Closed ghost closed 3 years ago
I use the latest release from the zip and externals are downloaded.
ok I've been trying to use D2X on multiprocess. I just tried on singleprocess and now it worked in 15 min. Idk why that is but now it works at least.
Can you share your computer specs? It might be the case that your computer cant run 3 dandere2x sessions at once, and that its actually slowing stuff down.
Intel i5-7500 (4c 4t) 16 GB DDR4 3000 GTX 1060 6GB Everything SSD
I'm taking a stab at this, and I'm not sure how python's multi-thread works, but I use a bunch of threads in dandere2x-single process, so 4c 4t might not be enough. For reference, there's an ffmpeg instance + dandere2x_cpp instance + 5 subthreads for each dandere2x process, so we might run into resource-allocation racing conditions by using the multi-process mode on your setup.
Yeah that could be the problem indeed. Well, this just gave me another reason to buy the Ryzen 9 5900X lmao. Btw. I love your program! I wish I could participate in it in the future :)
Should I close this thread or keep open if you want to make optimizations?
I'll close it out here. For what it's worth, I'm on a 3900x right now :] and it's super satisfying seeing all the cores get used lmao.
So I have been trying to upscale a clip using D2X. The clip is about 20 sec long and it seemed to be doing fine until this point:
I have waited for 2 hours with nothing new. So then I checked if Dandere is doing something but Task Manager doesn't say so :/
And when I exited Dandere this error shows up. (It seems to be working but at the same time not? -> It is hung up at split_video2 but it crashed at split_video1?)
Oh, and it doesnt matter if I use caffe, conv-cpp or vulkan.
Idk should I wait even longer?