-
FCP X hardware performance
-
Did you have Premiere set to use highest bit depth (32 bit)? I’ve found renders out of X to be super clean – colors come out exactly like you create them and Blu-rays look better than anything I’ve rendered out of any NLE. I don’t mind waiting a little longer if the quality is as good as it gets.
Lance Bachelder
Writer, Editor, Director
Irvine, California -
This exactly what I have found. Playback on a MacPro 3.33GHz, 6 Core 2010 Macpro and 24 GB shows the CPU idling at almost 80% when playing back a three camera angle Pro Rez 1920 x 1080 multi cam clip. Playback constantly drops frames and is even worse when an LHI AJA Kona card is brought into the mix. I remember this became much worse when FCPX 1.03 and 1.04 were released. By contrast a 2.93 GHz 7iCore IMac with only an 800 Firewire and 12 GB of RAM plays back the same clips without any problems
-
I’ve got journeyman chops but one thing I have learned is to avoid TIFFs (and jpegs where possible). I always convert to PNG files. Smaller and preserves transparency. I know that was very helpful in FCP7. And it has been my experience in X so far as well.
Actual mileage may vary, but I would avoid TIFFs if possible.
-
[Michael Hadley] “I’ve got journeyman chops but one thing I have learned is to avoid TIFFs (and jpegs where possible).”
JPEGS, yes, but I use TIFFs all the time. Its a venerable format that allows for layers and transparencies. I’m not sure what your problem has been with it, but I’ve been using it for years without issue.
-
[Oliver Peters] “Basically, Apple argued for years with the PowerPC (G5) that pure processor speed didn’t matter. In fact, most objective tests proved this to be flat out marketing BS. Speed does matter and it’s true with Xeons as it was with PowerPC chips.
If performance is more important than elegance, you WON’T go with an Apple computing product. Want to get the most out of 3D apps or After Effects? Run it on an HP under Win 7 64. But, for editing, Apple makes machines that are powerful enough for most of our needs.”
And this has been the crux of the issue for years, and truly became apparent when Apple moved to intel and more direct comparisons could be made not too long ago.
Apple has always favored “system stability” over raw power. Windows seems to have caught up, I don’t know, the windows machine here controls our SAN and that’s all it has to do, but it was true not too long ago from my anecdotal connections with MacOS and Windows.
Windows machines using the same application where usually much faster, but crashed much more often.
So then the argument became that Macs had better “uptime” and the notion of speed then became about overall job speed instead of direct CPU speed. This is probably not true anymore.
-
[Jeremy Garchow] “Apple has always favored “system stability” over raw power.”
I disagree with this statement for three reasons.
First, stability and power are not opposing sides of a tradeoff. Not only is it possible to engineer for both, it’s almost always necessary — what high-performance system can sacrifice stability?
Second, Apple does offer raw power. Apple has twice gotten earlier access to faster Xeon CPUs than was available to the competition (the 3.0 GHz quad-core Xeon Clovertown that powered that 2007 Mac Pro, and the Nehalem Xeon 35xx and 55xx series that powered the 2009 Mac Pro). Apple has also offered top-tier Xeon processors with each new generation of Mac Pro. Every Mac Pro has been very competitive, both in terms of absolute CPU power and cost, at the time of its launch.
Third, Apple has done some really interesting work in advanced computing, especially related to clustering (XGrid and QMaster) and symmetric multiprocessing (early multiprocessor PowerPC systems and today’s Grand Central Dispatch).
I could certainly agree that Apple does not prioritize performance, as they consistently fall behind the competition on mid-architecture updates, as well as continuing to tolerate their often-noted GPU and PCIe deficiencies, but they do have a long history of (pardon the pun) power computing.
Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog – What I’m thinking when my workstation’s thinking
Creative Cow Forum Host: Live & Stage Events -
[Walter Soyka] “First, stability and power are not opposing sides of a tradeoff. Not only is it possible to engineer for both, it’s almost always necessary — what high-performance system can sacrifice stability?”
So why had it never happened?
