battle@umbc3.UMBC.EDU (Rick) (09/12/90)
For your entertainment; The 80386 can run in the virtual 8086 mode. Is it outrageous to assume that Intel may build the 80586 to run in the virtual 80386 mode? If so, each session could address 16 Meg of RAM and with a fast enough cpu clock, one computer could be a workstation and a file server, IE. Novell, all at the same time. Any comments?
37KGLLQ@CMUVM.BITNET (Tony Papadimitriou) (09/13/90)
>The 80386 can run in the virtual 8086 mode. > >Is it outrageous to assume that Intel may build the 80586 >to run in the virtual 80386 mode? > >If so, each session could address 16 Meg of RAM and with >a fast enough cpu clock, one computer could be a workstation >and a file server, IE. Novell, all at the same time. > >Any comments? The sad thing is that each 80386 virtual mode would also have to support the virtual 8086 mode and that means DOS will never leave us alone :-)
paulc@hp-lsd.COS.HP.COM (Paul Carroll) (09/14/90)
> Is it outrageous to assume that Intel may build the 80586 > to run in the virtual 80386 mode? > > If so, each session could address 16 Meg of RAM and with > a fast enough cpu clock, one computer could be a workstation > and a file server, IE. Novell, all at the same time. > > Any comments? Yeah. The 80386 (DX) addresses 4 Gigabytes of memory, not 16 Megabytes (that is, 32 address lines, not 24). You must be thinking of either the 80286 or 80386SX. I would figure an 80386 or 80486, running in protected memory with paging turned on, would meet the criteria you indicated above. The secret is to get the OS to use the capabilities of the chip. Does anyone know if the SYS V Unix that runs on the 386 uses these capabilities? Or does it run in a strictly flat memory space? The real question is: what would you do to the 80586 or 80686 to make it stand out from the 80486? Just adding more peripherals wouldn't seem to be enough. Is added speed the only thing left without requiring a completely new architecture or OMF? ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ + Paul Carroll "I don't believe there is a single + + HP Logic Systems Division man, woman, or child alive in + + paulc%hp-lsd@hplabs.hp.com America today that doesn't enjoy a + + lovely beverage!" - David Letterman + ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
toma@tekgvs.LABS.TEK.COM (Tom Almy) (09/14/90)
Well now that people are purely speculating on the 80586, which will have real mode, virtual 8086 mode, 80286 protected mode, 80386 protected mode, virtual 80386 mode, and protected 80586 mode, how about adding a 68000 mode to appeal to those folks that think you can only write programs for a Motorola processor. Tom Almy toma@tekgvs.labs.tek.com Standard Disclaimers Apply :-)
kaufman@delta.eecs.nwu.edu (Michael L. Kaufman) (09/16/90)
In article <7680003@hp-lsd.COS.HP.COM> paulc@hp-lsd.COS.HP.COM (Paul Carroll) writes: > The real question is: what would you do to the 80586 > or 80686 to make it stand out from the 80486? Just The thing to do is to make the 80586 virtualize itself. Have a v386 mode would be cute, but would not solve any great problems. Having a v586 more would solve most of the currect ones. Michael
feustel@well.sf.ca.us (David Alan Feustel) (09/17/90)
Intel could provide an option permitting each segment to have its own
page table.
--
Phone: (work) 219-482-9631 MCI mail: DFEUSTEL BIX: FEUSTEL
E-mail: feustel@well.sf.ca.us {ucbvax,apple,hplabs,pacbell}!well!feustel
USMAIL: Dave Feustel, 1930 Curdes Ave, Fort Wayne, IN 46805-2710danielce@ecr.mu.oz.au (Daniel Ake CAROSONE) (09/17/90)
In article <7680003@hp-lsd.COS.HP.COM>, paulc@hp-lsd.COS.HP.COM (Paul Carroll) writes: > you indicated above. The secret is to get the OS > to use the capabilities of the chip. Does anyone > know if the SYS V Unix that runs on the 386 uses these > capabilities? Or does it run in a strictly flat memory > space? When I saw the details of the i386 and i486 adressing, segmenatation, virtual memory within segments, etc. I said to myself: "This chip is designed to run MULTICS." All we have to do is get rid of messy-dos :-) For those who are not aware: MULTICS was an operating system designed by CS people with a 'wishlist'. It was given up on, as the technology at the time was not good enough to run it. UNIX started life as a single- user much-simplified version of multics, with a much simpler (flat) memory model. Now is the time to resurrect it!.
wsinpdb@svin02.info.win.tue.nl (Paul de Bra) (09/17/90)
In article <7680003@hp-lsd.COS.HP.COM> paulc@hp-lsd.COS.HP.COM (Paul Carroll) writes: >> Is it outrageous to assume that Intel may build the 80586 >> to run in the virtual 80386 mode? ... What we need is a chip, say 80n86, which has a virtual 80n86 mode. This would mean that if the system is running operating system A, you can have a virtual operating system B running on it, or vice versa, and both will see a real 80n86 processor. So running DOS under Unix would yield a true 80n86 dos session for instance (but you probably won't want to run dos on such a machine anymore). If this is too difficult (wouldn't surprise me), now that we already have an 80386 and 80486 I would even settle for an 80n86 which offered virtual 80(n-1)86 capabilities. My first hope for the 80486 was that it might offer a virtual 80386 mode, or even only a virtual 80286, but Intel obviously didn't think that was important, though many people have expressed the need for this. If a new chip offers nothing more than just a higher speed, many users are not going to care, as what is available now is more than fast enough for most of us. Paul. (debra@research.att.com)