RevengeFNF Posted March 6, 2015 Posted March 6, 2015 Just to check if someone noticed this also. Some charts using New Relic on my server with average response time of IPB 3.4 and IPS 4.0. IPS4 seems to have a much higher average response time, 10 times higher in fact. This on the same server using the same database and the same resources. Can someone replicate this?
Makoto Posted March 6, 2015 Posted March 6, 2015 Is this on a fresh installation or an upgrade? Have you allowed the background conversion tasks to complete if it's an upgrade?
RevengeFNF Posted March 6, 2015 Author Posted March 6, 2015 Its an upgrade from a backup of my live board. Background Tasks are all completed.
Management Lindy Posted March 6, 2015 Management Posted March 6, 2015 What apps do you have installed on your suite?
sobrenome Posted March 6, 2015 Posted March 6, 2015 I was expecting a much better performance from IPS 4 compared to IPS 3.4.7.
RevengeFNF Posted March 6, 2015 Author Posted March 6, 2015 What apps do you have installed on your suite?Just forums and calendar.
Will Munny Posted March 6, 2015 Posted March 6, 2015 I've noticed significantly longer server response times. Previously I was consistently getting <= 0.2 seconds. It's now between 0.3 and 0.5 seconds.
sobrenome Posted March 6, 2015 Posted March 6, 2015 Probably there are some bugs that still need attention.
Izaya Orihara Posted March 6, 2015 Posted March 6, 2015 On my own naked eye and based on user experience, I haven't noticed much difference in either personally. I will be looking into it personally.
RevengeFNF Posted March 8, 2015 Author Posted March 8, 2015 Ok, now instead of measure the performance with data from my server, lets measure the time my browser takes to load the "View New Content".I'll use a plugin for Chrome called App Telemetry. IPB 3.4.7 IPS 4.0 Invision IPS 4.0 So, my site with IPB 3.4.7 load in 0,64 seconds, which is blazing fast. My test site with IPS 4.0 load in 1,76 seconds, which is still pretty fast, but we do not have that feeling of blazing fast. Invision IPS 4.0 site load was even slower than my server, with 2,08 seconds to load. My live and test site use the same server with exactly the same resources for which one. The database is the same.
esquire Posted March 8, 2015 Posted March 8, 2015 On this site I'm noticing significantly longer times to view last posts from within individual forums which list the last post of each thread/topic. Not sure why that should be the case but it is.
Gnuru Posted March 8, 2015 Posted March 8, 2015 I can confirm longer server reaction times compared to 3.4.7, especially if there blocks like newest threads on the page.Google pagespeed complains too...
sobrenome Posted March 8, 2015 Posted March 8, 2015 It seems that IPS 4 is indeed slower than 3.4.7...
asigno Posted March 9, 2015 Posted March 9, 2015 Yep it's slower. I'd have expected lots of other issues which were also in the 3.4.x to have been resolved too such as the amount of http requests.
Management Charles Posted March 9, 2015 Management Posted March 9, 2015 Yes it is slower. We have not denied that. In fact you can see on our Feature Plan that performance is the next thing we are doing. We are aware of many areas we can make much better.
RevengeFNF Posted March 9, 2015 Author Posted March 9, 2015 Yes it is slower. We have not denied that. In fact you can see on our Feature Plan that performance is the next thing we are doing. We are aware of many areas we can make much better.In this case its seems its a problem with slower php execution. I get fast responses from mysql. So IPS4 seems very well optimized in terms of queries to the database.
Skipy7 Posted March 11, 2015 Posted March 11, 2015 In this case its seems its a problem with slower php execution. I get fast responses from mysql. So IPS4 seems very well optimized in terms of queries to the database.From my basic tests does seem that way, would also be nice if IPS consolidated the rendered out CSS and JS files as we have it now there are 17 TCP requests just in that.
Recommended Posts
Archived
This topic is now archived and is closed to further replies.