Re: [Nuke-users] Nuke ignoring GPU?
Hi Nick, If Kronos is unable to process on the GPU, it will fall back to the CPU to finish processing. The most likely reason for this is that the GPU ran out of memory, but it could be that some other error occurred. When this happens it will have to start the render again from the beginning on the CPU, so it will take longer than rendering on the CPU from the start would have done. I think this fallback could be what you're seeing; you should be able to confirm (or otherwise) by watching the CPU activity while the GPU-enabled Kronos is processing. If it is really running on the GPU, the CPU use should be minimal. If the GPU is not being used, I'd recommend you send your script and machine specs in to Support so they can investigate why not. Regards, Lucy On Mon, Dec 5, 2016 at 5:38 AM, Nick Guth <nick.g...@gmail.com> wrote: > I only have a single GPU on this machine (980Ti). Still trying to figure > it out, but no combination of settings seems to make it process faster than > my CPU. > > ___ > Nuke-users mailing list > Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ > http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users > -- Lucy Wilkes Lead Software Engineer - Nuke Engine The Foundry 5 Golden Square, London, W1F 9HT Tel: +44 (0)20 7479 4350 Web: www.thefoundry.co.uk Email: l...@thefoundry.co.uk The Foundry Visionmongers Ltd. Registered in England and Wales No: 4642027 ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
Re: [Nuke-users] Creating a custom renderer using blinscript
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=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 <madshl...@gmail.com> 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 > Nuke-users@support.thefoundry.co.uk, 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: l...@thefoundry.co.uk The Foundry Visionmongers Ltd. Registered in England and Wales No: 4642027 ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
Re: [Nuke-users] user knob arguments (colour, size….)
Hi, The Shuffle node uses a Custom knob with its own custom UI. I'm afraid you would need to do the same in order to change the colour and appearance of knobs (such as a checkbox). Regards, Lucy On Thu, Mar 17, 2016 at 10:43 AM, Ben Dickson <ben.dick...@rsp.com.au> wrote: > I'm not aware of a way to, say, colour the checkboxes of a Bool_knob, via > the Python API nor the NDK. The Shuffle node presumably uses a special knob > type (the implementation of which would have access to the Qt widgets to > style them specially)? > > You could potentially make a custom knob type, but this is likely far more > work than it is worth (I have not tried this, but I wonder if you would > lose things like expression linking and on?) > > On 17/03/16 14:15, Info Mail Bag wrote: > >> hey Darren >> >> cheers >> >> but is this not more to do with setting knob values like blur size x,y >> values. >> >> I am looking some thing to be able to set GUI values :: sizes, colours >> fonts and the like. >> >> or is this done though various python modules? >> >> >> On 17/03/2016, at 2:36 PM, Darren Coombes <darren.coom...@me.com >> <mailto:darren.coom...@me.com>> wrote: >> >> >>> http://www.thefoundry.co.uk/products/nuke/developers/63/pythondevguide/basics.html >>> >>> * >>> *Thanks.* >>> *Darren Coombes* >>> ** >>> * >>> * >>> *Check out some of my work...* >>> * >>> * >>> *www.vimeo.com/darrencoombes <http://www.vimeo.com/darrencoombes>* >>> * >>> * >>> * >>> * >>> * >>> * >>> *Mob: +61 418 631 079 <tel:+61%20418%20631%20079>* >>> * >>> * >>> *Skype: darrencoombes* >>> *Twitter: @durwood81* >>> * >>> >>> On 17 Mar 2016, at 2:26 PM, adam jones <adam@mac.com >>> <mailto:adam@mac.com>> wrote: >>> >>> Hey all >>>> >>>> I am looking for the documentation regarding userKnobs. >>>> >>>> Although it would be cool to get a list of all user knob arguments. >>>> for gizmo building >>>> >>>> I am currently looking for the info on how to colour the BG / >>>> interior of a check box, like the shuffle node. >>>> >>>> >>>> >>>> regards >>>> -adam___ >>>> Nuke-users mailing list >>>> Nuke-users@support.thefoundry.co.uk >>>> <mailto:Nuke-users@support.thefoundry.co.uk>, >>>> http://forums.thefoundry.co.uk/ >>>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>>> >>> ___ >>> Nuke-users mailing list >>> Nuke-users@support.thefoundry.co.uk >>> <mailto:Nuke-users@support.thefoundry.co.uk>, >>> http://forums.thefoundry.co.uk/ >>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>> >> >> >> >> ___ >> Nuke-users mailing list >> Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ >> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> >> > -- > ben dickson > 2D TD | ben.dick...@rsp.com.au > rising sun pictures | www.rsp.com.au > ___ > Nuke-users mailing list > Nuke-users@support.thefoundry.co.uk, 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: l...@thefoundry.co.uk The Foundry Visionmongers Ltd. Registered in England and Wales No: 4642027 ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
