Radiance 6 timeline?

[LittleBuddha]

Hi.

Have noticed that there is a new version of Radiance out, that promises new features and a speed up.

I see that @mikkel has been prowling the radiance forums.

Is there a timeline on an update here?

Regards,
LittleBuddha

[mikkel]

Hi @LittleBuddha,

We have not yet discussed shipping Radiance 6 with the installers, but I have been using it myself without issues except for the bug that I reported on the Radiance forum. This has been fixed in the version released a week ago.

I replaced the version in the Radiance directory, C:\Program Files\ladybug_tools\radiance. Setting the workers to 1 in HB Recipe Settings and adding “-n x” in HB Radiance Parameter worked as expected.

[LittleBuddha]

@mikkel

worked as expected.

That sounds great, will give it a try.

When you say “worked as expected”, I expect that you mean the same as 5, but faster?

Yours,
@LittleBuddha

[LittleBuddha]

Ran a model, that works on pollination installer.

Crashed:

1. Solution exception:
   The recipe failed to run with the following error(s):

2026-03-20 13:02:20 ERROR: “DirectSky” failed. See below for more information:
rfluxmtx -I -aa 0.0 -ab 1 -ad 5000 -ar 0 -as 2500 -c 1 -dc 0.25 -dp 64 -dr 0 -ds 0.5 -dt 0.5 -faf -lr 4 -lw 2e-06 -n 15 -ss 0.0 -st 0.85 -y 2242 - “sky.dome” -i “”“scene.oct”“” < “grid.pts” | rmtxop -ff - “sky.mtx” -c 0.265 0.670 0.065 > “results.ill”
2026-03-20 13:02:24 ERROR: “TotalSky” failed. See below for more information:
rfluxmtx -I -aa 0.0 -ab 2 -ad 5000 -ar 0 -as 2500 -c 1 -dc 0.25 -dp 64 -dr 0 -ds 0.5 -dt 0.5 -faf -lr 4 -lw 2e-06 -n 15 -ss 0.0 -st 0.85 -y 2242 - “sky.dome” -i “”“scene.oct”“” < “grid.pts” | rmtxop -ff - “sky.mtx” -c 0.265 0.670 0.065 > “results.ill”

These are the settings I am running with.

And the error persist at -n 1

Running Radiance 6.1a (2026-03-13).

Any ideas?

[mikkel]

@LittleBuddha

When you say “worked as expected”, I expect that you mean the same as 5, but faster?

Yes, but depending on the number of sensors and parameters it might be slower when setting the workers to 1 since the results – which is one huge matrix when using 1 worker – will be sent to rmtxop and this will now be a bottleneck. This can probably be fixed by using the new rcomb tool that Greg presented at the 2024 Radiance Workshop.

@LittleBuddha

And the error persist at -n 1

Running Radiance 6.1a (2026-03-13).

Any ideas?

Are you able to share the log file and the Grasshopper file with internalized geometry?

C:\Users\[USERNAME]\simulation\[MODEL_NAME]\annual_irradiance\__logs__\logs.log

[LittleBuddha]

Sorry @mikkel, this one is my fault. I am not a smart man!

I removed the Rad5 folder, before putting in the Rad6 components. And some essential files were therefore missing. Copied them in, and now it runs.

Turns out, in my case, with only 620ish points, Rad6 uses 1.5 -2 times longer with workers set to 1 and n to 15, than, workers set to 15 and n to 1.

Here is my model, if you want to play with it.

PV geometri - rad6.gh (75.2 KB)

[mikkel]

@LittleBuddha

Turns out, in my case, with only 620ish points, Rad6 uses 1.5 -2 times longer with workers set to 1 and n to 15, than, workers set to 15 and n to 1.

It’s probably be because the recipes are not geared for this change in Radiance (yet). Likely, the ray tracing itself is quicker, but everything else will be slower with 1 worker because only one task run at a time.