I'm trying to figure out how to benchmark performance for transcoding multiple videos at once using ffmpeg. I have no experience with this.

Does anyone have ideas or suggestions for how to do that? Or know were to read up on it?

I searched forever and couldn't find anything that helped. One guy asked the identical question on some Q&A board and got zero responses, which was disappointing to say the least.

Also, ffmpeg is a strange beast. It's the most powerful video encoding software in the world. It's used by Vimeo, supposedly by YouTube, and every transcoding service on the internet. Yet there still seems to be a lack of meaningful coverage. The official documentation is ok (and improving), and there are some decent answers on SO, but that's about it. No established best practices, few tutorials, hardly any discussion. That seems odd.

videohelp.com is pretty solid usually

Any chance someone wants to take a crack at an issue I'm having? I'm trying to get a value from a check box through an AJAX script -- should be very simple, but something is going wrong. First time I've ever posted a question on Stack Overflow and not had it answered within minutes. Posted it yesterday and only had one answer and the person deleted it after I pointed out their flawed logic.

Anyway, you can find the question here. I will be forever grateful if someone knows the issue!

Answered it for you

You are awesome, apparently I should've posted here before wasting several hours yesterday Thanks!

Dont play a lot of poker these days, but the Full Tilt welcome back bonus is almost too good to pass up on....

Dunno mine yet, have to clear the 600$ midyear first heh

it all depends on the effective rate the bonus clears at though. it ends up being equivalent to rakeback, usually at a slightly better rate than normal rb. their new flat rb rate is 25% -- i'd be suprised if this bonus is worth more than 35% rb.

not at all sure, but I don't think bonus deducts from rakeback any more?

if that's the case i stand corrected
but my standard assumption is that all poker rooms are incorrigibly stingy

Well if you grind out the bonus..better do it right away. I think there will be fish with money stuck on the site who will play it fast but not reload...then it'll be a regfest.

I found this poster amusing:

http://csillustrated.berkeley.edu/PDFs/posters/50-CS-Quotes-poster.pdf

Some samples:

"Perl – The only language that looks the same
before and after RSA encryption."

"In My Egotistical Opinion, most people's C
programs should be indented six feet
downward and covered with dirt."

Link no worky

Hi Guys,

Can anyone recommend an ad service... I need something that is very quick to setup (i.e. no long approval process), and displays mobile ads (via a webapp).

Thanks

Anybody have good advice for load testing?

I've used jmeter and selenium before and both kind of blow.

For really basic, throw x requests at this page with y simultaneous connections, AB is what you want: http://httpd.apache.org/docs/2.2/programs/ab.html

I want something a little higher level. I'd like to be able to write up a few common use cases and have them run over some period of time.

You can send data (GET or POST) with AB... not sure about running over a period of time, but this would be relatively easy to script.

This still feels super low level (maybe I'm misunderstanding?).

As a really simple example I might want to login, do an action that sets up some state and returns an ID (so say creates Foo object with fooId), than perform a second action that uses that ID (so go to page /foo/fooId).

In a more complicated example I might actually want to do this on a page that creates Foo dynamically with an ajax request.

What about Canoo?

I didn't even see it as an option.

I'm testing out The Grinder. It seems like what I want (pretty good feature set) but allows me to write jython if necessary. My main complaint from using JMeter was that you were stuck doing development in a GUI and using horrible regexs everywhere.

Could you not achieve that with a set of Unit Tests?

I'm not sure what you mean. We have tests that check for correctness - but I want to do lots of these actions at the same time and see how the system responds.

So basically we have a web app. The web app does some stuff and also sends jobs to a backend machine (well group of machines) that does work asynchronously. So a use case is something like a user logs in, starts a task, checks the progress, and does something when the task finishes.

I want to have lots of "people" doing this at the same time over some period of time. We want to know where things break - the app server, our worker machines, on certain tasks, etc.

Ah, ok... I see what you mean.

JMeter is the only tool I know of which does this, but I think you have already mentioned it.

For my class in "Computer Graphics" i had to choose a subject to write/program about.
I chose "OpenGL for Android", now i have to come up with an actual example which i'm going to program.
So if someone has any ideas (doesn't have to be super simplistic) about what i could do with this subject i'd like to hear them!