I’ve used the GPU acceleration within the options in SpectraLayers for some time but I have the dumb feeling that the application doesn’t use as much anymore compared to installations before version 11.0.6,
My system is a notebook using an AMD Ryzen 9 3995 16 core desktop CPU with 64 GB ram and an NVIDIA RTX 2070 card with 8 GB ram. The system has CUDA libs installed and also libs for OpenCL
Please drive your developers to add more GPGPU instructions to the code. Calculating on graphics cards is so much faster…
There is a lot to improve in the code, imho.
SL is a great concept, but it feels so buggy and somehow incomplete.
The VST-interface, for instance, is unusable for my usecases, because of the stutter step lagging. It should work at least as in Wavelab.
1 Like
Laptops are always limited thermally
SL will work fine with this. Have you checked in “Edit | Preferences | System” under “Miscellaneous | AI Processing Device” that the GPU card is selected, and not the CPU?
SL only need the basic NVIDIA driver installed to work, preferably the “studio” driver, and at that, only the basic driver itself, no add-ons.
My Notbook has a dedicated heat pipe and it’s loud during render jobs. It manages heat quite well.
I already mentioned that GPU options have been activated
The studio driver is in use
what processes are you running for which this an issue? Unmixing modules? Unmixing music? VST applications to a portion of the freq spec?
Manual tasks are absolutely incredible IMO
but I am NOT unmixing music
I’m doing manual NR and re-balancing and SL is mind-boggling good at my needed tasks; Unmix Noisy Speech is a godsend and processes quite quickly on my 7 year old i7 desktop machine
1 Like