don@gp.govt.nz (Don Stokes, GPO) (11/24/89)
In article <6808@pt.cs.cmu.edu>, lindsay@MATHOM.GANDALF.CS.CMU.EDU (Donald Lindsay) writes: > My trusty VAX Architecture Handbook states that a process may have at > most 2 GB of virtual memory. This means that the VAX will have maxed > out. To go further would mean that the physical memory would be > larger than a virtual space. IBM is already there! > > That's happened to DEC before - with the PDP-11 and the DEC 20 both. > I hope they, at least, remember how painful it was to all concerned. I don't need my Architecture Handbook to say that 8-) Of course that is 2GB *virtual*. The limits for user code don't change with a change in physical memory, but the more memory, the less paging, and the better the performance - which is the whole purpose of providing lotsa core. They used to build VAXes with 256KB of physical memory, and they still addressed 2GB virtual. If you want a bigger address space, buy a Cray. But my VAX/VMS internals manual gives the page table entry's page frame number size as 21 bits = 2^21(pfn) * 512(page size) = 1GB. Which means, under the current architecture, the machine maxes out at that point. But, that doesn't actually matter terribly, as the pfn database is the OS's problem, not user code - VMS could easily be modified to use a different pfn layout without breaking very much existing code. Of course they could pull filthy tricks and include the owner access mode bits to the page number...... The guts - not a problem. Don Stokes ZL2TNM / / vuwcomp!windy!gpwd!don Systems Programmer /GP/ Government Printing Office PSI%0530147000028::DON __________________/ /__Wellington, New Zealand__________don@gp.govt.nz________ Programming is built on teamwork; it allows you to blame somebody else.
ggw@wolves.uucp (Gregory G. Woodbury) (11/26/89)
In article <593@gp.govt.nz> don@gp.govt.nz (Don Stokes, GPO) writes: >In article <6808@pt.cs.cmu.edu>, lindsay@MATHOM.GANDALF.CS.CMU.EDU (Donald Lindsay) writes: >> My trusty VAX Architecture Handbook states that a process may have at >> most 2 GB of virtual memory. > >Of course that is 2GB *virtual*. > >But my VAX/VMS internals manual gives the page table entry's page frame >number size as 21 bits = 2^21(pfn) * 512(page size) = 1GB. As sometimes occurs, both are correct. The VAX architecture features a split instruction and data space, and there are page tables for each of the spaces. 1 GB of instruction space + 1 GB of data space = 2 GB of virtual address space. This is not to mention the other bit (of course) that indicates that you are accessing the 2 GB of virtual space reserved for the system, where all the real interesting stuff is -- the real memory, the i/o devices and the page tables themselves. -- Gregory G. Woodbury Sysop/owner Wolves Den UNIX BBS, Durham NC UUCP: ...dukcds!wolves!ggw ...dukeac!wolves!ggw [use the maps!] Domain: ggw@cds.duke.edu ggw@ac.duke.edu ggw%wolves@ac.duke.edu Phone: +1 919 493 1998 (Home) +1 919 684 6126 (Work) [The line eater is a boojum snark! ] <standard disclaimers apply>