Dual G5 FCP render benchmark

No replies
Joined: Aug 31 2002

Here's the G5 figures we've been waiting for!
" This test has been skewed somewhat by the fact that on any computer FCP4 renders up to 40% faster than FCP3 or FCP2.

(FCP 2.0 or higher, render files sent to an internal HD.) Place a 10 second clip (DV25) on the timeline and apply a gaussian blur with a setting (radius) of 25. Cut the clip in half and apply a 5 second cross dissolve on the cut. Time from the click of "Render All" to the point at which the render status window disappears.

The Results:

G3 SP 300 4:45
G3 SP 350 4:35
G3 SP 400 (Blue and White) 4:06
G3 SP 350 (w/ G4 550 ZIF Upgrade) 1:45
G4 SP 400 1:40
G4 SP 800 (iMac) 1:25
G4 SP 450 (OS9.2.2/QT5) 1:21, 1:16
“ (OS10.1.5/QT6) 1:12
PB Ti/667 (1st generation) 1:12, 1:21, 2:05, 2:30
PB Ti/500 (Rev. A) 1:11, 1:12,1:15
G4 SP 500 1:10
G4 SP 733 (Original ver./not Quicksilver)1:08
G4 SP 533 1:05
G4 SP 867 (Quicksilver) :49
G4 DP 450 :47
G4 DP 500 :40
G4 DP 533 :39
G4 DP 800 :37 (FCP 3), :29 (FCP 4)
G4 DP 1gHz (Quicksilver) :28, :29, :30
G4 DP 1gHz (DDR RAM) (OS10.2.1) :28

G5 DP 2gHz :12

Or you could simply play the file in :10 seconds, because it plays it in RT in high quality with no problem.
I'll have to come up with a harder test! I personally went from 4:06 to :12 in four years."
From: http://www.creativecow.net/forum/read_thread.php?threadid=412299&forumid=8

Note that Final Cut Pro is intesively optimised i) for Altivec and ii) for dual processors - being some 70% faster on a dual setup of the same CPU speed. I'm not sure if Logic has been similarly optimised so far.

Furthermore these figures are for an un-G5-optimised version of FCP, so the speed increase is about proportional to CPU speed increase.
What further benefits will proper 64-bit code-optimisation of the OS and app bring??!!!