Cpu

When selecting a new machine, what type of cpu-architecture really makes sense when running win xp + vvvv?
Are multicore processors being supported at all?
And what about compatibility with 64-Bit Systems?

Personnaly we use quad core which is really flexible and powerful.

In vvvv it is usefull for HD video streaming for ex.

The possibility of managing the 4 differents core is usefull too, to attribute diffrents instances of vvvv.

So I would recommend a Core™ 2 Quad Q9550 or better the new i7 Core™ i7 920.

We often use machine with old quad Core™ 2 Quad Q6600 which are really nice .

About 64 bits , no infos !

And what about compatibility with 64-Bit Systems?

We often use 64bit XP, no problems so far.

If you apply a patch to the vvvv.exe you can use up to 4GB of Ram on 64-Bit systems. Just get yourself a copy of Microsofts imagecfg.exe, place it into the vvvv directory, open up the commandline and run the following: “Imagecfg -l vvvv.exe”

Ok, thanks.
But Björn, does 64-Bit XP have any positive impact on vvvv’s performance at all?
I mean, does just run or does it run better/faster?

it just runs. Except for the additional memory I don´t know of any advantages.

added valuable info to System Recommendations

Video is decompressed across multiple cores as are freeframes, s theres another reason…

hmmmm
i wonder if it’s still usefull to patch vvvv.exe on win7 x64?
i try that cmd:
Imagecfg -l vvvv.exe
but return an invalid switch (-l is not in the list)

the -a 0x1 switch works but hmmm

no need to assign vvvv.exe to a specific processor permenantly
since it’s done by design?

There are at least two different versions of imagecfg one that supports the l-switch and one that doesn´t. I googled a little but couldn´t find a valid download source for the exe that supports -l.

But there seems to be another way. (The first step isn´t necessary)

I haven’t tried this yet but catweasel posted about running 2 vvvv instances in separate cores to offload the interface onto its own core… http://vvvv.org/users/catweasel

thanks bjoern & gilbi
added the alternative method to the System Recommendations

64bit wont make a difference, as the application needs to be compiled with a different instruction set (x86-64).
Since every value in VVVV is a 64bit value, there’s a chance there might be some speed increase.
But the real slowdowns aren’t in the arithmetic, so i wouldn’t expect it to make a useful difference.

thanks for your light elliot
i even don’t know how to make a performance/benchmark test in vvvv
to evaluate the difference…