Path: utzoo!utgpu!jarvis.csri.toronto.edu!torsqnt!tmsoft!masnet!canremote!william.pipher From: william.pipher@canremote.uucp (WILLIAM PIPHER) Newsgroups: comp.sys.ibm.pc Subject: Re: another neophyte Message-ID: <90020304365666@masnet.uucp> Date: 3 Feb 90 09:18:00 GMT Organization: Canada Remote Systems Limited, Mississauga, ON, Canada Lines: 89 bolton@cg-atla.UUCP (Lee Bolton) bP>Orga: Agfa Compugraphic Division, Wilmington, Mass. USA asks: bP> Included with the system, was a utilities disk with 3 bP>important bP> programs on it. bP> MICEMM4B.SYS a (sic) memory manager. bP> RAMDRIVE.SYS a disk emulator. bP> RELOCATE.EXE a ROM to RAM translator. bP> device=c:\bin\micemm4b.sys 0 (256) /ERROR bP> device=c:\bin\ramdrive.sys 256 /a << --- what is 'a'? bP> First, RELOCATE.EXE is supposed to copy BIOS from ROM to RAM. bP> neat. The little 4 page document is a bit too terse on the bP> subject. Like it doesn't happen to mention just what exactly bP> gets moved, or from whence to where. a couple of sentences bP> mumble about BIOS, EGA BIOS, and/or a 64K frame buffer?? Normally with a 386, the ROM bios can be copied to some address above 1 meg in extended memory. Then, the superior memory management of the 386 is used so that the bios now in RAM appears to the system to be where it always is. The effect of this is a) there will be less extended memory available for other tasks and b) your computer could run faster. Mine does. Usually this is set-up defined in extended cmos registers, rather than by software, although Desqview QEMM uses the software approach, too. See if you can implement this Bios SHADOWING through your system set-up configuration. bP> MICEMM4B.SYS is supposed to make the top 384K available as bP> extended memory. I presume that you mean that MICEMM4B can place an expanded memory page frame into some "space" between 640K and 1 meg. If my presumption is correct, and if the driver is too stupid to find its own address to map it's frame, you'll have to help. Do you have NORTON SI? (or similar utility)? This is what mine says: DOS reports 703 K-bytes of memory: A search for active memory finds: 640 K-bytes main memory (at hex 0000-A000) <<< space in use 128 K-bytes display memory (at hex A000-C000) <<< space in use 32 K-bytes extra memory (at hex C000-C800) <<< too small 84 K-bytes extra memory (at hex CF00-E400) <<< O.K. put page 960 K-bytes expanded memory here CF00-DF00 ROM-BIOS Extensions are found at hex paragraphs: C800 CB80 CC00 bP> live there. MICEMM4B.SYS is supposed to take a pile of bP> parameters. First is a number describing a starting address bP> as counted in paragraphs from the bottom of what's referred to bP> as 'discontiguous memory'. I have no idea what they might be asking for (number of paragraphs from A000:0? from B000:0? from C000:0? who knows?). An intelligent response would be to use CF00 (as per above example, or whatever), and hope that that's what they meant. bP> The second parameter is supposed to be the size of the pool to bP> be managed (in bytes). Presumably this is asking for the total size of extended memory to be allocated to the expanded driver (I'm mind-reading again). This will be size of extended memory minus size of shadowed rom minus size of ram drive minus 10000 hex for the EMS page. bP> I cannot get the ramdisk and the relocated BIOS to co-exist. bP> No matter how I set up the memory manager, it faults out bP> with a memory protection error when I invoke the relocator. bP> It's not unreasonable to think it possible to put BIOS bP> wherever it has to go and let the ramdisk have the rest is it? Should be OK. I would expect that the ROM shadowing should be enabled first, then the expanded memory driver, then finally the Ramdrive. Try specifying /E as an arg to ramdrive.sys, not '/a'. This will place it up in extended memory out of the way: device=c:\bin\ramdrive.sys 20 128 /E ******** continued in next message ****************** --- ~ DeLuxe 1.11a19 #3744 william.pipher@canremote.uucp