home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!olivea!mintaka.lcs.mit.edu!zurich.ai.mit.edu!ara
- From: ara@zurich.ai.mit.edu (Allan Adler)
- Newsgroups: sci.math.symbolic
- Subject: Re: long symbolic runs
- Message-ID: <ARA.92Sep1200345@camelot.ai.mit.edu>
- Date: 2 Sep 92 01:03:45 GMT
- References: <JAFFER.92Aug29231912@camelot.ai.mit.edu> <1850@nikhefh.nikhef.nl>
- Sender: news@mintaka.lcs.mit.edu
- Organization: M.I.T. Artificial Intelligence Lab.
- Lines: 23
- In-Reply-To: t68@nikhefh.nikhef.nl's message of 1 Sep 92 07:46:33 GMT
-
-
-
- I would not rule out the possibility of backing up 200 MB or more. If one has
- a tape drive, one can save stuff in pieces in files of a certain size and
- then back them up on the tape drive, then delete them to make more room
- for files to hold more pieces. As long as one has enough cartridges
- and is willing allow the time to back up to a tape, one can handle
- gigabytes.
-
- On a SUN, the tape drive is reasonably efficient. On the PC's I've seen
- that have tape drives, they are as slow as molasses.
-
- In reply to the original question about what algebra systems provide
- facilities for automaticaly saving computations, I think, but am not
- sure, that some authors of REDUCE packages may have toyed with this
- feature. My reason for thinking so is that the source code for the
- Groebner package and the Distributed Polynomial package says something
- about the checkpointing stuff having been deleted. I'm told that
- the practice of periodically saving partial results to disk is called
- checkpointing.
-
- Allan Adler
- ara@altdorf.ai.mit.edu
-