Benchmarks Survey - Background Rendering vs. foreground

So last night I had to get 5 short projects up on YouTube for a client to review.I thought all 5 videos could be done in a background batch and they would all be uploaded in less than an hour. Not exactly. Here were the results:

---

1080p, 2 minute 23 second timeline, every frame needing to be rendered
YouTube Share > Background > This Computer - 1:22:43 (1 hour 22 minutes 43 seconds)

1080p, 2 minute 15 second timeline, every frame needing to be rendered
YouTube Share > Background > none - 6:47 (6 minutes 47 seconds)

If I do my math right foreground sharing is more than 10 TIMES FASTER. Wozers.

---

If you see some benchmarks floating around the web and twitter about FCP X taking hours to render a simple sequence this is probably why. As we have heard before foreground rendering uses the GPU while background is CPU only.

I would be curious how other users systems compare.

* Test Notes: MacPro 2006, 2.66, 4 cores, Radeon 5770, 13GB Ram. No other processes were running, these times were taking from the Share Monitor "i" button. These were one pass, 960x540 setting in Share > YouTube. I am not using Compressor, but since Compressor doesn't use the [strike]CPU[/strike] GPU in my early tests the results would be similar. These projects were pulling footage from the same interview, with a few ken burn's stills in each. Background rendering was off.

This great information to know!

tpayton wrote: I am not using Compressor, but since Compressor doesn't use the CPU in my early tests the results would be similar. These projects were pulling footage from the same interview, with a few ken burn's stills in each. Background rendering was off.

Which result fits to (single) Compresor? I could try to do a export over Compressor cluster if you have way to share the project with only the required media?

Opps I had a typo. I meant to say that Compressor currently doesn't use the GPU. I fixed that in my original post.

My offer for Cluster Export is still up.
Say can i use your Information for my Board?

I had a cluster going with compressor 3, but haven't got it figured out yet with compressor 4. Qmaster is now part of the compressor app. There is a setup guide in the compressor 4 docs on how to deal with storage. I believe that all you need to do is have the storage location of your event mounted on each machine, or at least accessible.

Let us know how it comes out.

Do you know about this web page?
support.apple.com/kb/HT5333

I think he was mentioning how to configure a cluster unless I misunderstood.

Yes all workstations will have to have access to the exact same storage, otherwise one of those workstations would have to copy media over to work on it, then copy media back.

The original question was speeding up Compressor, which lead to creating clusters. Clusters will speed up work depending on the codec and size of the jobs. If you're doing 2-3 minute clips, a network cluster may not show any improvement. If you're transcoding long form video projects, or multiple projects, a network cluster will help.
help.apple.com/compressor/mac/4.0/en/com...Manual-90863DPR-6A13