I recently got a RX580(MSI 8gb gaming x) and hoping a significant rendering improvement.
I have been comparing rendering with GPU and without, it turns out my PC takes 1:03 to render a 3min clip using CPU alone (95%+ CPU and 0% GPU utilization). Then 1:27 when I enable the hardware encoding (~75% GPU and less than 10% CPU utilization) the same clip.
My CPU is an older i5-4690 CPU running at 3.5GHz, 16gb of RAM. running 20.02.17 Shotcut.
To be honest, I am quite disappointed seeing the graphic card upgrade didn’t help. I am wondering if I haven’t configured things properly for rendering? any suggestions, inputs are welcome. Thanks.
It can depend on the resolution and frame rate in addition to the hardware. The memory bus between the CPU (where video is decoded and processed) and the GPU is good (PCIe x16) but not awesome. So, this can be a bottleneck. Meanwhile, x264 for CPU encoding is highly optimized. So is x265, but HEVC is inherently much heavier when utilizing the features that makes it compress better. So, your hardware encoder might only be more beneficial for HEVC.
1 Like
What kind of storage are you using? 75% gpu load sounds like it wasn’t being fed fast enough. I’d also consider using a longer clip, it’s possible something like turbo boost kicked in during the cpu render but not the gpu render and wouldn’t be sustainable on say a 30 minute clip leading to a different outcome.
This topic was automatically closed after 90 days. New replies are no longer allowed.