Hi.

I'm considering looking for a G5.
But how much faster is it actually compared to a G4 1.67GHz.
The G5 I can have right now is dual 2.0 GHz. Or should I wait for a dual 2,3GHz if that makes quite a difference.


Manfred


[ Edited by asrael22 30.08.2017 - 21:43 ]

I can offer to run some benchmarks on my 2,3 GHz G5 MorphOS/Linux-wise, so you can compare the numbers. If these numbers are not already available in some thread.

Quote:

Yeah, why not. As long as I can run the same benchmarks on my G4. Preferably on MorphOS.

I see you have a G5 11.2.
You don't use that with MorphOS, or?

> I see you have a G5 11.2. You don't use that with MorphOS, or?

MorphOS has not been released for PowerMac11,2 :-)

Quote:
Not with the 11,2, but with the 7,3 which is also in my signature.

[ Editiert durch ernsteiswuerfel 31.08.2017 - 01:35 ]

Quote:

And MorphOS runs with two Cpu? To get the best of G5.

Quote:

Well, no. Should be well known that MorphOS does use one core/CPU only.

But still one G5 Core/CPU should be faster than the a single core G4.
Question is how much.

Quote:

I've run tests using Cinebench 11.5 ( under MacOS 10.4.11 ) , and:

- PowerMac G4 - 1.25 GHz - CPU Score - 0.15 (2GB Ram DDR 333)
- PowerBook G4 - 1.67 GHz - CPU Score - 0.21 (1GB Ram DDR 333)
- Powermac G5 - 1.8 GHz (Single CPU/Single Core ) - CPU Score 0.31 (2 GB RAM DDR 333 )
- PowerMac G5 - 2.0 GHz ( Dual CPU ) - CPU Score 0.72 (4GB Ram DDR 400 - MacOS 10.5.8 )

I guess Piru's benchmarks still remain valid:

mplayer

lame

And if you want one for a 2,7 GHz G5 - let me know.

Quote:

Please do!

Quote:

And may not be, but it is a neat machine.
I'm rebuilding the watercooling for my 2.5 GHz Quad core right now.

As to the 1.67 GHz laptop you mentioned, its more than adequate for anything you'll do under MorphOS.
I've retained a 1.42 G4 PowerMac with a registered copy of MorphOS that I use regularly.

The G5 just does everything a little faster.

Quote:

Well, I've repeated the LAME test a few times, but it won't go any faster than 6 seconds. As for the Mplayer benchmark - I don't know what Prometheus trailer Piru was using. I've tested a few and the results vary from 19 to 62 seconds for video decode. Maybe (just maybe) it's the one that gives me 38.616 seconds, but that's a lottery.

Better run a demanding Javascript benchmark (like Kraken 1.1) on OWB and share your results. Mine is 108643.5ms.

> the 1.67 GHz laptop [is] more than adequate for anything you'll do under MorphOS.

I can't speak for him and his needs, but I'm not sure 1.67 GHz G4 is able to do realtime playback of each and every 720p video thrown at it. My 1.5 GHz G4 clearly wasn't, while my 2.3 GHz G5 definitely is.

> I don't know what Prometheus trailer Piru was using.

http://downloads.dvdloc8.com/trailers/divxdigest/prometheus-trailer.zip

Quote:

In that case my result is 38.581s.

Quote:

Generally, I don't use MorphOS to watch videos.
Even if the G5 can handle 720p video, the cpu power used to generate the video on a G5 is still excessive, particularly when compared to other operating systems.
So, whether on my G4 or my G5, I don't waste cpu power on this task.

If we ever get gpu assisted decoding, or something else to offload this task, maybe I consider it.

Quote:

Well. I just want some more speed.
But for 10% (just a value) more it's not worth buying a G5. Even though I've always wanted one.


Manfred

Quote:

Where you got that 10% from?

Piru's benchmarks, comparing 1,67 GHz PowerBook with 2,0 GHz PowerMac G5:

Mplayer:
G4 - 82,2 sec.
G5 - 49,1 sec.

That's 40% faster.

Lame:
G4 - 14 sec.
G5 - 8 sec.

That's 40% faster.

hi all

With my G5 2.7G, Kraben 1.1 (OWB javascript benchmark) give me : 115514.9ms


Mplayer (svn r37401) with video prometheus in RAM.... = 42.857s

[ Edité par beworld 31.08.2017 - 18:48 ]

Quote:

You'll definitely get more than 10%.
It isn't a direct scaling (due to cpu differences), but the clock speeds are a fair indication of the improvement.

Quote:

I just wanted to express that 10% is not enough for me buying new hardware.
40% ist quite something.


Manfred

Anecdotally I can say that booting MorphOS 3.9 on my powerbook g4 1.67ghz vs my powermac g5 2.3ghz feels remarkably faster on the g5. Launching applications, in particular games which rely on a decent GPU, feels orders of magnitude faster on the g5.

I have no time comparison or anything. This is just my unscientific experience. :)

Quote:

From my experience I can confirm that there are quite some 720p vids that the G4/1667 (Powerbook 5.6) is not capable to replay fluently w/o delays/skips. Overall the Powerbook 5.6 @ 1667 MHz does not perform better than the 1.5 GHz Mini, they are pretty much on par.

> It isn't a direct scaling (due to cpu differences), but the clock speeds are a fair indication
> of the improvement.

Yes, according to the Lame and MPlayer benchmarks quoted/reported in this thread, a G5 with 20% higher clock speed than a G4 has 40% better performance than the G4, or in other words: the G5 delivers 17% better performance per clock than the G4.

> 40% ist quite something.

Absolutely. Updating from 1.5 GHz G4 to 2.3 GHz G5, like I did, even amounts to 80% performance gain, which is quite a jump.