Weird corruption effect when resizing and deinterlacing with DGSource

[Rocky]

No! But they get very busy in the fall caching food.

[Rocky]

When I install our new 4090 I'm gonna upgrade to the latest driver. Wanna take bets on if this is still broken? Technically, I may not have to upgrade but I need 525+ to make CUDA compute-sanitizer work, so I may as well go all the way.

https://docs.nvidia.com/compute-sanitiz ... index.html

[Bullwinkle]

My bet? Still broken.

[Curly]

Speaking of compute-sanitizer:

https://forums.developer.nvidia.com/t/i ... dec/282064

CUDA guys are way more responsive than NVDec guys.

[Curly]

Hmm, right when I prove it's their problem, he goes quiet. Well, let's give him some time.

[hydra3333]

If this is it, https://forums.developer.nvidia.com/t/i ... /282064/47 they have been nearly 2 weeks.
(not complaining, just observing)

[Sherman]

Curly kisses butt pretty good when he needs to.

[eXtremeDevil]

Tested with latest drivers, still broken for me.

[Rocky]

Thanks for the info. I'm not surprised. The workaround remains: do resizing in your script after DGSource().

Maybe we need an integrated DGResize() in DGSource() implemented CUDASynth-style.

[hydra3333]

Thanks for the info. I'm not surprised. The workaround remains: do resizing in your script after DGSource().

OK, noted, thanks.

Maybe we need an integrated DGResize() in DGSource() implemented CUDASynth-style.

why not. Choices of resizer methods, including the current one ?

[eXtremeDevil]

Thanks for the info. I'm not surprised. The workaround remains: do resizing in your script after DGSource().

Maybe we need an integrated DGResize() in DGSource() implemented CUDASynth-style.

Latest firmware, same issue. I'll keep an eye to this thread in case you implement that DGResize(), I'm sure it will be interesting