What Should I Consider When Pushing Overclocking for Professional Film Editing? -- Guidance Required
What Should I Consider When Pushing Overclocking for Professional Film Editing? -- Guidance Required
Hello everyone,
I’ve just finished my initial build for a new editing rig, and it turned out to be a more economical choice compared to buying one off the shelf. As someone who works professionally in film editing, I found it hard to justify purchasing a custom setup when I could assemble my own.
First of all, thank you all for your support on this forum – there’s such valuable information here, and I wouldn’t have managed without the community’s help. Great job everyone!
I’ll share my specifications below, but my main question is: I’ve fine-tuned my CPU to a stable configuration using the standards and tests shared in the forums and guides. However, even with daily tasks like editing 4K footage, it doesn’t match the intensity of those stress tests. Should I push the overclock further, or should I keep it more conservative to protect the CPU’s longevity? Since I don’t play games, my goal is to maximize value while avoiding actions that could shorten the CPU’s lifespan.
Here’s a quick summary of my hardware:
- CPU: i9-9900K
- Motherboard: Gigabyte z390 Designare
- RAM: Crucial Ballistix Sport LT 62 GB (4 x 16 GB) DDR4-3200
- Cooler: Noctua NH-D15
- OS: Windows 10
- Case: Carbide 275R
- Current OC: 4.9 GHz @ 1.30v (Medium LLC) AVX offset -1
I ran Prime95 and RealBench tests, and the results were promising:
- Passed stress tests with no restart issues.
- Temp range: 75–85°C max, 74–81°C avg.
- Max voltage: 1.315
- RealBench showed stable performance (74–80°C avg) and slightly higher maxV than during stress tests.
For data gathering, I used HWinfo and CoreTemp.
A few notes:
- Two front intake fans keep the CPU a bit warmer than others.
- I’m thinking about moving the bottom fan to the top of the case so it can blow air over the HD trays and PSU more effectively, above and slightly behind the Noctua cooler as a secondary exhaust.
- This might improve real-world cooling, or maybe keeping the airflow in is more important.
While I’m currently working on rendering and exporting 2K video, I saw temperatures between 60–65°C across all cores (fans didn’t reach high speeds). Notably, maxV was higher than during stress tests (1.329–1.364). Could this suggest I should tweak the CPU voltage?
I know the Noctua cooler will cap my cooling performance, but if my workloads are lighter than the stress tests, can I increase the overclock? I’m balancing caution with the desire to get the most out of my investment and preserve the CPU’s health.
For anyone else who needs more details or references, I really appreciate your help. This has been a valuable learning experience!
Best regards,
[Your Name]
As soon as I noticed the thread, an image formed in my mind of a tired editor eagerly awaiting the final render-out to complete, just as the midnight deadline loomed with only a few minutes left before the computer crashed.
Haha, it's common for filmmakers to face similar challenges. Most professionals rely on redundancies and backups, yet this brings up an important question: how much does OCing affect stability? I understand the usual guidelines apply, but it seems stock intentionally restricts the CPU's capabilities. If you conduct thorough testing, are we measuring a 10% or even a more significant increase in instability?
The remaining half of the equation reflects the actual performance improvement you can expect within your specific tools and process. For a certain OC level, reductions in rendering times could range from 5% to 50%, but it's likely closer to the lower side of that spectrum. The trade-off comes with increased risk of instability.
Despite my strong dislike for Prime95 small FFT and the 'power virus', I still believe it’s the right test to put a 'production' system through before letting them profit from it. If it can’t handle 6-8 hours, reduce its performance until it improves. Depending on the installation or if it isn’t the main user, I recommend even running SABATOGE over it by covering the case and conducting the test. You need to anticipate someone using it for their coat rack.
Not really.
It also relies on the specific software you're using.
Benchmarks work well for comparing two different systems or configurations.
Not ideal when aiming for a simple "10% in Cinebench equals 10% faster rendering."
Except, if I understand correctly: Cinebench might be somewhat artificial, likely due to its interface control for a dependable benchmark tool and the score not being easily connected to real-world use. However, the rendering engine (Cinema 4D) is essentially the same as Maxon's product, so benchmark outcomes generally align with actual performance in their context. In short: scores don’t match render times directly, but higher scores consistently indicate quicker renders.