Pegasys TMPGEnc 3.0?

[Leoslocks]

Some news at xBIts leads me to believe that 3.0 may be out second quarter this year.

. . .it is not clear when the apps with tweaks for SSE3 actually emerge.

[Baldrick]

http://www.pegasys-inc.com/en/press/04_0204.html

[marvel2020]

Well it looks like Pegasus has jump into bed with Daddy Intel giving a 2 finger salute to AMD owners.

[triphop]

Its somewhat contradictory and heres why:

TMPGENC is a reasonably priced product (IOW its downright inexpensive compared to some of its competition) and here they go mating it up to a processor that is only just being launched and will probably carry a 20% price premium over competiting Athlon chips. I guess with the premium that Intel charges for its CPUs, you don't have much cash left over for high-end encoding software

How hard is it to simply produce two or three versions (recompiled from same source code (with #ifdef/#endif blocks [nasty!] or static linked processor dependant code)) for each of the major processors. The main executable then simply runs the processor specific module that best matches the processor in the customers machine.

But then I never understand all this marketing stuff anyway. You might find that Intel has paid Pegasys to do this in a very similar fashion that they paid PC manufacturers to have that stupid "Intel Inside" sticker on thier boxes.

[FOO]

Nobody said that it WON'T WORK on AMD did they ?
Nobody (sane) could sell software like that

[triphop]

Nobody said that it WON'T WORK on AMD did they ?
Nobody (sane) could sell software like that

Some time back CCE crashed on AMD Athlon processors (known "feature") and Premier Pro requires SSE instructions - this counts out some of the older Athlon chips.

[MrMungus]

How hard is it to simply produce two or three versions (recompiled from same source code (with #ifdef/#endif blocks [nasty!] or static linked processor dependant code)) for each of the major processors. The main executable then simply runs the processor specific module that best matches the processor in the customers machine.

Pretty hard when people don't pay for the software they use.

[triphop]

Pretty hard when people don't pay for the software they use.

Sure - of course. But should I buy a piece of software that will not run optimally on my processor? And I am not talking about warez.

[andkiich]

We found that when the SSE2 instructions are enabled with the Opteron and Atholon64 processor, execution speed decrease. The reason is if SSE and SSE2 floating point instruction and integer instruction are used in a situation where they are dependant each other, performance dramatically decrease*. This is a particularity of the Opteron and Athlon 64 processor and it is possible a change in the future version brings improvements. In each parts of TMPGEnc are codes using such instructions, especially the DCT or NR filters.

This is from the article that Baldrick linked. So it does appear that it will work with the AMD chips, but since it is not optimized for them, then it will run slower. And there is potential for "future version brings improvements".

[Leoslocks]

I see that TMPGEnc has an update for 2.5 up now also. I have been looking for This.

[CFster]

Why bother. CCE is much faster, and gives better DVD results in my case...

[MrMungus]

Sure - of course. But should I buy a piece of software that will not run optimally on my processor? And I am not talking about warez.

Whether you use it or buy it or not is up to you. My point is that maybe they'd have more $ for development for your processor if more people paid for it. That's all. Don't make me break out my nunchucks.