Hi. Please advice me what system will has higher perfomance (for FHD 60fps):
- Great CPU i7-6700HQ + integrated Intel 530
- Average CPU like i5-8250 + discrete Nvidia MX150
thank you in advance!
Have you looked at this comparison?
If you are planning on doing GPU rendering, make sure you get a GPU that supports B frames.
There are three types of performance: preview, export (final image generation within Shotcut), and encoding.
Preview benefits the most if you get 4 of the fastest cores you can find. Shotcut preview will not use every core on the system on purpose because the memory and caching allocated for each thread would eat up all but the beefiest workstations, among other reasons.
Similarly, export benefits the most if you get 8 of the fastest cores you can find, like an i7-7700K. Shotcut does not thread enough to use every core on a 64-core system. It kinda caps out around 8. So, if you have 8 cores at 4.2 GHz, that is 33.6 GHz of export processing, whereas 16 cores at 2.4 GHz that’s thread-limited to 8 cores is only 19.2 GHz of export processing. The faster cores are… well… faster. (Note than any left-over cores not used by Shotcut can be used by an encoder like libx264.)
Software encoding performance is heavily dependent on the codec you are writing. If you’re writing H.264 or H.265, those thread very well to 16 and 32 core boxes and the higher core count wins the GHz war. However, if you’re writing a low-overhead codec like lossless or intermediate codecs, then Shotcut processing becomes the bottleneck and you’re back to needing the fast 8 cores to speed that part up the most.
Hardware encoding also benefits from 8 fast cores so that Shotcut can deliver a frame to the hardware encoder as fast as possible.
However… major caveat… hardware encoding tends to create large files with poor quality if anything less than an RTX 20xx GPU is used (compared to x264 software encoding). Personally, I still use the software encoder because I don’t have RTX, and I would probably still use the software encoder even if I did because I can push the quality and resolution much higher than hardware can.
Outside of encoding, GPUs play a very minor role in filters and preview UI. The integrated graphics chips have been able to easily handle the workloads I’ve thrown at them. Most filters run in CPU mode anyway.
In this light, an Nvidia MX150 gains you basically nothing. It will be underutilized for preview and will be less-than-adequate if your goal is high-quality encoding. People with different priorities (like fast workflow at lower quality, or live streaming) will have a different opinion of course, and that’s totally valid. It depends on your use case.
Personally, I would vote for the faster cores of the 6700HQ.
4 Likes
This topic was automatically closed after 90 days. New replies are no longer allowed.