ID 312943 - Nuke does not clear its cache after dealing with larger than standard resolution files, such as 50,000 x 30,000.

Follow

Problem summary
Nuke does not clear its cache after rendering with larger than standard resolution files, such as 50,000 x 30,000. When working with a standard resolutions such as NTSC video up to 8k cube maps Nuke will clear it's cache and work as intended. 


Customer reported version
nuke.10.5v2

Customer reported platform
10.11

Steps to reproduce

1) Launch Nuke and import the the attached script. (A checkerboard with a custom resolution of 50,000 x 30,000 with 16 write nodes attached rendering out null.null files)

2) Navigate to the render tab and select "render all write nodes" (f5)

3) You will notice that the time after each render will get progressively longer after each render, due to Nuke not clearing its cache. 

Workaround
Navigate to your init.py file and insert the following line:

nuke.addAfterFrameRender(nukescripts.cache_clear, ())

The init.py file is located within your .nuke directory. If it does not exist, feel free to create it in a text editor.

You can locate it via this link: Q100048-Nuke-Directory-Locations

Reproduced by support
Nuke 11.0v2 - Windows 7 - Linux CentOS 6.9 - MacOS 10.12

Nuke 10.5v6 - Windows 7 - Linux CentOS 6.9 - MacOS 10.12 

Nuke 10.5v1 - Windows 7 - Linux CentOS 6.9 - MacOS 10.12

Nuke 10.0v6 - Windows 7 - Linux CentOS 6.9 - MacOS 10.12

Nuke 10.0v1 - Windows 7 - Linux CentOS 6.9 - MacOS 10.12

Nuke 9.0v9 - Windows 7 - Linux CentOS 6.9 - MacOS 10.12

Nuke 9.0v1 - Windows 7 - Linux CentOS 6.9 - MacOS 10.12

Nuke 8.0v6 - Windows 7 - Linux CentOS 6.9 - MacOS 10.12

Nuke 8.0v1 - Windows 7 - Linux CentOS 6.9 - MacOS 10.12

Nuke 7.0v10 - Windows 7 - Linux CentOS 6.9 - MacOS 10.12

Earliest version tested
Nuke 7.0v10

Expected behaviour
Due to the large file size associated with this bug, I assumed Nuke would slow down at render time.


Actual behaviour
Nuke did slow down to a crawl after not clearing its cache after rendering such a large file repeatedly.   


    We're sorry to hear that

    Please tell us why