> > Thank you Bryan - I was just consulting Ben Forta's book on > > cflock and to quote "Forcing code to single threaded use only > > can seriously impact system performance." Do you think this is > > an issue in this case? > > Not necesarily. I ran a regression test where I used different > locking methods to read 10,000 application variables (same locking > issue as session variables) and the results were somewhat > surprising; using one CFLOCK tag outside the loop it took 450 > ms while using 10,000 individual locks inside the loop took > only 600 ms. I performed this experience using CF Server 4.5 > on my laptop (Pentium 3 with 256 RAM). As you can see, the > effect of one lock or even of 20 simultaneous locks is > practically nill.
I'm assuming from the above that you're using a single script with a loop inside it, and you're calling that script once from your browser. If my assumption is incorrect, I apologize in advance.
You'll want to be extremely careful about using looping-based tests to determine server performance. Typically, they don't really tell you what you want to know. Rather than running one script which does something 1000 times, a much better test would be to have one script run by 1000 concurrent users - it'll tell you a lot more about real-world performance.
Dave Watts, CTO, Fig Leaf Software http://www.figleaf.com/ voice: (202) 797-5496 fax: (202) 797-5444
CF-Talk
Dave Watts