-
FCP X hardware performance
-
[Oliver Peters] “Clearly it looks like: a) X is optimized for the i7 processor, b) actual processor speed is more important than the number of cores, c) the more VRAM the better.”
I’m not so sure about processor “speed” so much as the capability of the processor at the time they were designed.
How would a 2010 (not 2011) quad i7 BTO iMac compare to a 2010 MacPro 12 core. If it’s truly i7 optimization, you’d still see the iMac and MacPro of the same year, leading. Or, to be more accurate, compare i7 and Xeon of the same number of cores and same Ghz with 1GB VRAM to see the role the CPU is playing.
-
[Craig Seeman] “I’m not so sure about processor “speed” so much as the capability of the processor at the time they were designed.”
While in general I would agree, I do have to say that Apple has never taken good advantage of multiple cores (as in hyper-threading, multi-threading, etc.) for most of the apps, especially editing. In most benchmarks, a faster speed quad will outperform a slower 8-core of the same vintage processor. At least in the Mac Pros, which is the only place in an Apple product where this comparison exists.
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.
– Oliver
Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com -
FWIW – I was just able to run a test with the same clip and filter in PProCS6 on the 12-core Mac Pro. Export/render using the AME function. Took approx. 2:30, which is under half the time that the equivalent function took in FCP X using the exact same machine.
– Oliver
Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com -
[Oliver Peters] “I was just able to run a test with the same clip and filter in PProCS6 on the 12-core Mac Pro. Export/render using the AME function. Took approx. 2:30, which is under half the time that the equivalent function took in FCP X using the exact same machine.”
Champagne for the house! Put it on Dennis’ tab.
David
AllinOneFilms.com -
[Oliver Peters] “While in general I would agree, I do have to say that Apple has never taken good advantage of multiple cores “
Agreed.
I am wondering how an i7 and Xeon of the same vintage compares though.
If each have same cores and speed, Xeon should still be faster in some functions. I’d love for someone(s) to do such testing.My concern about tests of different models from different years are that it may well be that, for example, a 2011 Quad i7 iMac is faster in some functions than a 2008 Quad or maybe even Octo Xeon MacPro.
If Apple updates both the iMac and replaces the MacPro this year, I’m sure we’ll see such tests.
Tangentially, on another forum someone was comparing FCP7 and FCPX export on a MBP 13″ and, without going into all the details, FCP7 was faster. My own speculation is that FCP7 doesn’t really care about the GPU whereas the more resource hungry FCPX isn’t given much with the integrated GPU of the MBP 13″. A claim that FCP7 is faster would have to be qualified to the system. Heck I bet FCP7 is much better on a Core2Duo than FCPX. Given how little RAM is used FCP7 might even appear faster on system with only 4GB RAM. FCPX would starve on such a system.
Basically my point is, when doing speed tests there’s several things that need to be considered. I would be carefully saying FCPX is “optimized” for i7 unless it’s compared to a like Xeon, not older ones.
For example, would expect an 8 Core 3GHz Xeon 2006 MacPro to be faster than a 4 Core 2.8GHz i7 2011 iMac since 2006 MacPro has higher GHz and more cores?
I might guess that the 2011 iMac would be faster (at least with regards to many FCPX functions) because of the improvements in the new i7 processors compared to 5 year old Xeons. I don’t know this for a fact but that’s why I bring up vintage as an important consideration when testing.
-
[Craig Seeman] “I am wondering how an i7 and Xeon of the same vintage compares though.”
Well, if you look back at the tests I ran, The 12-core MP and the iMac are fairly close in vintage.
[Craig Seeman] “without going into all the details, FCP7 was faster”
I’ve done similar tests and seen similar results.
[Craig Seeman] “For example, would expect an 8 Core 3GHz Xeon 2006 MacPro to be faster than a 4 Core 2.8GHz i7 2011 iMac since 2006 MacPro has higher GHz and more cores? “
I think speed and cores per processor matter. In other words I have a sneaking feeling that you aren’t getter double the bang just because you have two chips instead of one.
[Craig Seeman] “I might guess that the 2011 iMac would be faster (at least with regards to many FCPX functions) because of the improvements in the new i7 processors compared to 5 year old Xeons.”
I would agree. I think in the case of FCP X specifically, the GPU is a huge factor.
– Oliver
Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com -
I believe playback AND rendering is being handed off th the GPU as in CS6. The real reason the iMac is winning. Put something better in your 12 core than a 6970 and it should win.
FCP 7 only utilized 4gigs of RAM and 1 processor. Didnt utilize GPU. The best machine would be 6 core 3.66 Mac pro for it. Motion, shake and compressor however utilize all cores and Motion also utilizes the GPU I think.
Many tests show X also to render slower than 7, but they’re usually comparing it to an 8 bit render. I think the renderer and RT in X is of a higher caliber and 10bit. Or something to do with core foundation…
-
[Bret Williams] “Put something better in your 12 core than a 6970 and it should win”
I’m afraid you confused the two cards. This 6970M IS the card in the iMac. It’s a 2GB VRAM card. The 12-core has a 5870 card (1GB VRAM). This is the top-of-the-line Apple-approved card. You would have to get a higher-end NVIDIA and throw out the card bought from Apple.
– Oliver
Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com -
[Bret Williams] “but they’re usually comparing it to an 8 bit render. I think the renderer and RT in X is of a higher caliber and 10bit. Or something to do with core foundation…”
Actually, no. That’s based on codec and effect. MB Looks rendering ProResHQ in and out is 10-bit. No difference between 7 and X in that regard.
– Oliver
Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com -
[Bret Williams] “FCP 7 only utilized 4gigs of RAM and 1 processor. Didnt utilize GPU. The best machine would be 6 core 3.66 Mac pro for it. Motion, shake and compressor however utilize all cores and Motion also utilizes the GPU I think. “
I don’t know where that myth comes from that FCP only uses on CPU. Totally untrue.
——
“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
Reply to this Discussion! Login or Sign Up