cytron@uiucdcs.UUCP (04/27/83)
Relay-Version:version B 2.10 gamma 4/3/83; site burl.UUCP Path:burl!spanky!ihnp4!ixn5c!inuxc!pur-ee!uiucdcs!cytron Message-ID:<1959@uiucdcs.UUCP> Date:Tue, 26-Apr-83 23:26:10 EDT #R:uiucdcs:27800005:uiucdcs:27800006:000:750 uiucdcs!cytron Apr 26 11:32:00 1983 I think the strategy used in the BCC500 was also used in a computer developed at the University of Hawaii. Here's the swapping idea: When a page fault occurs, all the user's pages are dumped to a contiguous portion of secondary storage (disk), thus freeing up lots of main memory. At the end of one disk rotation, the page which caused the fault is found, and swapped in, and then the pages that were swapped out en masse can be swapped in and the user can run again. This scheme requires high I/O bandwidth, but little main memory. The first reference I have to it is the BCC500 which appeared in the base note. It looks like the Univ. of Hawaii may have picked up on it, too. Anybody know of any references? Thanks. ...pur-ee!uiucdcs!cytron
wapd (04/27/83)
Would someone please post a one-paragraph description of what the BCC500 is and who makes it ? Never heard of it. Bill Dietrich houxj!wapd
kepecs (04/28/83)
leave the discussions about swapping in net.singles