Much higher RAM usage (2-3 times) compared to FastSDCPU when using the exact same models/settings

[JohnAlcatraz]

Much higher RAM usage (2-3 times) compared to FastSDCPU when using the exact same models/settings

[FSSRepo]

Currently, im2col is being used for convolutions, which consumes a very high amount of RAM during the VAE phase.

I have been working on a kernel that merges im2col and matrix multiplications to avoid materializing a lot of data in memory, although that entails a 40% performance reduction. So far, I am only doing this for CUDA; for CPU it will be more difficult and will likely have a negative impact on performance.

[JohnAlcatraz]

Currently, im2col is being used for convolutions, which consumes a very high amount of RAM during the VAE phase.

But I did my comparison with TAESD instead of the VAE, so I think that means the VAE isn't used at all? TAESD is super lightweight already.