Re: [Nuke-users] Email not in spam test - please respond
Hi Howard, It looks as though Gmail is the common factor. Just to let you know that we use it here and so your messages might not be getting through to anyone at The Foundry... They go to spam for me - just checked and there were a few in there! Regards, Lucy On Mon, Mar 14, 2016 at 5:08 AM, Carlos Trijueque <li...@ctrijueque.com> wrote: > A little progress. It doesn't go straight to Spam but Gmail told me that > it is because of my filters (I have your email address whitelisted) and > still flags it as Spam. > > ./charlie > ___ > Nuke-users mailing list > Nuke-users@support.thefoundry.co.uk, 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: l...@thefoundry.co.uk The Foundry Visionmongers Ltd. Registered in England and Wales No: 4642027 ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
Re: [Nuke-users] Performance mode makes GUI performance terrible?
Hi Jose, I'm not surprised you are seeing this. In performance mode there will be extra synchronisation (locking) happening in order to update the timers. The more threads you are running with, the more of a performance hit you will see. I would expect the effect to be especially pronounced if you are running Nuke on a machine with dual (or more) processors, as the synchronisation will be particularly expensive there. This mode is indeed intended for trouble-shooting. Personally I find it most useful for command-line renders. The difference in Nuke 9 is that we fixed several issues with the performance timers to make the reporting more accurate. This could well explain why you now find the performance mode to be slower. For example, any nodes with calls to sample (for example Transform, LensDistortion, ScanlineRender) were not being accounted for properly prior to Nuke 9, so you are now likely to see more locking with these nodes than before. Regards, Lucy On Sat, Mar 5, 2016 at 1:23 AM, Jose Fernandez de Castro < pixelcowbo...@gmail.com> wrote: > Just tried it adding the -n flag but it still slows down to a crawl. Is no > one else seeing this with medium to big scripts? > > On Fri, Mar 4, 2016 at 5:15 PM, Randy Little <randyslit...@gmail.com> > wrote: > >> All postage stamps turned off? >> On Mar 4, 2016 4:33 PM, "Jose Fernandez de Castro" < >> pixelcowbo...@gmail.com> wrote: >> >>> I know it's for troubleshooting, but when I'm in Performance mode I'm >>> not even able to do that as I can barely navigate the DAG, and even >>> connect/reconnect nodes without everything freezing for a few seconds. The >>> data output of course is useful, but it would be better to actually be able >>> to debug interactively. >>> >>> On Fri, Mar 4, 2016 at 12:29 PM, John Mangia <j...@johnmangia.com> >>> wrote: >>> >>>> I believe that mode is intended more for troubleshooting of slow >>>> scripts rather than daily use. If you didn't use performance metrics you >>>> could probably figure out where things are slowing down by looking at the >>>> usual culprits (large bbox, scanline render, vectorblur, furnace, convolve, >>>> etc) >>>> >>>> Sent from my iPhone >>>> >>>> On Mar 4, 2016, at 3:19 PM, Deke Kincaid <dekekinc...@gmail.com> wrote: >>>> >>>> You could just render a frame in the shell and spit out an XML with the >>>> performance metrics using the "--Pf filename" flag. Then you just have >>>> data to create metrics from rather then having to look through the >>>> nodegraph manually. >>>> >>>> On Fri, Mar 4, 2016 at 9:11 AM, Jose Fernandez de Castro < >>>> pixelcowbo...