OpenMark results on PowerPC Macs

[zen.state]

Originally Posted by SuperJudge

Does the GMA 950 even support Core Image?

Oh yeah. 100%. It screams at anything 2D. It supports every GPU feature of the OS. Snow Leopard's GUI screams and even Lion performs pretty well. Keep in mind this is a 3D test.

I only see it crippled when I try to play games which was twice and then I stopped. My G4 games far better performance-wise and I do all my gaming on it. The GMA has shared memory and no hardware rendering vs. the 6200 that has hw rendering and it's own dedicated 256MB DDR2.

[Nameci]

Originally Posted by SuperJudge

Proof that buses are not necessarily the limiting factor:

PowerMac 11,2 (2.3GHz DC G5, PCIe), 4GB RAM, GeForce 6600 w/ 256MB VRAM

Lol, my PB G4 15" DLSD scored higher than the Nvidia 6600?

If that is the score for 6600, I want that QFX 4500 soon...

[zen.state]

Originally Posted by Nameci

Lol, my PB G4 15" DLSD scored higher than the Nvidia 6600?

Even more interesting is the 12" 1.5GHz PB with a Geforce 5200 that only scored 1411.

[SuperJudge]

Originally Posted by zen.state

Oh yeah. 100%. It screams at anything 2D. It supports every GPU feature of the OS. Snow Leopard's GUI screams and even Lion performs pretty well. Keep in mind this is a 3D test.

Fair enough. I forget that OS X is sanely 2D on its desktop environment sometimes.

Originally Posted by Nameci

Lol, my PB G4 15" DLSD scored higher than the Nvidia 6600?

Yeah, man. The 6600 is a dog of a card for 3D. The only thing to really commend it is the fact that it can drive a single dual-link display which was no mean feat when it came out. Definitely not a gaming card, though.

[Nameci]

Then my ATI Radeon 9700 Mac Edition on my dualie 1.42 PM G4 would score higher than the 6600 I bet.

[SuperJudge]

Originally Posted by Nameci

Then my ATI Radeon 9700 Mac Edition on my dualie 1.42 PM G4 would score higher than the 6600 I bet.

Almost certainly.

[zen.state]

Originally Posted by SuperJudge

Fair enough. I forget that OS X is sanely 2D on its desktop environment sometimes.

Yes and 2D performance in my experience is a little more influenced by CPU than GPU. This is why my G4 with 6200 has better GUI scores in Xbench than many G4's with better GPU's and faster AGP. Because the G4 7448 is a screamer.

[Nameci]

Plugged or unplugged, it's the same score for ATI Mobility Radeon 9700 128mb on a PowerBook G4 15" 1.67GHz DLSD.

1440x960x32
Score: 5644

1280x800x32
Score: 5644

[zen.state]

Originally Posted by Nameci

Plugged or unplugged same score for ATI Mobility Radeon 9700 128mb on a PowerBook G4 15" 1.67GHz DLSD.

Score: 5644

Considering your 9700 is a mobility version vs. a 6600 desktop card your score being even just a little higher like it is is pretty respectable performance. It's the fastest Apple PowerPC laptop ever.

I found it very surprising my 6200 in AGP 2x equaled SJ's 6600 in AGP 8x. Or are the dual 2.3 G5's PCIe graphics?

[Nameci]

If SJ has the 256MB Nvidia GF 6600, then it's either the DC 2.3GHz or the Q2.5GHz. It's a PCIe card.

[zen.state]

Originally Posted by Nameci

If SJ has the 256MB Nvidia GF 6600, then it's either the DC 2.3GHz or the Q2.5GHz. It's a PCIe card.

I noticed he listed dual 2.3 but missed he did list it was PCIe. Thats amazing that my slower 6200 in a G4 on AGP equals that.

I have this same one but with a black PCB as the page mentions: "This card's PCB can be blue or black".

[SuperJudge]

Originally Posted by Nameci

If SJ has the 256MB Nvidia GF 6600, then it's either the DC 2.3GHz or the Q2.5GHz. It's a PCIe card.

That is 100% correct.

It's amazing to me, too, that an AGP 2x card is equal in performance to a card on PCIe. Go figure, huh?

[Nameci]

It'probably a good justification to upgrade the GPU to a QFX 4500...

[zen.state]

Originally Posted by Nameci

It'probably a good justification to upgrade the GPU to a QFX 4500...

Did you see this post on the first page? Very good score!

Originally Posted by MrJonnyJones

PowerMac G5 Quad. I know, I need bigger screens!

