RAW files from the Nikon Z9 are Raw processed with DxO Pure Raw 2's Deep Prime process and then exported back into LRC when I work on them. I run a Mac Studio with an Ultra Chip and run Adobe LRC/PS, Dxo Pure Raw 2, DxO Photo Lab 5, Capture One 22, Phocus, Nikon NX Studio, AND all 3 Topaz AI products + LCP and DVR18 (using Rosetta) for Video work.Īll my images are imported and managed using LRC.
and a more streamlined installation process on the Mac." Agian implying they also have done something.
Similarly - Topaz State "Native Apple M1 silicon support - Experience 20-30% performance improvement. If you go to DxO's they market PR2 to run "– 4 times faster on an Apple Silicon Mac", which implies they have done something to use the M1's capabilities. You will have to look on their web pages. BUT NO I do not now if these have been optimised to run on Mac Silicon. Another thing is that even though my Linux laptop has a faster GPU (GTX 1060) normal desktop operations seem slow compared to M1 Air. What I notice is that it takes a while for my low-end M1 Air to load a job into the NE, but when it starts, it finishes rapidly. Other YouTube videos show the Ultra isn't super helpful for photographers.Īs I recall, there were hints that some Topaz applications can spread a job between NE and GPU but numbers were inconclusive, within the range of sampling error. M1 Ultra was not measured, but there's very little difference between M1 Pro and Max. Here's a video that answers some of your DxO questions. Do either of these programs have the ability to split up a batch job, processing some of the photos using Neural Engine acceleration, and others using GPU acceleration, in order to keep the SoC as busy as possible and reduce total time? If so, how well does this work for each of the four SoCs? Does DXO PL5 DeepPrime take advantage of the extra Neural Engine on the M1 Ultra?ģ. The M1, M1 Pro, and M1 Max all appear to have the same Neural Engine.
How much in the way of extra GPU cores (M1 Pro, M1 Max, M1 Ultra) does it take before the GPU path on a higher-end processor matches, or significantly exceeds, the performance of the Neural Engine path on the M1?Ģ. On the M1, users have reported that DXO PL5 DeepPrime runs faster using the Neural Engine Path than using the GPU path. Does anyone know from personal experience if the DXO DeepPrime or Topaz DeNoise noise reduction programs are updated to take advantage of the power of the latest M1 processors (M1 Pro and M1 Ultra?ġ.
0 kommentar(er)
