Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Macbook Pro VS Mac Mini: iMovie Rendering speeds

K

krypticos

macrumors regular
Original poster
Nov 9, 2007
237
0
hello all,

i just purchased a mid 2011 mac mini i5 for $300 bucks which i thought was a really good deal. anyways on to my reason for my post. i didnt really know where to post this so i thought i would try here first. so i do simple edits in youtube and i made like a little benchmark test think with an edit in a select video with a select edit. for more detail on the benchmark thing. its a 1080p video that is 2min55seconds long. i add a transition in at 13 seconds in. then time how long it takes to spit out a file.

so i ran this test on my hackintosh i7 2600k and it took 5min 45 second. ran same thing on macbook pro early 2011 with i7 2635qm took 1min 26 seconds. ran same test on mac mini mid 2011 with i5 2415m took 1min 25 seconds.

ok so my question is why did the mini and the macbook pro take the same amount of time. i would think the macbook pro would accomplish this faster since it is a superior machine in the since of cpu with 4 physical cores and 8 total threads vs 2 physical cores and 4 threads. unless that isnt the limiting factor in this question.

Does imovie use quicksync to render videos, because that is really the only thing i can think that is the same between the 2 machines. i know that this seems to be a stupid question to me but i just want to understand how these to machines perform the same task at the same rate of speed.

i mean for what it boils down to my macbook pro with 16gb of ram a 240gb ssd with a quad core i7 performs just as good as a mac mini with 6gb of ram and a 5400rpm hdd and a dual core i5. for this one task that i do a lot.

Thanks for your help in advanced
Anthony
 
Last edited:
after setting here thinking about it. it has to be using quicksync that is the only way it would render in the same amount of time. is there a way to disable quicksync on the macbook pro? to force opencl rendering on the amd card vs quicksync.
 
krypticos said:
after setting here thinking about it. it has to be using quicksync that is the only way it would render in the same amount of time. is there a way to disable quicksync on the macbook pro? to force opencl rendering on the amd card vs quicksync.
Click to expand...

Perhaps it has something to do with the video card.
 
It would be good to find out if it's possible, because the API for Quicksync does not focus on high quality when in comes to h264 processing. It's meant more for speed.

It's fine for Youtube uploads and such, but for archive quality video it's not good.
 
Altemose said:
Perhaps it has something to do with the video card.
Click to expand...

Video cards have pretty much nothing to do with rendering.

If I had to guess I'd bet that as a consumer app iMovie doesn't take much advantage of multiple cores so your render isn't really using the power that the MBP has. Both machines are also writing really quickly to the SSDs so there's no difference there.

You might try a bigger file and see if there's more of a difference.
 
fig said:
Video cards have pretty much nothing to do with rendering.

If I had to guess I'd bet that as a consumer app iMovie doesn't take much advantage of multiple cores so your render isn't really using the power that the MBP has. Both machines are also writing really quickly to the SSDs so there's no difference there.

You might try a bigger file and see if there's more of a difference.
Click to expand...

I would try a 15 minute 1080p video.
 
Try posting a screenshot of the activity monitor while you render the video. It will let people see exactly how many threads are being used for the render. (do it on both systems. It could be that iMovie is only using a couple of cores and Turbo Boost rather than all of them and Hyperthreading).
 
Last edited:
An activity monitor screenshot would be helpful. I don't think iMovie takes advantage of the additional i7 cores, so the rendering differences comes down to CPU speed. I forget what Ghz both machines run at. FCP seems to use a max of four cores, where video transcoders like Handbrake use all available (8 for the i7). If you want to compare the raw power of the two machines, use handbrake to convert a 30 minute video. The i7 will probably do that in half the time. Handbrake is available as a free download.

Otherwise, its a software limitation. The software must be programed to use multiple cores if available. If you run several programs simultaneously, more of the i7 cores will be used up and you will see more of a performance difference between machines.
 
There is a way to disable Quick Sync. QS only works with single-pass encoding. It will not work with two-pass encoding. I don't know about iMovie (don't use that) but in FCPX you can choose between Faster Encode (single-pass) and Better Quality (two-pass).

The difference in speed between hardware encoding and non-hardware encoding can be quite significant. An older machine encoding single-pass H.264 may outperform a newer machine encoding two-pass. In fact, since Quick Sync is not available on Xeon CPU's (workstation class), I would not be surprised if an i5 Mac mini would even outperform a 4-core Mac Pro in single-pass.

Two things may be happening here. Either the Hackintosh is encoding 2-pass, or Quick Sync is not supported in Hackintosh machines (like how OS X does not support TRIM on 3rd party SSD's). The latter is less likely, but I wouldn't be surprised if it was the case.

Other apps, like Handbrake and Compressor, also have 2-pass (or Multi-pass) encoding options to try out and compare. There's actually some nice info about this here.

lol, saying multi-pass makes me think of this

leeloo_multipass.jpg




EDIT: I misinterpreted the OP :D so to answer the OP, yes, it's very well possible that two computers using the same CPU architecture yield similar results in single-pass. However, the MBP will be faster in two-pass encoding.
 
Last edited:
thank you all for your great replies.

from what i have been finding online it looks like they both are using quicksync and sadly with imovie 10 you can't export using anything else but quicksync on quicksync enabled macs from what im seeing.

which for me is good because youtube videos get compressed to hell and back anyways so taking 3 min to render a video is really nice i just expected the mbp to be faster until i realized that it was using quicksync so it doesnt really matter.

now if i was making a film or something that needed quality i would switch back to imovie 9 for multi-pass encoding.

Again Thank You All
Anthony
 
krypticos said:
so i ran this test on my hackintosh i7 2600k and it took 5min 45 second. ran same thing on macbook pro early 2011 with i7 2635qm took 1min 26 seconds. ran same test on mac mini mid 2011 with i5 2415m took 1min 25 seconds.
Click to expand...

I find it odd that the hack seems to be exponentially slower.

Care to tell more about your build?

edit: just saw your last post
 
Last edited:
dylin said:
I find it odd that the hack seems to be exponentially slower.

Care to tell more about your build?

edit: just saw your last post
Click to expand...

My hackintosh build

Motherboard is a z68ma-d2h-b3 rev 1
CPU is i7 2600k
16gb of ram
Main drive is a 7200rpm HDD
Gpu-2xGTX 660 ti
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back