Does the new Background Export use all cores now?

It used to be the case that if you Exported Media on the foreground, FCP X used all your CPU and GPU power to do it. If you sent to Compressor or used the Share Monitor, it used less resources and had much bigger exporting times. Like the background render did.

Now there are 2 things new in FCP X 10.0.6.
First, we have full background export with 'bundles' (batches) and the ability to work further while exports are going.

Secondly, there's an article on the site here ( www.fcp.co/final-cut-pro/news/954-10-0-6...l-cut-pro-x-timeline ) where it says the Background Render now used the GPU, leading to some massively shorter render times.

So my question... Does the background Export now take full utility of all the CPU and GPU cores like it used to do in an Export on the foreground? Meaning, it automatically scales the power it frees for the export in the background depending on the work you are still doing in the program, but using ALL further power besides that?


And I guess the background render still stops whenever you touch the timeline?

Does anyone have an answer?

Much Appreciated,

I'd like to know more about this also.

I have noticed on my MBP i7 Quad-core that I only use 45-50% of my CPU power whilst exporting via share options..

And the same with a MP 8-core / 12-core machines.

The CPU's take about 1min to crank up to 45-50% once a export is started.

From the hardware standpoint, that 45-50% figure is using all physical cores and no virtual cores.

Most Macs have Intel processors with Hyper-Threading (HT). HT is a nice way of running 2 simple tasks concurrently, however falls flat on its face if you ask it to run 2 intensive tasks.
Take an 8 core system, thats another 8 cores from HT, giving you 16 in total.
Since video work is intensive, scrap the virtual cores and you are left with 50% of your total.

The processor usage is calculated using all cores available, which is 16, thus with only 8 being used shows a maximum of 50%.
Yet this method seems to allow you to happily browse the web and continue working in fcpx.

Anyway, as odd as it sounds, by only using physical cores the export should be faster.
Also take into account the processor has a fixed size of cache on the chip and trying to deliver the data to all cores when your processing large amounts will slow it down even if all virtual cores are used.

At least, that is my understanding and technical input.