@gmail.com> wrote: >>>> >>>>> Has anyone noticed that enabling Performance metrics makes Nuke really >>>>> slow? When I enable it on a medium to heavy script, the information on >>>>> performance mode is mostly useless, as I can barely navigate the DAG as >>>>> everything freezes and updates too slowly. This didn't use the case with >>>>> the early implementations of Performance mode, so I think something might >>>>> be causing the GUI to slow down heavily. I rarely use performance mode >>>>> this >>>>> days because of this issue. Anyone else seeing this? >>>>> >>>>> -- >>>>> Jose Fernandez de Castro >>>>> >>>>> ___ >>>>> Nuke-users mailing list >>>>> Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ >>>>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>>>> >>>> >>>> ___ >>>> Nuke-users mailing list >>>> Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ >>>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>>> >>>> >>>> ___ >>>> Nuke-users mailing list >>>> Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ >>>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>>> >>> >>> >>> >>> -- >>> Jose Fernandez de Castro >>> >>> ___ >>> Nuke-users mailing list >>> Nu
Re: [Nuke-users] Nuke locking up with 1 cpu busy
Hi Daniel, I've answered on your forum thread. I also wondered whether expressions might be causing this ( http://community.thefoundry.co.uk/discussion/post.aspx?f=189=120025=1008064 ). If you have a script and repro steps, please do send them in to our Support team so that we can take a look. Regards, Lucy On Fri, Mar 4, 2016 at 4:34 PM, Elvis Au <el...@blueskystudios.com> wrote: > Hi Daniel: > > This is a shot in the dark without seeing your actual script but the > behavior matches something I've seen before. Do you have expressions/python > code that execute anytime there is an update to the tree? You may need to > check your script editor to see this. If there is code being executed every > time there is an update to the tree any where that will also explain why it > gets worse as the tree gets bigger (more nodes = more code to chew through) > And as far as I know, that coded is being processed by one proc. > > I put a request for Foundry to look at this a while back: > > Bug 35703 - Expressions evaluated in unrelated parts of tree > > Development gave the following details in response: > > "...this is a known bug. The root cause is the way Nuke handles > expression links between knobs. There is currently no way of knowing if > changes in any part of a script will affect the result of a particular > expression without evaluating that expression." > > > In the past to get around this, I've copied over the roto node to an empty > script only to copy it back after I've made my edits. > > > Elvis > > > > > > - Original Message - > From: "Daniel Hartlehnert" <dah...@gmx.de> > To: "Nuke user discussion" <nuke-users@support.thefoundry.co.uk> > Sent: Friday, March 4, 2016 11:00:16 AM > Subject: [Nuke-users] Nuke locking up with 1 cpu busy > > Hi, > > sorry for cross-posting this from the forums, but i hardly got any > response there. > I tend to get this behavior in Nuke quite often with bigger scripts and it > just gets unusable. > The symptom is Nuke completely locking up when dbl-clicking a node to open > its properties. Usually it happens when dbl-clicking a roto or rotopaint > node. > Looking in the system monitor, i can see that 1 (out of 12) cpu’s is 100% > busy, and maybe, if you wait long enough, it will eventually be ok. > But i have waited for several minutes sometimes, still no go. > So the only option is to force quit out of Nuke. > The odd thing is, i can sometimes work in the same script for hours, then > suddenly the symptons start to show up. I even had it affect other nodes as > well. So dbl-clicking a merge node for example too ages. > I am on Nuke 8.0v6 on Linux. > > So my question is: are there any operations in Nuke that are single > threaded esp. regarding Rotos that could trigger this behavior? Have other > people seen this as well? > > Regards, > Daniel___ > Nuke-users mailing list > Nuke-users@support.thefoundry.co.uk, > https://urldefense.proofpoint.com/v2/url?u=http-3A__forums.thefoundry.co.uk_=CwIFaQ=hHkK43hKb5uKUGaYMJ-p252cFdfVBHtyegDvswk59fU=ilRVisJEHQmq9__Xya9HaQsNGSKiRcQWFhzTKtLfVGE=xwcgBNYY4WZEQKkwg6glCgaJgv8rIYpAIVFYKItVAy8=J36lFYVResxadI3h1dkBqxJPG3pj9G1mENp3es6CKCA= > > https://urldefense.proofpoint.com/v2/url?u=http-3A__support.thefoundry.co.uk_cgi-2Dbin_mailman_listinfo_nuke-2Dusers=CwIFaQ=hHkK43hKb5uKUGaYMJ-p252cFdfVBHtyegDvswk59fU=ilRVisJEHQmq9__Xya9HaQsNGSKiRcQWFhzTKtLfVGE=xwcgBNYY4WZEQKkwg6glCgaJgv8rIYpAIVFYKItVAy8=GQr-0hea83arFuOwsdNB3_x8LEiHprW51yRZlmY0HMk= > ___ > Nuke-users mailing list > Nuke-users@support.thefoundry.co.uk, 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: l...@thefoundry.co.uk The Foundry Visionmongers Ltd. Registered in England and Wales No: 4642027 ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
Re: [Nuke-users] Total render time in GUI?
