HTOA in Houdini 18.5.462 does not render anymore...

[alainh3d]

HTOA does not render anymore.... I lots of errors suddenly:

1"Failed to save output to file "Traceback (most recent call last)

I have not used Houdini env but intead a json file in the packages folder and it's been working great for the last 3 months until yesterday.

Or this error:

2 node "driver_houdini" is not installed

or

(null): could not set STRING parameter "name"

I cleaned up my file to the bare minimum and tried to delete object per object but no luck...

I got it working for 5 minutes today but that's it... So I saw in another post that we could send you a PML file, which is what I'm doing now:

https://we.tl/t-hZMcHMVgXy

I was using GPU so i tried CPU but it's the same. What else can I do?

this is my env file:

#
# Houdini Environment Settings
#
# The contents of this file are read into the environment
# at startup. They will override any existing entries in
# the environment.
#
# The syntax is one entry per line as follows:
# VAR = VALUE
#
# Values may be quoted
# VAR = "VALUE"
#
# Values may be empty
# VAR =
#

# Example:
#
# HOUDINI_NO_SPLASH = 1

OCIO = "C:\Program Files\OpenColorIO_1\config.ocio"

[Stephen.Blair]

It's the Render View that doesn't work, right?
Rendering to mplay, or to a file on disk does work.

I'll take a look at the Process Monitor log.

[alainh3d]

No. When it bugs it does it for "rendering to a file" also...

But I might be in luck as i just copied all my objects 1 by 1 in a fresh scene and got the CPU to work on RenderView... I'm trying to render a sequence out on file right now... That's working again... After several reboots...

I'm trying the GPU... And it's also working! Man, I don't know what's going on?! Let me know if you see sometinhg in the log, but maybe something corrupted in the file?

Thank you for responding so quickly, though... It feels good to know that someone out there is ready to help 😉

[alainh3d]

And i should add that GPU is blistering fast comparing to CPU! that's great!

[Stephen.Blair]

node "driver_houdini" is not installed

That's never about the scene itself. It's something weird going on with HtoA in Houdini, where driver_houdini disappears for some reason. Very hard to track down...

[Stephen.Blair]

driver_houdini is not used when you render a file to disk. If a render to disk fails, it must be something else.

There is no reference to driver_houdini in the Process Monitor file.
Was this log for a render to disk? And what was the error when you tried to render?
You'll need to check the Arnold ROP for the actual log

[alainh3d]

ok. thank you Stephen.

[alainh3d]

I tried to post the ROP error log here but it creates a new question? so to make sure I'm sending you a text file:

AH_ROP_errro_log.txt

[alainh3d]

i meant to write AH_ROP_error_log not "errro"

[Stephen.Blair]

Ok, that's a generic GPU error:

00:00:06 1066MB ERROR | [gpu] an error happened during rendering. OptiX error is: Unknown error (Details: Function "_rtContextLaunch2D" caught exception: Encountered a CUDA error: cudaDriver().CuEventSynchronize( m_event ) returned (700): Illegal address, file: <internal>, line: 0)
GPU 0 had 8410MB free before rendering started and 4767MB free when crash occurred

Does it happen all the time? There's no way to know exactly what the problem is, not from just the error message.

[alainh3d]

i reopened the original file and here is the error:

Error

Command Exit Code: -6

00:00:00 76MB | [hick] Received: AA: 2 2
00:00:00 76MB | [hick] Received: begin
00:00:00 76MB | [hick] Batch mode
00:00:00 76MB | log started Thu Oct 14 08:11:04 2021
00:00:00 76MB | Arnold 6.2.0.1 [903992ac] windows clang-10.0.1 oiio-2.2.1 osl-1.11.6 vdb-7.1.1 clm-2.0.0.235 rlm-12.4.2 optix-6.6.0 2021/02/09 10:40:42
00:00:00 76MB | running on CORSAIR-II, pid=24516
00:00:00 76MB | 1 x AMD Ryzen Threadripper 3970X 32-Core Processor (32 cores, 64 logical) with 130960MB
00:00:00 76MB | NVIDIA driver version 472.12 (Optix 60801)
00:00:00 76MB | GPU 0: NVIDIA GeForce RTX 3080 @ 1710MHz (compute 8.6) with 10240MB (8410MB available) (NVLink:0)
00:00:00 76MB | Windows 10 (version 10.0, build 19043)
00:00:00 76MB | soft limit for open files raised from 512 to 2048
00:00:00 76MB |
00:00:00 76MB | [hick] Start
00:00:00 76MB | [hick] Received: log_verbosity: 7942
00:00:00 76MB | [hick] Received: log_console_enable: 1
00:00:04 404MB WARNING | [polymesh] /obj/Env/polygons: discarded 3 duplicate deformation keys
00:00:04 404MB WARNING | resetting parameter nlist on /obj/Cholesterol/polygons (found: nans or infs)
00:00:04 405MB WARNING | [polymesh] /obj/Cholesterol/polygons: incomplete vertex normal specification; ignoring normals
00:00:04 408MB WARNING | [polymesh] /obj/Veins_BG/polygons: discarded 3 duplicate deformation keys
00:00:04 399MB WARNING | [polymesh] /obj/Vein/polygons: discarded 3 duplicate deformation keys
00:00:06 486MB WARNING | [aov] output redefines AOV "Z" as type "RGBA": ignoring and using previous type "FLOAT"
00:00:06 486MB WARNING | [aov] filter "/out/BloodFlow_a1:mitnet_filter" does not support INT inputs
00:00:06 1066MB ERROR | [gpu] an error happened during rendering. OptiX error is: Unknown error (Details: Function "_rtContextLaunch2D" caught exception: Encountered a CUDA error: cudaDriver().CuEventSynchronize( m_event ) returned (700): Illegal address, file: <internal>, line: 0)
GPU 0 had 8410MB free before rendering started and 4767MB free when crash occurred

[alainh3d]

Hi Stephen,

It seems like 1 object in my scene needed to have a material assigned at Geometry level (in Render tab) in order to render and work, even if there is a material node on the object itself...

I also uninstalled the latest Nvidia Game ready driver and reverted back to the studio driver and things are working... So maybe that was the problem?