I've been following this thread (somewhat) and I just have to comment on the following:
Jose wrote:
> Well, after doing that, it is still consuming a lot much more CPU than
> CGI-WIN.
> How do I know that ? Using NT Perfomance Monitor, running the servlet,
> I see peaks of 100 % of CPU. Running the same transaction again, same data,
> via CGI-WIN, I'll have at most 50 % of CPU (56 %, to be more accurate).
Without further information, what this indicates to me is that the CGI-WIN I/O
implementation is not as efficient as the JVM I/O implementation. The big question is
not whether the CPU is at 50% or 100%, but how long does it take to read the file?
Personally, I like it when the computer works harder, and I don't have to wait as
long... ;-)
Bottom line: If the CGI-WIN actually reads the file faster, fine. Regardless, you
are letting only one of many indicators of system performance misguide your thinking.
CPU utilization on a well-balanced system is not typically a significant factor in I/O
performance.
Bart Cassady
Database Analyst
Corporate Technology Services
Associated Engineering Group Ltd.
[EMAIL PROTECTED]
___________________________________________________________________________
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff SERVLET-INTEREST".
Archives: http://archives.java.sun.com/archives/servlet-interest.html
Resources: http://java.sun.com/products/servlet/external-resources.html
LISTSERV Help: http://www.lsoft.com/manuals/user/user.html
