IES-VE 2017 - file "not found in RAYPATH"

[eur...@staenglengineering.com]

Hi all,

I know this issue has been around since the start of Accelerad, but I'm trying to run Radiance via IES-VE 2017 (Feature Pack 4), and I'm getting this error:

rpict -w -t 1 -vtv -vp 9.601 -10.516 42.228 -vd -11.201 7.849 -10.427 -vu 0 0 1 -vh 80.000 -vv 60.000 -x 1280 -y 1280 -x 64 -y 64 -ps 6 -pt 0.10 -pj 1.0 -dj 0.35 -ds 0.30 -dt 0.05 -dc 0.80 -dr 2 -dp 128 -st 0.50 -ab 2 -aa 0.15 -ar 64 -ad 128 -as 64 -lr 6 -lw 0.005 -av 0.00 0.00 0.00 -af C:\Users\Eric\AppData\Local\IES\tasks\2957\Radiance\image4.amb -e C:\Users\Eric\AppData\Local\IES\tasks\2957\Radiance\image4.log C:\Users\Eric\AppData\Local\IES\tasks\2957\Radiance\image4.oct

rpict: 0 rays, 0.00% after 0.0000 hours

rpict: system - File cuda_compile_ptx_generated_camera.cu.ptx not found in RAYPATH.: No such file or directory

I've tried editing the RAYPATH directories, as shown here, but it doesn't help. The Accelerad folder permissions are set such that everyone has full access. I also attached a screenshot showing that the file in question is in the Accelerad lib folder. Does anybody know what is causing this problem?

[Nathaniel Jones]

First check that the demos included with Accelerad run successfully. If they do, then your RAYPATH variable is set up correctly.

Are you running Accelerad from within IES<VE>? I'm not an IES user, but if Accelerad works outside IES and not inside, then my guess is that IES sets its own environment variables when it runs. This document from IES support seems to support that theory. It looks like you can change the RAYPATH used by IES in the registry.

Alternatively, you could copy the files from the Accelerad lib folder into the Radiance lib folder.

Nathaniel

[eur...@staenglengineering.com]

Yes, Accelerad ran successfully for all of the demos. Yes, I am trying to run Accelerad from within IES<VE> but haven't gotten it past the directory error. I could only find one RAYPATH entry in the registry, and it matched the environment variable. IES doesn't have a dedicated Radiance lib folder. It only has a generic "app\radiance" folder. I put all the Accelerad lib files there. Still no luck. I'll let you know if I find a solution. Please let me know if you have any ideas too. 

Thanks,
Eric

[Nathaniel Jones]

The Radiance lib folder should contain the file rayinit.cal. It could be renamed, so possibly that is the apps\radiance folder.

In case you edited any registry or environment variables after opening IES, you may need to close it an open it again.

Because the issue seems to be IES not reading the RAYPATH correctly, I would suggest contacting IES technical support for a solution.

Nathaniel

[eur...@staenglengineering.com]

Thanks, I've sent a query about the issue to the IES support team. We'll see if they have a solution.

[eur...@staenglengineering.com]

The folks at IES say this isn't something they test for or support.

[Nathaniel Jones]

I'm asking IES for a trial license to take a look at this. Apparently they use a bespoke Radiance version, so it may have different behavior.

However, from the error you received, it's clear that IES was calling the Accelerad version of rpict, so I'm surprised that it didn't find the .ptx files, particularly if they were in the same folder as rayinit.cal.

Do you have a simple reproducible example that you can share with me by email to test this out?

Nathaniel

[Nathaniel Jones]

I was able to get Accelerad to work with IES. The steps were:

1. Copy the contents of C:\Program Files\Accelerad\bin into C:\Program Files (x86)\IES\VE 2018\apps\radiance
2. Rename rpict.exe to rpict_old.exe, and rename accelerad_rpict.exe to rpict.exe. The same would apply to rtrace and rcontrib, but I've only tried rpict.
3. Copy the .ptx files from C:\Program Files\Accelerad\lib into C:\Program Files (x86)\IES\Shared Content\radiance
   

The only issue is that I'm not aware of how to edit the special parameters Accelerad uses for parallel irradiance caching through the IES interface. The defaults were pretty badly tuned to the example model you sent, resulting in poor ambient lighting rendering.

Nathaniel

[eur...@staenglengineering.com]

I followed the steps you listed and it worked for me, too. I got some strange renderings using the default "Medium" and "High" quality settings (like all surfaces rendered as black except for windows, or large splotchy shadows), but after some customization of the radiance parameters it improved a lot. 

Thanks again for your help. Hopefully this will help other IES-VE users as well.

Eric

[Nathaniel Jones]

Hi Eric,

Can you let me know what settings turned out to improve the rendering quality? That may be useful information to share around as well.

There's also some information on this forum about past experience with the rendering settings in DIVA, which may be similar.

Nathaniel

[eur...@staenglengineering.com]

I'm still playing with the parameters because I think the accuracy is overkill for my purposes, but here's what I used. Overall it's a bit grainy, but that can be mitigated scaling the image down afterwards.

[Nathaniel Jones]

I see that you've turned off irradiance caching by setting ambinet accuracy to 0. That will certainly get rid of the artifacts that Accelerad can introduce by not being able to modify the special parameters, but it also requires that you use a substantially smaller limit weight value in order to account for multiple ambient bounces. Normally, limit weight should be about 1/(ambient divisions), but with ambient accuracy set to zero, you might want to use 1/(ambient divisions^2) or smaller, or you may want to make limit reflections negative to turn on Russian roulette. Unfortunately, all of these will result in longer rendering times, and may cause stack overflows because IES will not let you tune the -g parameter.

Alternately, you might try a large ambient accuracy value like 0.2.

I agree that and ambient divisions and ambient super samples are probably set to greater accuracy than you need. Unless you expect large variations in brightness across your scene, you can probably turn ambient super sampling off by setting it to zero.

Nathaniel

[eur...@staenglengineering.com]

Accelerad has been incorporated to the 2019 version of IES-VE released today. I have not yet tested it, but I thought some people would like to know.

Eric

[IES user Rad]

Hi, did you manage to try the incorporated IES accelerad functionality in the end? Any success in terms of achieving effective GPU utilisation?