tested the cpu, probably the latter.
By the time I got a fairly humble Pentium 200mhz a few years later it scored
680 and a modern pc would be several tens of thousands on the same test. It still amazes me though how much we managed to do back then on machines with 512kb of memory, 20mb hard disks and so slow that it would take one a month to just load a modern operating system, not that it possibly could of course.Every increase in computer speed seems to get immediately swallowed up by the bloatware that is then able to run on it and no one bothers to write anything efficiently.
That's why there are rules for writing portable code. Follow them and most of the issues go away.
MBQ
Are you sure? My 1512 had an 80186 in it.
In article , Andy Burns writes
I'm getting used to Gnome3, though whether I *like* it...
See my earlier post.
Giving XP more than 2GB made it do exactly what yours is doing.
Thrash.
Pulling the VM back to 2GB gave me a usable system.
suggest you rip some ram out, or upgrade to a better OS.
No, YOU can't even write portable code.. Many other people have managed quite nicely.
No device driver is ever portable, because its interfacing to a specific piece of hardware.
You are unlikely to stick an INTEL chipset, using INTEL type I/O calls onto a motorola CPU using memory mapped I/O are you.
Mind you, in your case the answer is probably yes..and then blame it on the compiler.
All of the big-endian little-endian problems do, and the problems of integer size.
All of the problems of actual hardware interfacing of course do not - how could they? Only a complete doofus like Dennis would expect them to.
That of course is what you had a BIOS for, and later on, an operating system for.
Yep SUN went Motorola and then SPARC. They were always a 'solid hardware' company and were always BSD UNIX.
I am not sure they ever made an INTEL Sun, although solaris was ported to PC hardware.
In message , "dennis@home" writes
Done my share, yes if you mean circuit design rarely designed the boards, that's more a job for window lickers like yourself
I'm sure you were boasting that you had some time ago, as though it was something special
I don't suppose ANYBODY understood any of the boards you "created"
Is that why you were put out to grass?
I don't really think that we can compare your sad, turgid existence with mine
You just don't get it, do you , you sad old retard
I employ staff so that I don't actually have to work - they work, I get paid. Not something your little mind can understand
I have a 4GB XP (well media centre) system here with a Q9550 in it. Runs very nicely without any thrashing...
We fitted NEC 8086s as they ran faster than the Intel ones.
You sound a bit like the guys from GEC computers I met at a disk conference (yes they had conferences where they discussed disk design). I knew a bit about disks and got volunteered to go there and see what was going on. they were all going on about RLL encoding and really basic disk interfaces. they laughed when I said we were going to go SCSI and didn't care how the data was encoded.
We had to wait 6 months to get our first disk but I know it was the correct decision.
Who used the segmented MMU then? AFAIR the segmented MMU on the 286 was not on the 386. As in you couldn't define segments and what privileges could access them as R/W, etc.
The original C compiler was actually three pass.
But is the database portable?
So you agree its not portable then.
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.