Slashdot Mirror
MacWorld Magazine Benchmarks the G5s
La Temperanza writes "Macworld has released yet another set of benchmarks of the full line-up of G5 desktops, along with Dual 1.42GHz and single 1GHz G4s. The results are very interesting indeed, and I think I can safely say they're not biased in the G5's favor." I dunno, it should not come as too much of a shock that a dual G4 can beat a single G5 in many tests.
I am a believer of momentum and curves.
Two words that together will suck up all the resources of a machine. I think you'll see plenty of home users maxxing out their G5s once they start doing home videos. The market may swing back to the home users from corporations because the general home users do have a few apps that will need it.
The obvious ways this thing should be different are huge memory moves: the true independent DDR and fast bus means this thing can move a DVD's worth of data in ten seconds. The other way this should be better is that the processor should be able to have multiple floating moint commands being processed at once (in addition to altivec). neither of these are showing up in the app-based benchmarks.
these difference should be huge and impossible to miss. something is wrong. maybe some debug codes in the new OS or the compilers are crippling the G5.
Some drink at the fountain of knowledge. Others just gargle.
Is there really much need for so much desktop power?
In a word, Yes.
Think about all the compute-intensive tasks that were overnight jobs a few years ago, that have become real-time or near-real-time work today. Did you ever use a video editing system that made you wait to render transitions? (And I do mean *wait*.)
There are many situations where any double-digit improvement in processing speed translates directly into thousands of dollars of productivity per user per year.
-jcr
The only title of honor that a tyrant can grant is "Enemy of the State."
If using it, or even looking at it, gives you joy, and you have the bucks, what's the harm?
:-(.
You're helping keep Apple in business, so it can make more cool things, so you can buy them. If we stop buying them, then they can't make cool things anymore
That being said, for my purposes, anything that increases real time capacity and reduces rendering time in Final Cut is bound to pay off big-time. And, judging by the rest of the responses, most serious PowerMac users feel the same way.
D
No, it shows that as soon as you start swapping, a 20x slowdown is not unexpected.
Not dirt cheap perhaps but $600 is damn inexpensive considering what you're getting for your money; it's not cutting edge, but a 450 DP G4 is nothing to sneeze at.
Window resizing is non-trivial. For things like spreadsheets, you could just render a larger chunk to an off-screen buffer and clip it correctly for the screen. For most apps, however, you have to calculate the positions of the widgets (which are relative to the size of the window), and then draw them. For this to appear smooth, you must do this at least 25 times per second while the user drags the size. GPU power has very little to do with it except for in special case, like resizing a video clip (which is done entirely on the graphics card).
I am TheRaven on Soylent News
Let's not forget that these systems are running nearly identical binaries, which, while it seems fair, is not.
The binaries are optimized for the G4. Optimization for the G5 will create quite different binaries which could run _much_ faster on the G5.
While these tests are a great comparison for performance we'll see today, apps compiled with newer G5 optimizing compilers will push the top numbers even a bit farther, as will future OS updates. Users with G5 a year from now might look back on these numbers and wonder why they were so low...