Interesting. Couple of questions. What sort of processes did you find to be slower? Was it all pages or stuff that used cfquery or cfloops of cffile or whatever? Were you using the same web server? What load server did you use? Full version of CFMX or developer Version ? (anyone know if dev version uses SMP?) Which OS ? How many CPUS ?
Again I'm not arguing against your statement, just like to know more about this stuff. Cheers Justin -----Original Message----- From: Paolo Piponi [mailto:[EMAIL PROTECTED]] Sent: 12 February 2003 17:55 To: '[EMAIL PROTECTED]' Subject: RE: [ cf-dev ] CFMX is second class Oh, yes. Absolutely. In fact, I had seen many reports beforehand of poor CFMX performance and I wondered myself how much was down to first time performance. Unfortunately, my tests were based on real comparisons after compiling. Unless, of course, CFMC decided to recompile some of the 40-odd cfm pages that I access on each hit. If so, same goes. Paolo -----Original Message----- >From: Justin MacCarthy [mailto:[EMAIL PROTECTED]] Sent: 12 February 2003 17:40 To: [EMAIL PROTECTED] Subject: RE: [ cf-dev ] CFMX is second class I'm not saying your are wrong. Just wondering how you tested this? I presume you discounted the first time the page was loaded? Thanks Justin -----Original Message----- >From: Paolo Piponi [mailto:[EMAIL PROTECTED]] Sent: 12 February 2003 17:40 To: '[EMAIL PROTECTED]' Subject: [ cf-dev ] CFMX is second class After a month upgrading a 4000 page application (1% pages needed fixing) and testing and after spending 2 days trying to install CFMX on our development server, I have come to the conclusion it is too slow. I even backed out of CFMX and ran CF5 once again on the same machine and my results show that CFMX is between 15-30% slower. I give a range here because whilst CF5 was consistent in its reporting, CFMX was very erratic and frequently reported very poor times. Although not what my decision was based on (nor the figure above) when monitoring the activity, a web page run on CF5 peaked at 16% CPU and the same page on CFMX peaked at 32% - on the same machine. The benefits of upgrading (some bug fixes, CFCs, etc.) were not important enough. I was expecting for at least the same performance and was hoping for something better. I am about to report back to my 'superiors' that CFMX is a no-go. However, after sending countless messages to you guys bugging you for support I thought I might pass it back to you one more time for comment. Paolo PS Unless a future upgrade recognises this performance issue and we have another window of opportunity we will not be upgrading for a very long time. It seems that Macromedia have been so obsessed with competing for all round integration they have forgotten what they are here for - to chuck out some HTML. If this is the direction Macromedia are now going, although very reluctantly, we might end of ditching CF, at least for some projects. CFMX is a second-class web application. I seem to remember a company by the name of Netscape that went down a similar road. -- ** Archive: http://www.mail-archive.com/dev%40lists.cfdeveloper.co.uk/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] For human help, e-mail: [EMAIL PROTECTED] -- ** Archive: http://www.mail-archive.com/dev%40lists.cfdeveloper.co.uk/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] For human help, e-mail: [EMAIL PROTECTED] -- ** Archive: http://www.mail-archive.com/dev%40lists.cfdeveloper.co.uk/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] For human help, e-mail: [EMAIL PROTECTED] -- ** Archive: http://www.mail-archive.com/dev%40lists.cfdeveloper.co.uk/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] For human help, e-mail: [EMAIL PROTECTED]