[Walter Soyka] “Second, Apple does offer raw power. Apple has twice gotten earlier access to faster Xeon CPUs than was available to the competition (the 3.0 GHz quad-core Xeon Clovertown that powered that 2007 Mac Pro, and the Nehalem Xeon 35xx and 55xx series that powered the 2009 Mac Pro). Apple has also offered top-tier Xeon processors with each new generation of Mac Pro. Every Mac Pro has been very competitive, both in terms of absolute CPU power and cost, at the time of its launch.”
But the OS isn’t optimized for it, Walter. We have had this discussion sooo much. Instead of singling me out for it, reread what Oliver said, and have a look here: https://forums.creativecow.net/readpost/335/35669
Apple is not in the CPU race, my guess is from their standpoint, they don’t have to be and take a different philosophy about what “speed” means.
[Walter Soyka] “Third, Apple has done some really interesting work in advanced computing, especially related to clustering (XGrid and QMaster) and symmetric multiprocessing (early multiprocessor PowerPC systems and today’s Grand Central Dispatch).”
But does it really go faster? Sure, for some things I’m sure it does, but do you have tests on this sort of thing? What editor setup XGrid? QMaster is very hard to mange when using multiple machines. It does work well on single machines (virtual cluster), but it is for one application, Compressor. It used to be for Shake. FCP Legend doesn’t use it, and X only uses it when using Compressor. How does it help me when I’m Ae or any other application that is relying mostly on CPU cycles?
Macs have always been slower even if they have the same parts. I’m sorry, it’s true.
[Walter Soyka] “I could certainly agree that Apple does not prioritize performance, as they consistently fall behind the competition on mid-architecture updates, as well as continuing to tolerate their often-noted GPU and PCIe deficiencies, but they do have a long history of (pardon the pun) power computing.”
You mean “power computing”. Sure, they use the same parts, but the actual speed tests always tell the tale.
-
[Jeremy Garchow] “But does it really go faster? Sure, for some things I’m sure it does, but do you have tests on this sort of thing? What editor setup XGrid? QMaster is very hard to mange when using multiple machines. It does work well on single machines (virtual cluster), but it is for one application, Compressor. It used to be for Shake. FCP Legend doesn’t use it, and X only uses it when using Compressor. How does it help me when I’m Ae or any other application that is relying mostly on CPU cycles?”
Errr… you can use Qmaster to distribute render jobs for AE. Also for Nuke rendering. And it speeds things up significantly.
——
“You also agree that you will not use these products for… the development, design, manufacture or production of nuclear, missiles, or chemical or biological weapons.”
iTunes End User Licence Agreement -
[Frank Gothmann] “Errr… you can use Qmaster to distribute render jobs for AE. Also for Nuke rendering. And it speeds things up significantly.”
You mean through the multiple Ae instance thing? If you only have one computer, what’s the point? Is it going to help? Are you saying that Macs are faster at distributed rendering than similarly specced Windows machines?
Distributed rendering is one thing, it is a way to split up a job to as many machines as you have. It would really depend on what kinds of renders you are doing. If they are short, less complicated renders, it can take just as long to copy all the necessary files to the clusters as it does to simply render the video.
It’s all relative, and it still has nothing to do with how fast OSX runs over windows on the same hardware.
Jeremy
-
You should mail that to Apple.
I’ve seen some similar results (I posted in a thread about an article of Apple about when to optimize Media, recently), and to be honest, it’s awful.Yes having so much real time effects and background rendering and importing is nice, but having export times that are FAR longer than the competition isn’t up with their marketing of having a 64bit new code multicore future NLE. A lot of the culprit seems to be in working with XDCAM but If FCP7 could do it fast, FCPX should too.
Maybe the announced native MXF support later this year Will help.
But Olivers test proved that EVEN when working with Prores (THE code to get best performance in FCPX?) you often get slow exports in comparison with the competition.
And/or Red Giant has terrible implementation of their plugins for FCPX, also possible.
End result is the same though.
Reply to this Discussion! Login or Sign Up