What hardware do you use for editing with CinGG? Low-end; mid-range; high-end or very high-end workstations (Xeon/Threadripper, dual GPU, etc)?
What hardware problems have you encountered? What do you think is the ideal configuration?
1) High-end:
Gigabyte Z790 AE AX/ Intel 12-core i7-12700KF/ Intel Arc A750
2) Mid-range:
MSI Z170-A PRO/ Intel 4-core i7-6700K/ NVIDIA GeForce GTX 960
(Dell XPS 13 9370/ Intel 4-core i7-8550U/ Intel UHD Graphics 620)
I use an older DELL workstation
XEON 14c/28t
64 MB ECC
SDD
GTX 1080 TI
The Proxy is perfect and so you fit cin to your hardware.
Working with original files is IMHO unrealisitic - no matter what editing software. Working at ProRes or RAW or even H265 or H264 etc. That doesn't make sense for the editing / preview process IMHO because not only of the high compression but because of the frame complexity by design of these codecs too.
DNXHR pre-import formatting is quite a good thing but won't be necessary if you use the proxy.
All the interface and how cin handle files is super efficient and straight forward.
Super clean implementation of FFMPEG. So nearly any overhead.
Rendering with 28 threads is not super efficient. They all are being used but not to 100%
BTW: I can say that 8 thread CPU are fully used to 100%. So i can assure multi-core rendering is possible using multiple cores to 100%. But the CPU sclabilty is limited at the actual situation.
I think 16 threads would be a sweet spot at the moment. I don't know where the bottleneck is. I hope it is FFMPEG and future development would enhance the render process with higher CPU core counts.
For effects and fades I sometimes use the background rendering.
When you need to evaluate a critical fade or effect - just pre-render. Some effects are so time critical and I'm not sure even the fastet hardware would manage a certain frame count in realtime.
I tested the GPU rendering too and logged CPU time and temperatures.
Both types produces nearly the same render times.
The GPU method had the advantage the CPU was never throtteling and stayed much cooler.
The GPU temps raised but to a normal stress value.
So a more elegant strategy, but in reality not faster than the 14 SMT cores
I think I should check the render farm out.
If a large render job came in I would be well prepared.
I could run it it virtually from one computer right?
Are you aware that all links to the manual are broken? From Google, from other posts and from the menu.
@cincity Unfortunately, the server is currently undergoing maintenance and many parts are unavailable. The only way is to use the manual included in the appimage: launch the CinGG appimage, move the mouse cursor over any tool (for example, the “Label” button) and press “Alt+h” to open the manual at the “label” section. From there, you can go to “contents” and select “render farm” from the index. Study it carefully because it is a complex tool.
Also keep in mind that having all threads at 100% causes the CPU temperature to rise significantly. User Fairy54 posted a script that runs the threads alternately in order to maintain an acceptable temperature and achieve maximum rendering efficiency. Unfortunately, the script is located on the Bug Tracker and is also inaccessible.
Or, if you trust the website, I have posted the manual in PDF format here: https://limewire.com/d/9aKhF#5sfXRuI2E5
