scrolling is not smooth and in an uncontrolled manner (a swipe brings you somewhere much further within the collection 2-3 swipes can bring you at the end of the collection two down and two up will not bring you to the same images/row) thumbnails start popping up (being displayed one after the other) only after import has finished images get imported without showing any thumbnails during import process Open a preview (selecting one image and pressing "w") in a 5 images per row grid (by using trackpad or mouse wheel)ħ.
_Note:_ Due to default setting, OpenCL is not activatedĦ. Do a fresh install of 3.2.1 / 3.0.2 without changing any configuration. Clean config/cache (delete ~/.config/darktable and ~/.cache/darktable)ģ. Remove any previous /Applications/darktable.appĢ. about 800 RAW images (.NEF) of a Nikon D750ġ. DMG from Darktable's official GitHub releases Darktable 3.2.1 / 3.0.2 directly downloaded as. ope … nCL devices: 'Intel(R) UHD Graphics 630' (device 0), 'AMD Radeon Pro 5500M Compute Engine' (device 1) Maybe that’s why the auto magic wasn’t happening. Thanks both for your help.ĮDIT: I read in the manual, “A sufficient amount of graphics memory (1GB+) needs to be available for darktable to take advantage of the GPU.” My intel gpu has 1536mb while nvidia only has 512mb, so I guess sticking with the integrated gpu is the way to go anyway. When resolution = 1 it takes ~ 1.6 secs, and when the scope is waveform it still takes ~15 secs, so no doubt just changing from waveform to histogram has been the biggest change for me. The changes I made were to run your opencl script, use histogram instead of waveform, and reduce resolution of preview image to 1/4. Compared to ~16 secs in the doc I uploaded earlier, this is a huge improvement. So, no real advantage it seems in having discrete. However, running -d perf on similar processing as earlier test gives pixel pipeline processing ~0.4 secs for both cards. I thought the overview told me what I had, but it did not tell the whole story. It’s funny, I have known the importance of opencl for some months, and spent quite a bit of time researching gpu’s for my next machine (linux), but did not know my own machine.
If one wants best performance there‘s no shortcut … Theres a whole chapter in the manual on OpenCL and tewaking.
0 kommentar(er)
