OSM Performance- Not sure what forum to put this in...
Hi all,
I know that the OSM player is pre-release, but I'm concerned about what ultimate performance it is capable of attaining. If my experience is typical, I am very hesitant about deploying with this framework. Hopefully, most of the issues I'm seeing are caused by debugging hooks that are still in the code.
When I installed it on a test Drupal site and started playing videos, they were really choppy, so I decided to do a little benchmark using the Mediafront.org front page (to make sure it was not my set-up or configuration) and different browsers and platforms. This is obviously not a controlled test, but it does shed some light on the current status of the player in my environment.
Just as an FYI, we have a commercial grade 100 Mb internet connection (Cogent Communications fiber) and both test systems branched from the same gigabit switch.
Here are the very rough CPU utilizations for the two platforms I could immediately test. They are not "state of the art" systems but do mirror pretty well what a lot of users may have.
Is anybody else seeing similar issues? If not, I guess I'll have to get out my exorcism paraphernalia.
Thanks,
Rich Beyer
_____________________________________________________
Win XP Pro SP3/ Pentium 4, 2.8GHz, 1GB RAM (Flash 10.1 r82)
IE 8.0.6001.18702
HTML 5 & MP4- Right around 50%
Vimeo 75%-80% Choppy video, audio OK
YouTube 40%-50%
Generally, IE 8 had a decent and useable video playback performance. Buffer indicator did not advance during pause for Vimeo(or match Vimeo's indicator). No way out of full browser display in YouTube- controls disappear. Also, controls do not redraw when exiting full browser mode in other formats.
-----------------------------------
Safari 5.0 (7533.16)
HTML5 80-99%
MP4 90-99%
Vimeo 95+%
YouTube 90+%
For the most part, the playback was terrible-lots of dropped frames . Buffer indicator did not advance during pause with any format.
Firefox 3.6.8
HTML5 99%
MP4 40%
Vimeo 40%
YouTube 50-70%
Playback was choppy. Buffer indicator did not advance during pause on Vimeo.
________________________________________________
_________________________________________________
iMac G5, 2.1GHz, 1.5GB RAM OS X 10.4.11 (Flash 10.1 r53)
Safari Version 4.1.1 (4533.17.8)
HTML 50%
MP4 55%
Vimeo 80-85% Choppy Video and audio
YouTube 80-85%
----------Buffer indicator did not advance on pause except for YouTube
-------------------------------------------------------------------------------------------------------
Firefox 3.6.8
HTML 90+%
MP4 80%
Vimeo 90% choppy video and audio
YouTube 80%
!!!!! --- With Firefox, the logo layer flickered terribly on all formats except for HTML5.
As a comparison, I played the Richard Stallman Video (not sure if it is exactly the same one) directly from the YouTube page and the CPU% was around 55% and very good quality.
___________________________________________________
Travis--you are very welcome. It is the least I can do to help in whatever way that I can. Open source is not "free" in terms of your time, effort and money, so it is the responsibility of anyone that makes use of these wonderful creations to give back in some way.
If there is any specific testing that I can help with, please feel free to contact me at the email address listed in my forum account. I like trying to "break" things. I have access, at home, to newer versions of Windows (Vista and Win7 on dual and quad-core system)) as well as Mac OS X 10.6 (Mac Intel duo-core) and the older pentium and G5 systems at work.
Thanks for all the great work that has gone into this software.
Have a safe move. I'll look for the update in a few weeks.
Regards,
Rich
With all the new consumerization of IT across the globe more people are buying PC's, smart phones and tablets and it shouldn't take too long for actual courses are mandatory (to use, diagnose PC problems, installing software etc).
Sponsors
|
|
|
Rich,
First of all let me just extend my sincerest appreciation for your time on doing this. This is a HUGE help, and definitely gives me some insight as to what other poeple are experiencing with the OSM Player. My goal is to make this player top notch, and this certainly helps.
One thing that I will say off the bat is that it is really hard for me to determine which performance issues are related to the code I wrote vs. the software that is running the media. I am curious if the media events that this player handles is bogging down the player since there are several things that need to happen on every media update event. One way that we could verify this is to completely remove all media event handling ( where you would not get any progress indication in the control bar ), and see if that improves teh performance. If it does significantly, then that will tell me that I am doing something processor intensive in the handling routines.
Again, thanks for your efforts. I will let you know when I have an engineering build ready ( moving out of town so it may be a couple of weeks ).
Travis.
International Open Source Solutions
What can Alethia build for you?