It would be interesting to see metrics on large real world numbers such as 5-10K.
"This e-mail is from Reed Exhibitions (Oriel House, 26 The Quadrant, Richmond, Surrey, TW9 1DL, United Kingdom), a division of Reed Business, Registered in England, Number 678540. It contains information which is confidential and may also be privileged. It is for the exclusive use of the intended recipient(s). If you are not the intended recipient(s) please note that any form of distribution, copying or use of this communication or the information in it is strictly prohibited and may be unlawful. If you have received this communication in error please return it to the sender or call our switchboard on +44 (0) 20 89107910. The opinions expressed within this communication are not necessarily those expressed by Reed Exhibitions." Visit our website at http://www.reedexpo.com -----Original Message----- From: [EMAIL PROTECTED] <[EMAIL PROTECTED]> To: CF-Talk <cf-talk@houseoffusion.com> Sent: Wed Mar 15 16:07:31 2006 Subject: Re: Asynchronous vs linear file processing Thanks Doug. I'll contact you off list for the files to look over and to talk about an article. I've mentioned to Macromedia in the past (and will have to mention to Adobe) that the async processing is such a great tool that every version of CF should have it. Even if pro is limited to 10 threads, it's still a massive tool. What type of hardware did you run this on and what was the memory config? > Just an FYI... > I was so inspired by Michael Dinowitz's talk at CFObjective (and an > internal office debate) that I took the time yesterday to do some time > comparison's between file processing the "old fashioned way" (one line at > a time) vs performing file processing utilizing an asynchronous gateway. > Here are the results... > > The Test > > Configure a CF Event gateway pointing to a cfc that performs record > insertion, calling it "asyncTest" > > Read in an incoming flat file containing 236 records. > > For each row in the file, insert a record into a table. > > For each insert, the datetime stamp at that very instant will also be > inserted into a field in the row so that we know exactly when it was > inserted. > > Upon processing completion, by taking the difference in time between the > first row inserted and the last (min(time) and max(time)), we can see how > long it took to complete the process. > > > The Results > > Processing the old fashioned way: one line at a time. 3 runs were > executed. The average time for inserting 236 rows in this manner was 20.7 > seconds. The page was returned to the user in 21990 ms. > > > > Processing each row asynchronously: each individual row was handed off to > an asynchronous gateway. 3 runs were executed. The average time for > inserting 236 rows in this manner was 1.7 seconds (12 times faster!) . > The page was returned to the user between 67 ms and 90 ms. > > > > Of interest to note is the fact that setting the server's simultaneous > gateway thread count to 260 did not increase the throughput of the thread > executions any more than having it set at 30. > > Just wanted to share that inspiring info. > > > > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~| Message: http://www.houseoffusion.com/lists.cfm/link=i:4:235468 Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4 Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4 Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4 Donations & Support: http://www.houseoffusion.com/tiny.cfm/54
