Can ImageSharp take advantage of dedicated GPU?

[AlessandroMartinelli]

As per title.

By taking a look at issues, it seems the answer is yes. However I'd like to know up to what extent is the gpu used by the library.

Thanks

[aloksharma1]

As per title.

By taking a look at issues, it seems the answer is yes. However I'd like to know up to what extent is the gpu used by the library.

Thanks

did you find the answer to it, i need realtime image manipulation imagesharp is cross platform hence i am very eager to know this myself.

[AlessandroMartinelli]

@aloksharma1 sadly no, I never found an answer to it, thus I'm not using ImageSharp.

[aloksharma1]

whats the alternative? i found computesharp any other easy to use library like imagesharp?

[antonfirsov]

ComputeSharp has a very different purpose: it enables you writing shaders in C# on Windows. ImageSharp is a cross-platform general-purpose image-processing library.

[antonfirsov]

ImageSharp is CPU-only. I'm wondering what are your use-cases?

[aloksharma1]

my use case is real-time video feed processing, the video computations are done on cuda side then getting it on ImageSharp makes it loose performance due to cpu->gpu->cpu transitions of data this made me loose around half my fps on feed, anyways i also found that SkiaSharp can use Gpu so i have started implementing that now.

[antonfirsov]

real-time video feed processing

Feels like this needs a video processing library rather than an image processing one. Have you considered FFmpeg? Looks like there is a .NET wrapper.

[aloksharma1]

my impression of ffmpeg was a video encoder decoder, can it also draw on frames? i need to show detection boxes and some other info on frames.