Here's a question for you about encoding on the server. If two videos are encoding at the same time. Do they share the processing power. lets say we have 2 vids same size 1 encodes at 1 minute and then if 2 are going at the same time does it encode them both in 2 minutes? Thanks, T On Fri, Aug 13, 2010 at 3:08 PM, Ashley Sheridan <ash@xxxxxxxxxxxxxxxxxxxx>wrote: > On Fri, 2010-08-13 at 14:48 -0600, Tristan wrote: > > i have a list of people that uploaded their videos waiting to be encoded. > > these are the line items for the db > > id | member_id | created | status > > I just want to make sure that everyone gets a turn processing. > > I was thinking WHERE DISTINCT(member_id) and created (oldest) > > but i see some flaw in that. > > > if someone uploaded 20 vids and then someone else uploads 30 vids they'll > still have to wait until the 20 finish. > > Almost have to keep a list of the users being processed in another table and > if someone else comes along append them in. > > say we have 5 people that uploaded multiple videos. I want it to go like > > sandy, tom, jim, harry, star > > and then back to > > sandy, tom, jim, harry, star > > then if someone else comes into the mix. add them to that queue > > Thanks, T > > > You can probably do it with the one video list in the DB. Each time your > video encoder has a free spot it should obtain a list of videos from this > table from the last time period, both processed and waiting to be queued. > Look up the first one from this list and check to see if a video for the > same user has been processed within a specific time period. If yes, then > move to the next video in the list and check that in the same manner, and > repeat until you find one. If you don't find any, then you can just pick any > video from the list. > > You should be able to process several videos at the same time on a decent > server, and a 64-bit system should make mincemeat of an encode as it should > do that type of number crunching more quickly. > > Thanks, > Ash > http://www.ashleysheridan.co.uk > > >
