Hello ! I have been trying myself to integrate this GPU workflow for my project, but seems like most of the time I am having this issue when I render :
ERROR | [gpu] an error happened during rendering. OptiX error is: Unknown error (Details: Function "_rtContextLaunch2D" caught exception: Encountered a rtcore error: m_api.pipelineCreate( context, pipelineOptions, compileOptions, modules, moduleCount, pipeline ) returned (8): Link error) GPU 0 had 23230MB free before rendering started and 13426MB free when crash occurred
log.txt
here on the top is the attached log, but overall I think the issue is just the one in the code bar, where I am still assuming that it still have 13426MB of memory free even before it crashed..?
I think this may be the reason why I cannot do a batch render or even render sequence button.
I would really appreciate it if there is a fix on this -- as GPU rendering would help me a lot speed-wise, and I thought it should be fairly stable and i accept that some AOVs would not work,but not expecting it to be not able to do any render at all..
The scene has a bunch of aiStandIn as well for more info, but I thought that this should not be the problem..?
I am using Nvidia Geforce RTX 3090 and my driver is in Nvidia Studio Driver version 471.68
Thank you !
Unfortunately, there's no way to know what happened from just that message. The error message is generic, so all we know is that it crashed. It could be a bug in Arnold or in Optix (the GPU rendering engine).