Ahhh right. Quite interesting that it does not return random data under Windows. But yea, it should be initialized to zero. Would love to get my voice in on that request.
On Thu, Mar 31, 2016 at 10:46 AM, Lucy Wilkes <[email protected]> wrote: > Hi Mads, > > The behaviour you see on Linux is actually what I would expect. The output > image you declare in your Blink kernel will not be initialised, so any > pixels you do not write to can contain random data. Also, writes are not > atomic so if you write to the same output position multiple times, you will > get non-deterministic results since the processing order will depend on > your compute device. > > For the output initialisation, there's an existing feature request (which > I actually logged myself when I was experimenting with random writes): > > 159236 > <https://foundry.tpondemand.com/RestUI/Board.aspx?acid=08FCEF261C0AA0070B326053812D4CBA#page=bug/159236&appConfig=eyJhY2lkIjoiNUE2ODU3NUVFN0M1MzA0QzM5QTA5MENERkU0MjVFQkUifQ==> > : BlinkScript - add option to initialise destination image to zero > I suggested that random or ranged access output images should be > initialised to zero by default. Please let me know if you would like me to > add your voice to the request and I'll try to get the priority raised. > > In the meantime, if you put your renderer into a plug-in using the Blink > API rather than having it in a BlinkScript node, you would be able to > initialise your output image to zero. You would also be able to use images > with integer data, for which atomic writes are supported. I know it's not > as convenient as a BlinkScript, but I think it would solve these problems. > > Regards, > Lucy > > On Wed, Mar 30, 2016 at 10:32 PM, Mads Lund <[email protected]> wrote: > >> I have created a little custom 3D renderer using the Blinkscript node to >> do super fast voxel and point rendering. And it works perfectly in Windows, >> however in Linux I get a issue if I write to the same outputbuffer twice or >> leave some of the outputbuffer empty. (Causes corrupt output data) >> >> Does anyone know a way around this bug? >> >> >> -- >> Best regards. >> Mads Hagbarth Lund >> >> _______________________________________________ >> Nuke-users mailing list >> [email protected], http://forums.thefoundry.co.uk/ >> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> > > > > -- > Lucy Wilkes > > Senior Software Engineer > The Foundry > 5 Golden Square, London, W1F 9HT > Tel: +44 (0)20 7479 4350 > Web: www.thefoundry.co.uk > Email: [email protected] > > > The Foundry Visionmongers Ltd. > Registered in England and Wales No: 4642027 > > > _______________________________________________ > Nuke-users mailing list > [email protected], http://forums.thefoundry.co.uk/ > http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >
_______________________________________________ Nuke-users mailing list [email protected], http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
