iDVD encoded our 10 minutes of DV footage in 9 minutes. This was nearly twice as fast as the P4/2.53.
This sounds like a software issue, not a hardware one. I bet there is better DVD encoding software.
But does it really matter anyway unless you are encoding dozens of discs? You can just start the encoding process and do something else for 20 minutes.
This is news because it was severe enough to back out of the release.
It's like an automotive recall or toy choking hazard: DHTML could explode and make you go blind if you keep using Mozilla 1.2. Several Tripod users and viewers of garish movie promo sites have already sustained serious mental frustration. We are working with local authorities to prevent riots.
There are two new garbage collectors you can use starting with JDK 1.4.1. The Concurrent GC may reduce the pauses. The switch is -XX:+UseConcMarkSweepGC. I have been using the parallel one (-XX:+UseParallelGC) with good performance results on server programs.
The SR-71 was built with titanium to withstand the heat it generated at mach 3+ and that metal was new at that time for aircraft. Since the titanium had to expand there were ground leaks. Why don't you design a mach 3 jet that can fly at 85000 feet and isn't such a "hack"?
The more serious problems btw, are that it can't operate in the rain due to its coating I believe and its heat signature is huge.
Bear in mind that this is a 1960s era aircraft designed without CAD/CAM and it's still pretty damned impressive IMO.
Whoa there, buddy. Did you just start a sentence with a conjunction and end the very same sentence with a preposition? Have you no shame? There may be children here for goodness' sake.
I don't even have a site so what do I know? I'm a pothole on the information superhighway. I do know that Slashdot has been propagating bad spelling and grammar for years. Don't even get me started on its/it's and their/there/they're. It's not that difficult, people!
Yeah, WebLogic's CORBA stuff handled local calls more efficiently, but it still wasn't like a function call because of the overhead you mentioned. It's good that omniORB uses a different mechanism. That sounds pretty cool -- thanks for the info.
Sorry for the confusion, I didn't mean CORBA itself allowed batching, but that it could be done at a higher level in the application. At a simple level, you could batch keyboard events and mouse events (at the expense of smooth movement). Yes that would suck. (I just realized that CORBA overhead for mouse events might be really bad.) I think I would prefer a binary TCP protocol like X but higher level.
The only useful form of communication that CORBA supports is synchronous.
Well, you could do one way calls with callbacks, or you could create an event queue on the client side to batch up API calls before doing a CORBA call. I don't see this as a CORBA-specific issue or a fundamental problem any more serious than how a TCP/IP based protocol would have to deal with asynchronous issues.
I think 20ms is on the cynical side. On a good local network it should easily be < 5 ms, don't you think?
I realize that Fresco is high level and vector-based. The pixel question is just to see what the most extreme overhead would be. But then again, how would Fresco support games running in a window?
I'm just not sold on the idea of using CORBA for a component model in this manner. Gnome does this too so it's not a new idea to me. I have read many arguments, but I'm still skeptical. Why can't I have a proxy API that makes local library calls or CORBA calls, depending on what is needed? A language that doesn't want to call native code can use CORBA. There are also some "philosophical" issues about the realities of network transparency as I mentioned in another post.
I understand the design reason, but in practice it seems like you'd mostly be rendering to your local screen. In that case network transparency is just a performance burden.
Besides, we could debate whether network transparency even exists since local and remote resources are fundamentally different (network glitches don't affect local resources, and you generally need retry and error logic for networked resources).
I'm not trying to dis Fresco here, just think about the design tradeoffs. The problem with the X protocol is that it's low level, so even though it's a more efficient TCP-based protocol, you would be sending many more low level packets. In the end it could break even with the IIOP and marshalling overhead of CORBA since Fresco is high level.
You know I've been a computer hobbyist for many years and am interested in science and so on. I very rarely find "geek humor" funny, especially parodies and anything that I see repeatedly. I guess the technology in it is usually so superficial that it doesn't seem clever or witty.
