From: Nathan Whitehouse (email@example.com)
Date: Sat 08 May 2004 - 01:24:14 EST
Both RZope and Rho solve this problem.
it by using *nix's fork command, which takes as much
time as a memcopy.
Rho(http://rho-project.org) solves it by maintaining
a pool of idling R processes, which it maintains
active & can pull to do a calculation immediately.
Both of these are pretty immediate & allow
multistep/session calculations.(As compared to CGI
Here's a related question: Do any of the mentioned
(Rweb, R-Online, CGIwithR, RSPerl) support reusing the
same R process,
eliminating the startup overhead? This would be
useful to me as well.
Currently I use such a method on my computing cluster:
compute nodes run an R process/compute server that
listens at a socket
connection and subsequent commands from another
When the master process disconnects, the R processes
go back to
at the socket. Connecting to the R compute servers
takes < 2 milliseconds rather than the typical ~2
second R startup
Thanks for any tips.
Baylor College of Medicine
Houston, TX, USA
http://rho-project.org: rho- open source web services for R.
http://franklin.imgen.bcm.tmc.edu: Shaw laboratory, bcm.
Rfirstname.lastname@example.org mailing list
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
This archive was generated by hypermail 2.1.3 : Mon 31 May 2004 - 23:05:08 EST