Posted 7/25/03.
Updated 8/27/03.
Updated 9/13/03 with After
Effects 6.0
results.
Updated 11/17/03 with G5 aware versions of
Photoshop 7.01, Cinebench 2003, Unreal Tournament
2003, and AltiVec Fractal.
Many of you are
asking how the Dual 2Ghz G5 compares to the Dual
1.42GHz G4. I've gathered up test results from
various sources around the globe (New York, Italy
and Germany) along with tests run in our own
lab.
Photoshop 7.01
results with the G5 Plug-in.
|
G4/1.42GHz
MP
|
G5/2GHz
MP
|
G5
advantage
|
Non
MP-aware actions (seconds)
|
73
|
53
|
38%
|
MP-aware
actions (seconds)
|
35
|
28
|
25%
|
Bryce 5 render
of Beach Chair. (The advantage of the G5 can be
attributed to heavy use of L2 cache and
FPU.)
|
G4/1.42GHz
MP
|
G5/2GHz
MP
|
G5
advantage
|
Beach
Chair render (seconds)
|
37
|
22
|
68%
|
Cinebench
2003
CPU Render. (The G5 numbers were updated with
the new G5 aware version of Cinebench.)
|
G4/1.42GHz
MP
|
G5/2GHz
MP
|
G5
advantage
|
Multiprocessor
CPU Render (seconds)
|
107
|
51
|
110%
|
Adobe
After
Effects 6.0
Tryout.
(Ten frame render of Nighflight
sample file.)
|
G4/1.42GHz
MP
|
G5/2GHz
MP
|
G5
advantage
|
Ten
frame render of Nighflight
(seconds)
|
593
|
361
|
64%
|
BotMatch in
Unreal
Tournament
2003.
(Average of Anatalus and Asbestos at 1024x768 Max
settings. The G5's numbers have been update with
the latest patch of UT2003.)
|
G4/1.42GHz
MP
|
G5/2GHz
MP
|
G5
advantage
|
BotMatch
(frames per second)
|
30
|
65
|
117%
|
MacTechNews'
SpeedTest
benchmark
is new to me. I can't comment on the validity,
but I was most impressed by the OpenGL portion of
the test. It seemed more useful than the OpenGL
test on Xbench.
|
G4/1.42GHz
MP
|
G5/2GHz
MP
|
G5
advantage
|
CPU
rating
|
326
|
606
|
86%
|
QUARTZ
rating
|
238
|
379
|
59%
|
OPENGL
rating
|
360
|
565
|
57%
|
AltiVec
Fractal
test on a G5/2GHz MP (from New York).
65K
points
|
G4/1.42GHz.MP
|
G5/2GHz
MP
|
advantage
|
Gigaflops
|
10.8
|
14.8
|
37%
|
|