Optix_error_invalid_value

Omniverse General Discussion
1

I am getting this error in the sample marble project. This happens after I switch from RTX Real Time to RTX Path Traced.
I first get an error something in lines of “CUDA ran out of memory” (I have an RTX 2070 8gb)
Then I recursively get the following error. Seems that the denoiser breaks and my viewport quality drops significantly even if I switch back to Real Time mode.

“2021-03-16 14:41:02 [2,210,979ms] [Error] [rtx.optixdenoising.plugin] [Optix] [ERROR] (void*)outputIntensity is null
2021-03-16 14:41:02 [2,210,980ms] [Error] [rtx.optixdenoising.plugin] optixDenoiserComputeIntensity( denoiser, stream, &inputLayer[0], denoiserSetup.denoiserIntensity, denoiserSetup.denoiserScratch, sizes.recommendedScratchSizeInBytes) failed. Optix Error: OPTIX_ERROR_INVALID_VALUE.”

1 Like
2

Hello dirtpizza,
Thank you for the post, I will bring your issue to the proper channels for review.
Thank you for your patience.

1 Like
3

Thanks icolon! Any work around would be greatly appreciated. I really wanted to be part of the Create with Marbles competition.

1 Like
4

Hello!
I have the exact same issue as the OP - Any updates on what may be causing this?
Thanks!

5

I have created an internal ticket for this issue (for your reference OM-29252). I will update you when I know more information!

6

Hi,
I get this error message when path tracing. Image capture provided.
I use an rtx 2080.

2021-03-23 23:36:42 [Error] [rtx.optixdenoising.plugin] CUDA error 2: cudaErrorMemoryAllocation - out of memory)

Thank You.

Capture1
Capture11920×1080 1.55 MB

7

I would hazard a guess it’s exactly what it looks like - the scene is exceeding available vram on your card

I’m not sure Create is especially optimised on the memory footprint front based on comparing Vram used on similar scene data to another somewhat similar app

1 Like
8

Hi,
Raising path trace samples pr. pixel helped in redering.

thanks.

9

Hi All. I spoke with the dev team and they informed me that this error is because you are running out of GPU memory. They suggest using fewer assets for the lower end GPU cards. We will be looking into ways on improving this in the near future.

1 Like