[Nameci]

Yup, I have seen it, and I am salivating...

[zen.state]

Originally Posted by Nameci

Yup, I have seen it, and I am salivating...

So far thats the top score of the thread. I can't imagine any other PowerPC being able to beat that score since I imagine thats the fastest GPU you can get for the G5's.

[Nameci]

How about the GeForce 7800 GTX 512mb. Have anyone here have this card? Please post and we will see how it compares with the quadro FX 4500.

[ThunderSnake]

Originally Posted by Nova77

My G4 dual 1.42 MDD, 2 GB RAM (mac os 10.5.8)

Holy Cow!

Mine (MDD dual 1.42 @ 1.5 w/ Radeon 9800 & Leopard 10.5.8):



Wow! Is the GF 7800 GS really THAT much better than a Radeon 9800?

[Nameci]

I think so, there's the proof...

[Nova77]

Originally Posted by ThunderSnake

Holy Cow!

Mine (MDD dual 1.42 @ 1.5 w/ Radeon 9800 & Leopard 10.5.8):

Image

Wow! Is the GF 7800 GS really THAT much better than a Radeon 9800?

Yes it is
That bench just confirms others I've seen before buying the card.
Simply put, the 7800 GS is the 2nd best AGP card ever made. (the first one is the 7800 GT 512 MB).

"Holy Cow"... Can I borrow this one? funny as hell

[zen.state]

Originally Posted by SuperJudge

That is 100% correct.

It's amazing to me, too, that an AGP 2x card is equal in performance to a card on PCIe. Go figure, huh?

In Nova's world this isn't possible though.

How can AGP 4x over 2x possibly make a difference when even PCIe with a faster 6600 matches my score exactly. The 6200 is actually a 6600 with half the pipelines disabled.

[Nova77]

Originally Posted by zen.state

In Nova's world this isn't possible though.

How can AGP 4x over 2x possibly make a difference when even PCIe with a faster 6600 matches my score exactly. The 6200 is actually a 6600 with half the pipelines disabled.

1. Go see the techs that designed AGP 2X while searching for an improvement over PCI
Go see the techs that designed AGP 4X while searching for an improvement over AGP 2X
Go see the techs that designed AGP 8X while searching for an improvement over AGP 4X
Go see the techs that designed PCIe while searching for an improvement over AGP 8X

2. And just tell them all their efforts were useless because a card in AGP 2x = a card in PCIe in some benchmark.

3. See what happens.

[SuperJudge]

Originally Posted by Nova77

1. Go see the techs that designed AGP 2X while searching for an improvement over PCI
Go see the techs that designed AGP 4X while searching for an improvement over AGP 2X
Go see the techs that designed AGP 8X while searching for an improvement over AGP 4X
Go see the techs that designed PCIe while searching for an improvement over AGP 8X

2. And just tell them all their efforts were useless because a card in AGP 2x = a card in PCIe in some benchmark.

3. See what happens.

You have drawn the wrong conclusion. I would say that the take home message is that bus speed is not necessarily a limiting factor when discussing video card choices for a PowerPC Mac. The most important thing is to pick a card that doesn't suck.

[zen.state]

Originally Posted by SuperJudge

You have drawn the wrong conclusion. I would say that the take home message is that bus speed is not necessarily a limiting factor when discussing video card choices for a PowerPC Mac. The most important thing is to pick a card that doesn't suck.

Exactly. My argument from the beginning was that on PowerPC Macs the speed of that graphics slot or system bus are not anywhere near as limiting as a slow GPU and CPU. The overall PowerPC architecture is far more efficient even down to how it accesses the GPU.

In the PowerPC world specs often don't equate to real world performance.

[Cox Orange]

Originally Posted by Nova77

... Check out www.videocardbenchmark.net/ ...

interesting you link to passmark. I used this charts some time ago to get a vage feeling of how components compare to each other, but then there are cases where I wonder, if it says anything. What does it mean, when the highest score is (just example) 1000, and to others are 256 and 260. I mean you can't see from the scores, if it is just a small difference or if they are close as the number indicate. (though I got to say, last time I looked there, I was comparing HDDs and it did not satisfy me.)
Also at Passmark there are so many different PC-setups that all come together, not as easy as comparing allmost similar Macs, with just one part different.

So that this is not only Spam, I will post one of my Macs later.

PS: oh, and I forgot, maybe this is interesting to someone http://www.jcsenterprises.com/Japama...OpenGL%20.html and http://www.jcsenterprises.com/Japama...5E1780B95.html (last link lists cards which were tested in one same Mac and ranks them).