They are indeed accessible via Python: http://www.thefoundry.co.uk/products/nuke/developers/90/pythondevguide/performance.html Regards, Lucy On Fri, Mar 4, 2016 at 12:54 PM, Elias Ericsson Rydberg < elias.ericsson.rydb...@gmail.com> wrote: > There we go! Makes the process a lot easier, are these metrics accessible > through python or tcl? If so I'd love to know how to do it. > > Cheers, > Elias Ericsson Rydberg > > fredag 4 mars 2016 skrev Joel Byrne <by...@thefoundry.co.uk>: > >> Hi, chaps. >> >> One thing you could do is launch Nuke from the command line with the *-P* >> arg to turn on the performance metrics. When you add a Write and start >> rendering, you should see the *wall *count climbing steadily. The wall >> metric is the time taken if you were looking at a clock on the wall (that >> happened to measure time in milliseconds for whatever reason). >> >> There's some more info here if you're interested: >> http://help.thefoundry.co.uk/nuke/9.0/Default.html#comp_environment/organizing_scripts/using_script_profiling.html >> >> Cheers, >> Joel. >> >> On Fri, Mar 4, 2016 at 12:30 PM, Elias Ericsson Rydberg < >> elias.ericsson.rydb...@gmail.com> wrote: >> >>> Maybe not a super easy way, but right from the top of my head here's one >>> way is to do it. Have a process start in a before render script, then on >>> render finish start a second script that checks how long the process has >>> been running, store the result in a variable and kill the process. To view >>> the result, have python create or edit a text node and overlay the result >>> on your rendered image. >>> >>> There may be easier ways, I don't remember if the individual nodes in >>> nuke stores the time it takes to process them. In that case you would just >>> sum the upstream nodes. >>> >>> Just an idea, >>> Elias Ericsson Rydberg >>> >>> fredag 4 mars 2016 skrev Nick Guth <nick.g...@gmail.com>: >>> >>>> Is there an easy way to see total render time in the GUI when executing >>>> a write node? Cheers. >>>> >>> >>> ___ >>> Nuke-users mailing list >>> Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ >>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>> >> >> >> >> -- >> Joel Byrne >> Technical Author >> *The Foundry*. >> 5 Golden Square >> London, W1F 9HT >> Telephone: +44(203) 434 3190 >> Web: www.thefoundry.co.uk >> >> The Foundry Visionmongers Ltd. >> Registered in England and Wales No: 4642027 >> > > ___ > Nuke-users mailing list > Nuke-users@support.thefoundry.co.uk, 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: l...@thefoundry.co.uk The Foundry Visionmongers Ltd. Registered in England and Wales No: 4642027 ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
Re: [Nuke-users] Mac book pro GPU supoport not working
Hi Simran, Support for the AMD GPU in your Macbook is coming soon in the Nuke 10 release (http://www.thefoundry.co.uk/products/nuke/beta/10/). Regards, Lucy On Tue, Jan 26, 2016 at 1:03 PM, Simran Takhar <simran.tak...@gmail.com> wrote: > Hi, > I just recently bought mac book pro early 2016 January edition to work on > nuke but I am afraid that i am not able to get a GPU support can somebody > help ? > I am also attaching a screen shot of my configuration if it helps > > -- > Simran Takhar > > ___ > Nuke-users mailing list > Nuke-users@support.thefoundry.co.uk, 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: l...@thefoundry.co.uk The Foundry Visionmongers Ltd. Registered in England and Wales No: 4642027 ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
Re: [Nuke-users] Horrible freezing on Win 7 with any Motion Blur
Hi Andrew, There is a bug that could be related to this, where some scripts were rendering very slowly on 24-core machines. This has been fixed internally and should be making its way into a release soon. However, 20s sounds a bit more extreme than I would expect to see. Would you be able to send an example script that shows the problem, and full machine specs, so that we can see if it's the same thing? Also, if you run Nuke with fewer threads, e.g. "Nuke9.0 -m4" for 4 threads, how does the performance compare? Thanks, Lucy On Sat, Sep 12, 2015 at 8:17 PM, Andrew Mumford <a_mumf...@mac.com> wrote: > Having abysmal problems trying to do anything using “standard” built in > motion blur in Nuke. > > I’m talking about the standard motion blur in transform, paint, tracker > etc - basic stuff nothing to do with vectors or other schema. Any use of > motion blur knob above zero triggers multi second, say 20 or more, freezes > eventually slowing to the point where nuke locks up after every button > press in the gui. > > Is anyone else seeing this and does anyone know if there similar bugs > relating to Nuke on multi / multi core PC’s, 24 in my case, and multi > threading breaking / slowing down Nuke. > > Thanks for any feedback / sarcasm / Trumpism’s etc > > --- > Andrew Mumford > > a_mumf...@mac.com > 1 (310) 619 0208 > > ___ > Nuke-users mailing list > Nuke-users@support.thefoundry.co.uk, 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: l...@thefoundry.co.uk The Foundry Visionmongers Ltd. Registered in England and Wales No: 4642027 ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
Re: [Nuke-users] Video card nVidia GTX 780
Hi Valerio, The output you mention actually suggests that Nuke *is* recognising your GPU. This message in particular suggests that Nuke has found your device: Can't select GPU device no GPU available : 0 using device GeForce GTX 780 : 0 It could be better-worded but it means that while previously Nuke was unable to find a GPU to use on your system, it has now found one and is using it (the using device GeForce GTX 780 bit). The fact that you can see your GPU named in the default Blink device field also suggests that the GPU acceleration is working correctly. If you create a GPU-enabled node, for example Denoise, do you also see Local GPU: GeForce GTX780 in the parameter panel? If so, you should be able to select Use GPU if available (it's actually on by default) to use the GPU for processing. Regards, Lucy On Tue, Feb 24, 2015 at 11:50 AM, Valerio Oss - Pixel Cartoon ossv...@pixelcartoon.it wrote: Hi I just bought a nVidia GTX 780 video card, that I'm going to use in Nuke, After Effects and some 3D apps The exact model is ASUS GeForce GTX 780 STRIX OC Edition Direct CU II - 6GB RAM I am having some issues with Nuke, that is, it doesn't recognize it as a GPU. When I launch Nuke it says: Can't select GPU device no GPU available : 0 using device GeForce GTX 780 : 0 And hardware acceleration doesn't work. I tried to install the CUDA drivers from nVidia https://developer.nvidia.com/cuda-downloads but no way In the preferences (Performance/Hardware) I see: Default blink device: GeForce GTX780 My system is Win 8.1 64bit, and Nuke is 9.0v3 Any ideas why Nuke doesn't recognize it? Thank you Valerio -- - Non cedete mai, non guardate indietro, non ritenetevi mai incapaci. Riuscirete, senza ombra di dubbio. - We must be greater than what we suffer -- Valerio Oss PIXEL CARTOON - Via Taramelli 8/4 - 38122 Trento - ITALY Tel +39 0461 090596 Fax +39 0461 1533001 IT Cel Phone: +39 348 7401720 UK Cel Phone: +44 (0)7757188562 e-mail: ossv...@pixelcartoon.it web: http://www.pixelcartoon.it twitter: @ossvale IMDb: http://www.imdb.com/name/nm3645361/ -- ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, 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: l...@thefoundry.co.uk The Foundry Visionmongers Ltd. Registered in England and Wales No: 4642027 ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
Re: [Nuke-users] rendering with laptop - GPU or CPU? System crash on render
If you're rendering from the command line, the GPU won't be used unless you pass in the --gpu command-line option. If rendering from the GUI, you can change default blink device under Preferences - Performance - Hardware to Use CPU to never use the GPU, rather than turning it off for each individual node. Lucy -- 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: l...@thefoundry.co.uk On Sun, Jan 25, 2015 at 9:39 AM, Ron Ganbar ron...@gmail.com wrote: Happened to me a lot with my iMac, actually. I found that turning off all GPU acceleration solved this (thought caused longer renders). I opened the properties panels for each node that has GPU acceleration and turned that option off. Ron Ganbar email: ron...@gmail.com tel: +44 (0)7968 007 309 [UK] +972 (0)54 255 9765 [Israel] url: http://ronganbar.wordpress.com/ On Sun, Jan 25, 2015 at 11:33 AM, Darren Coombes darren.coom...@me.com wrote: Im rendering on macbook pro with nuke 9, with NVIDIA GeForce GT 650M graphics card, and I think the graphics cards on its way out.. After a couple of minutes rendering my sequence, the computer shuts down. If rendering using the computer you’re working on, does Nuke use the GPU for the rendering or the CPU? If its the GPU giving me problems, can i get it to render using the CPU only? It might be my only chance of getting frames out. Thanks. *Thanks.Darren Coombes* *Check out some of my work...* *www.vimeo.com/darrencoombes http://www.vimeo.com/darrencoombes* *Mob: +61 418 631 079 +61%20418%20631%20079* *Skype: darrencoombesTwitter: @durwood81* ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users The Foundry Visionmongers Ltd. Registered in England and Wales No: 4642027 ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
Re: [Nuke-users] Blink cross() function not working?
Answered on the nuke-dev list - my response is below in case anyone else has hit this issue and does not have access to that list. *Hi Steve,* *This is an error in the documentation - it should specify that the cross function is only defined for vectors with three components, i.e.: vec3 cross(vec3 a, vec3 b); //Returns the cross product of vector *a* with vector *b*.* *(This is because mathematically, the cross product is only defined for vectors in three-dimensional space). Sorry for the confusion - we'll update the documentation to make this clear. Please let us know though if you still have problems when dst(), A() and B() are all three-component vectors.* *Regards,Lucy* 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: l...@thefoundry.co.uk The Foundry Visionmongers Ltd. Registered in England and Wales No: 4642027 On Fri, Dec 5, 2014 at 3:50 PM, Steve Newbold s...@dneg.com wrote: Hi, Using the BlickScript node in Nuke 8.0v6, the following works (A B are input images), *dst() = dot(A(), B());* but this doesn't, *dst() = cross(A(), B());* and I got the following error message, *Error compiling kernel: Error at line:30 no matching function for call to 'cross'* The Blink docs have this, *scalar dot(vec a, vec b); //Returns the dot product of vector *a* with vector *b*.* *vec cross(vec a, vec b); //Returns the cross product of vector *a* with vector *b*.* So I'm not sure what is wrong with cross() or whether I'm using it incorrectly. Cheers, Steve ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
Re: [Nuke-users] GPU single/double precision
Hi John and Peter, The GPU-accelerated nodes in Nuke currently use single precision only. Hope that helps. Lucy 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: l...@thefoundry.co.uk The Foundry Visionmongers Ltd. Registered in England and Wales No: 4642027 On Tue, Oct 7, 2014 at 8:37 AM, Peter Hartwig peter.hart...@gmail.com wrote: Yeah i would love to know that too. Im building a new workstation and two 980 cards are still cheaper than a single k5200 On Tuesday, October 7, 2014, John Perrigo jpe...@gmail.com wrote: With the release of the new 900 series nvidia cards, it looks like single precision compute performance has increase significantly but double precision performance is less than the 700 series. Can anyone tell me what the GPU accelerated features in Nuke benefits from? -- Peter Hartwig Freelance VFX artist www.spyfactory.tv www.renderschool.com -- ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users ___ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users