cragl vfx tools | Nuke Python Tools and Plugins | smartRender

• blog
• tools
• studio
• about
• faq
• support
• login
• contact

Render time comparisons

The following shows some direct 1:1 real time render comparisons between smartRender and NUKE's default rendering method. We tested different nuke scripts of different scale with different machines and different operating systems. Each test is a real time recording of the exact same nukescript on the same machine. In our tests rendering with smartRender was always faster, sometimes even incredibly 4 times faster. As these comparisons are quite long feel free to simply skip through them and have a look at the comparison chart at the end of each video. The brave ones watch the whole video in one go :-D

How do I install smartRender?

Each cragl tool can be installed via cragl connect - our free tool to install and manage all of your cragl tools directly inside NUKE. If you haven't already done so please download connect.
By opening connect you will see a list of all cragl tools. You can drag and drop it to the right hand side of the connect window to reveal more infos. Just click on "buy". After purchasing you will receive an install code for your tool. Now simply click "install now" and enter the received install code. The tool will be downloaded and installed automatically for you. So you don't need to struggle with any install issue.

smartRender doesn't render as expected. How can I trouble shoot my work file?

If smartRender does not work properly on your working file then you should check a few things. Keep in mind that each working file is individual and has its own node structure so there is no general 'This is the way to do'. However, going through the following steps will ensure that your working file will render by a high chance.

1) Please check your job log. Any error should be reported in here. This will hopefully give you a good hint which nodes cause trouble rendering.

2) Although smartRender has been tested on all major platforms with all major Nuke versions, please double check that a default working file setup is working as expected on your machine: For that please create a very simple setup and render it on your Desktop. For instance, create a checkerboard and connect it to a Write node and set up the render path to let it render to a folder on your desktop. Try to render this simple working file using smartRender. Keep in mind that rendering such a simple working file might probably slower than rendering it the traditional way. That is because your rendering threads need to open the working file - which takes a few seconds. If your final rendering location is a server, then move forward with your simple setup and try rendering into your final render location on your server -- Or wherever your location is. This is to ensure that running in terminal mode, Nuke has the file permissions to render to the final render location.

If this simple setup renders successfully, then we are on a good way and have made sure that smartRender itself works as expected. Keep reading.

3) Please convert your gizmos to group nodes. Make sure that they are initialized through init.py files and not menu.py files. The same must be the case for any third party nodes and/ or plugins. The reason for that is simple: smartRender uses Nuke's terminal mode to render. The terminal mode does -not- read/ load any menu.py file. These module files are for interactive Nuke sessions only.

So please convert the gizmos to group nodes and making sure that nothing essential for your workfile to work is being loaded from a menu.py. If your working file still doen't render using smartRender then keep reading.

4) Probably a certain node/ some certain nodes refuse to render in terminal mode. In order to further trouble shoot your working file please try to render just a specific section from it. If this specific section renders, then try the next section. This should quickly give you a hint which nodes cause trouble when rendering in terminal mode.

5) smartRender is a terminal renderer: You can see what the terminal will execute by doing the following: Export your job log, open it and copy one of the lines of the terminal "input section" (1 line is one command - you will probably need to expand your file to fullscreen to grab one whole line). Paste the copied line into your terminal and have a look at the output. If there is any problem rendering it will appear in here.

6) Temporarily disable all other custom python scripts and plugins that you are customizing your Nuke with. From a technical point of view it is definitely possible that other custom scripts might cause issues with smartRender's rendering process. So temporarily disabling all other custom python scripts and plugins is a good way of checking that nothing interferes.

7) Please contact our support. Please include your working file (just the .nk file, no sources) and specify your used Nuke version and operating system including version. We are happy to help you.

I have problems rendering gizmos

If you use gizmos in your script please convert them to groups ("node" tab -> button "convert to group"). The reason behind this is simple: as the multi threaded rendering is executed in terminal mode each nuke session that processes a frame launches in terminal mode and doesn't load your menu.py file in which your gizmos are most likely loaded to the UI. Although an init.py file should be used for initialization of any gizmos, third party developers are free to add additional logic to their menu.py files - which won't be loaded when running Nuke in terminal mode.

If it comes to processing a frame, your nuke session will raise an error if there is a gizmo inside your processing tree that hasn't been fully initialized through a proper init.py file, the rendering will fail because Nuke doesn't know how to handle an unknown command. In order to avoid this issue, convert your gizmos to groups. It's anyway a good habit to convert all gizmos to groups that so your nuke script is flexible and is transferable to another computer on which the gizmo will probably not be installed.

My computer seems to freeze or is pretty slow after sending a job to render

If you set a high number of threads and your scene is quite heavy your computer is likely to slow down and blocking. This is completely normal as you told your computer to render several frames simultaneously. If you experience a complete slow down and your computer doesn't respond for a long time chances are high that your computer is getting some memory overflow. This is likely to happen on huge nuke scripts where you use multiple threads. To solve this try reducing the number of used threads for that job. Furthermore if you decide to use all available threads for rendering try to shut down other unneeded applications.

Does smartRender support rendering videos multithreaded?

No, but read on. Rendering a video multi threaded won't make sense, due to technical limits. This should be clear if you understand smartRender's workflow.

A file sequence contains multiple files. This is perfect for out sourcing to several threads which can render these frames independently from each other. When rendering a single video file though, it should become clear that this task can only be done by one worker because it operates on just one file.

Note though that smartRender contains functionality of auto transcoding your renderings. You can render a frame sequence multithreded and automatically let smartRender transcode it afterwards for you. For that please refer to smartRender's Auto transcoding after render documentation section.