Problem with fps

Like I said before don’t use SetProcessAffinity.

ah, so thats means just to use multiple instances ? ( sorry a little bit lost)

Yes just start multiple instances and let windows figure out how to deal with them.

ok; thank you. about DOPE i m using the work done by clement in contribution. if there is possibility to multi thread it it would be really really great

Well, you’ll have to ask @digitalwannabe about this. Or if you have some budget, ask some one else to look into it (maybe @woei ), it’s open source: https://github.com/digitalwannabe/d.o.p.e.

merci bjoern !!!

hi @bjoern , im pretty fighting with Sharedmemory DX11 2d node. Not arriving to have my image, despite i think structuration is ok and little test to send basic texture were ok.

I open my main, where there is the receiver and all variables job, then i launch 2d instance of vvvv where i send the final result as texture. The allocation pointer is sended in shared memory from the 2nd instance ( the player), being not the same at each opening
I believe that there is something not working with that pointer. Have you any advice ?
thxs

hey, sneaking in the conversation, have you considered using Spout for that?

hi sebescudie, no sneaking you are welcome !
i m using dx11.
global image needs to be 2,5 times a full HD for zooming inside. i would like to avoid any conversions. not sure SPOUT could handle it ?

To use Spout with dx11, you’d need the Spout dx11 pack. Looks like your setup would work :

image

ok, i will try it :) I used a long time ago SPOUT so lets see !!!
thanks a lot for this nice suggestion :)

Hum… the SPOUT sender cant create a share handle. does anyone have a solution ?

“texture not shared. no shared handle available from the connected texture”.

Where does your texture come from? I had the same error message when trying to connect a FileTexture directly to the Spout Sender. Guess the texture has to somehow rendered, using a TempTarget for instance.

spoutkaristouf

i m normaly using renderer mrt dx11.
i have put on side my project just to understand whats going on. because memory is very very busy.
so taking things from scratch:
from a same instance of vvvv the spout seems to work ok.
but using 2 different instances, now i have a ok for receiving and a very big lag ( refreshing every 3-5 seconds). I m suspecting teh shell execute option, or my spout setup exe placed in my dropbox . or the firewall ?

Re Christophe,

Just setup a quick test here : InstanceOne.v4p creates a Spout sender. When the patch start, it waits for five seconds and then starts the second instance that just has the receiver. Worked fine for me here!

I don’t think firewall has any incidence there.

Does the attached example help?

ChristopheSpoutMultiInstances.zip (7.0 KB)

Hi @karistouf, great to see your project growing!

As for your fps, I’m pretty confident this is not Dope causing your low framerate - please note that after you have computed the weights for your object you can store the output values, eg as a csv, and load it from there. You can therefore safely remove the BoundedBiharmonicWeights node. But even if you leave it in there it shouldn’t matter as long as you compute all the weights on load of the project and do not trigger it again later (as it is NOT async and will block the main vvvv thread…). I still recommend saving the weights to csv for consistency.
After computing the weights it’s only shaders, which doesn’t seem to be your problem. Where is your mesh coming from, Triangle? Note that this can/should also be precomputed, as indicated by the horizontal line in the help patches. Also the IndexedGeometryBufferTriangleList node (below the line) should only be set OnOpen:

We do have an interactive dope project running for a few years now with large images, lots of fine meshes, bones, cages etc. - it takes quite a while to load all the weights, meshes and images into ram/vram, but after that it’s solid 60 fps…

Best, Clemens

Hello Clemens, @digitalwannabe yes putting everything in io box is already done :) thank you !. The part with file loading is still there, untill i finish last chapters.

@sebescudie just succeed now to make liaison with spout using renderer target. BUT, its a very very ugly framerate ( 4 to 5 max).
is there a better way ? or should i re go on shared memory thing ?

thanks to all for your gentle help !!!

that does not sound right, just a spout sender should not be responsible for that :( do you get such a low framerate using my example? Both running at solid 120 fps here.

yes your sample is working.
well i have 6 chapters to test under the hand, and many shared values memories to try to dispatch ressources and not break all automations ( sounds, sensors, animations so on). ok i reboot and go back to see what is happening…

hum, ok rebooted machine.


… same lag in texture sharing