Messages in this thread | | | Date | Sat, 24 Jul 1999 15:36:57 -0500 (CDT) | From | Oliver Xymoron <> | Subject | Re: real-time threaded IO with low latency (audio) |
| |
On Sat, 24 Jul 1999, Bill Huey wrote:
> > > Oliver Xymoron wrote: > > > On an unloaded system, yes, you can be clever with mlock() and pre-fault > > everything before going into your critical code, but the people who are > > begging for RT performance for multimedia stuff don't understand that it > > means running with basically no load and giving up tons of memory and not > > touching the disk, etc.. > > Uh that's pretty typical arrogant *not-so-smart* Unix mentality. > Try capturing a 200 megabyte video stream without touching the disk. > > You'll fail flat out.
There's only one very small piece of video capture that needs to be realtime - ie worst case performance of a few milliseconds. That's the piece that sets up the DMA transfers for grabbed frames. That piece belongs in a driver, not an application. The remainder, ie queuing and unqueuing buffers, postprocessing, and spooling to disk, all can be done in application space and rely on the system's buffers. If it can't make it to disk without overrunning memory in that environment, it wouldn't make it anyway.
(currently working on a megapixel digital "video" system myself)
-- "Love the dolphins," she advised him. "Write by W.A.S.T.E.."
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |