Re: Speed of writing in Mathematica 3.01
- To: mathgroup@smc.vnet.net
- Subject: [mg11815] Re: Speed of writing in Mathematica 3.01
- From: bruck@pacificnet.net (Ronald Bruck)
- Date: Tue, 31 Mar 1998 02:28:33 -0500
- Organization: Univ of Southern Calif
- References: <199803260809.DAA06287@smc.vnet.net.> <6fi3j6$4nl@smc.vnet.net>
In article <6fi3j6$4nl@smc.vnet.net>, Sean Ross <seanross@worldnet.att.net> wrote: > James Bridges wrote: > > ... > > > > 23 seconds to write 1001 integers?! On a 233MHz G3! Everything else is > > whiff-bam fast, but writing is glacial. Now this is a simplified > > example of what I need to do but it contains one of the factors, namely > > that I have to individually write out elements of a list, hence the Do > > loop. Any hints? >> Reading and writing files has always been slow in Mathematica. The time > to write or read seems to depend on the numbers of objects Mathematica > is expected to deal with. Your example does seem abysmally slow. The > same code requires 0.22 seconds to execute on my Win95 pentium 166. > Usually, Mac's are faster than PC's. I wonder if there is not some > conflict in your system or something misallocated among system > resources that mathematica depends on. Granted that various speedups work, you should STILL check the BinaryFiles package. I haven't used it in awhile, but when I did I found it was much faster. I seldom use this, because most of the time my files have to be used by other programs (or vice versa, other programs prepare input for Mathematica). Still, there is something seriously wrong with Mr. Bridges' computer and/or Mathematica setup. On my PowerTower Pro 250, his program took just a little over 1 second. To be sure, this system has an ultrawide SCSI controller and drives, and 256MB of physical RAM gives Mathematica some room to move, but still... --Ron Bruck
- References:
- Speed of writing in Mathematica 3.01
- From: bridges@en.com (James Bridges)
- Speed of writing in Mathematica 3.01