Announcement

Collapse
No announcement yet.

CPU vs GPU image quality with Enlarger AI

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    DVDFab Enlarger AI for Copy CPU vs GPU image quality with Enlarger AI

    When I am converting a BD to UHD with Enlarger AI, an encoded in x265 is made. So, will better image quality be obtained using the CPU?

    I suppose that same answer will be valid for DVD to BD case.

    Thanks

    #2
    CPU produced better images in Ripper for DVD and BD sources, not sure if this is a practical choice for AI due to the time and volume of computations required
    Supplying DVDFab Logs in the Forum ...........................User Manual PDF for DVDFab v11................................ Guide: Using Images in Posts
    Supplying DMS Logs to Developers................................Enlarger AI FAQ.....

    Comment


      #3
      Originally posted by signals View Post
      CPU produced better images in Ripper for DVD and BD sources, not sure if this is a practical choice for AI due to the time and volume of computations required
      Many thanks. In my case I have an i9-10980XE (18 cores / 36 threads). It is a powerfull CPU. With it I can get from 6 to 8 FPS when enlarging BD to UHD. I normally open two instances of DVDFab Enlarger AI because with only one of it the CPU usage is aroung 85%. With two instances working in parallel the CPU usage is 100% constantly.

      But, in the future I plan to buy a RTX 3090 24GB. But I am concerned about power consumption. I don't know if the power consumption / performance ratio will be better or worse than for the i9-10980xe. When this CPU is in full load with Enlarger AI it consumes around 175W sustained. And I can get 6 to 8 FPS. But the RTX 3090 consumes much more in full load. nVIDIA says around 350W, which is double that in the case of the i9-10980XE. If I could get a lot more FPS than 8, then it would be worth it. And that's not counting image quality, which I already suspected would be better for CPU than for GPU.

      Thanks

      Comment

      Working...
      X