Bram Stolk

Ah, yes, I was feeding it a colour image. 

I was wrong about not caching, cuda_eval() does take less when called repeatedly for same code+data. 170ms when shift value changes, and 17ms when the shift value does not change....

I see in **jit.cu** that the call to `cuLaunchKernel()` passes null for the `kernelParameters` argument. So how are calls to the kernels parameterized? https://github.com/mitsuba-renderer/enoki/blob/master/src/cuda/jit.cu#L1372

> ...another random observation, switching off vsync makes a dramatic difference, but that's probably too wasteful... Yes. A hand full of frames lag hardly matters if you render at 1000+...

> PS: what system are you actually on? I just noticed that your mouse pointer looks more like Mac :D I need to do the same investigation there. Ubuntu

> > but I read sokol does not have a fullscreen toggle, is that still the case? > > sokol_app.h has a toggle-fullscreen feature now, but doesn't support switching the...

It turns out that it sometimes returns values well outside the -1 .. 1 range?

Did your computer clock change recently? On Wed, Jan 5, 2022 at 4:58 PM nothingshocking ***@***.***> wrote: > I have the same issue. Has there been any investigation or resolution...

The tool has just the harvester log at its disposal, when running a remote harvester. I checked, it simply does not have the req'd information. I think I will suppress...

Thanks for reporting. Does the dot turn cyan if you resize the terminal (w tool running) to be shorter in height?