Then people keep posting it. I often wonder if the demographic of "geek humor" is really freaks rather than people who have intellectual interests.
Technically Linux is just a kernel. You can't do much at all with just a kernel. You could print AAAA from one process and BBBB from another, but who would want to do that?
It seems more likely to me that AIX would distingusish IBM from its rivals. Everyone runs Linux.
It's possible that AIX is encumbered by technology that IBM licensed, but I don't know. I'd have to think IBM has ported AIX to x86. Likewise, I wonder why SGI hasn't chosen to port IRIX which has mature support for large scale multiprocessing.
Slashdot Mirror
This sounds like a software issue, not a hardware one. I bet there is better DVD encoding software. But does it really matter anyway unless you are encoding dozens of discs? You can just start the encoding process and do something else for 20 minutes.
-Kevin
-Kevin
It's like an automotive recall or toy choking hazard:
DHTML could explode and make you go blind if you keep using Mozilla 1.2. Several Tripod users and viewers of garish movie promo sites have already sustained serious mental frustration. We are working with local authorities to prevent riots.
-Kevin
-Kevin
-Kevin
-Kevin
-Kevin
Here's more detail:
tuning JVM v1.4.x+
The more serious problems btw, are that it can't operate in the rain due to its coating I believe and its heat signature is huge. Bear in mind that this is a 1960s era aircraft designed without CAD/CAM and it's still pretty damned impressive IMO.
-Kevin
-Kevin
Whoa there, buddy. Did you just start a sentence with a conjunction and end the very same sentence with a preposition? Have you no shame? There may be children here for goodness' sake.
-Kevin
I don't even have a site so what do I know? I'm a pothole on the information superhighway. I do know that Slashdot has been propagating bad spelling and grammar for years. Don't even get me started on its/it's and their/there/they're. It's not that difficult, people!
-Kevin
-Kevin
-Kevin
Sorry for the confusion, I didn't mean CORBA itself allowed batching, but that it could be done at a higher level in the application. At a simple level, you could batch keyboard events and mouse events (at the expense of smooth movement). Yes that would suck. (I just realized that CORBA overhead for mouse events might be really bad.) I think I would prefer a binary TCP protocol like X but higher level.
Well, you could do one way calls with callbacks, or you could create an event queue on the client side to batch up API calls before doing a CORBA call. I don't see this as a CORBA-specific issue or a fundamental problem any more serious than how a TCP/IP based protocol would have to deal with asynchronous issues.
I think 20ms is on the cynical side. On a good local network it should easily be < 5 ms, don't you think?
-Kevin
I'm just not sold on the idea of using CORBA for a component model in this manner. Gnome does this too so it's not a new idea to me. I have read many arguments, but I'm still skeptical. Why can't I have a proxy API that makes local library calls or CORBA calls, depending on what is needed? A language that doesn't want to call native code can use CORBA. There are also some "philosophical" issues about the realities of network transparency as I mentioned in another post.
-Kevin
Besides, we could debate whether network transparency even exists since local and remote resources are fundamentally different (network glitches don't affect local resources, and you generally need retry and error logic for networked resources).
I'm not trying to dis Fresco here, just think about the design tradeoffs. The problem with the X protocol is that it's low level, so even though it's a more efficient TCP-based protocol, you would be sending many more low level packets. In the end it could break even with the IIOP and marshalling overhead of CORBA since Fresco is high level.
-Kevin
(Okay, actually I think CORBA is gross, period.)
-Kevin
It's the ALPHA preview. Get it??
-Kevin
-Kevin
-Kevin
-Kevin
-Kevin
It's possible that AIX is encumbered by technology that IBM licensed, but I don't know. I'd have to think IBM has ported AIX to x86. Likewise, I wonder why SGI hasn't chosen to port IRIX which has mature support for large scale multiprocessing.
-Kevin