What's Better: Radeon 9600 Pro (OEM) or GeForce4 4600 "Titanium" (OEM)?

15 posts / 0 new
Last post
alk
alk's picture
Offline
Last seen: 3 years 6 months ago
Joined: Dec 20 2003 - 10:38
Posts: 369
What's Better: Radeon 9600 Pro (OEM) or GeForce4 4600 "Titanium" (OEM)?

The title pretty much sums it up.

I recently picked up an MDD G4 (dual 1.42, FW800) that came with the GeForce4 Ti. I've got a Radeon 9600 Pro (128 MB, 8x AGP pulled from a G5 and modified to work in a 4x AGP G4) in my dual 1 GHz QuickSilver. As the MDD is clearly the performance leader between the two, I want to put the better video card in it. However, reviews (all the ones I could find on Bare Feats) are generally unclear about which card is actually better. I understand that the GeForce card doesn't support Core Image, but that isn't a point of concern for me at the moment.

What will I be doing with it? Not much that either card can't do: A little Quake 3 Arena, some Warcaft III, Halo, and maybe even Return to Castle Wolfenstien.

Thoughts? Strong arguments one way or the other?

Peace,
Drew

Offline
Last seen: 11 months 3 weeks ago
Joined: Jan 19 2005 - 23:30
Posts: 700
Just my opinion, not based on

Just my opinion, not based on real stats, but i think the 9600 is more capable.

Hokusai's picture
Offline
Last seen: 3 months 2 hours ago
Joined: Dec 20 2003 - 10:38
Posts: 256
Hmm...

I don't personally know, but from what I've been able to gather they are about on the same level. This link might help best graphics cards table.

dankephoto's picture
Offline
Last seen: 1 year 5 months ago
Joined: Dec 20 2003 - 10:38
Posts: 1899
so test 'em!!

Yo dude, you have like both and you're like asking us?!!? Blum 3

Run something like one of the Q3A demos or the Halo perf. test and see what kind of frame rates you get. Then report back here.

For what it's worth, I'd always heard the Nvidia card is faster in classic MOS and the ATI card faster in X. I have only the Ti4600 myself, sadly.

Heh, but neither is exactly a screamer in such an old machine. Try anything recent and watch 'em choke.

dan k

madmax_2069's picture
Offline
Last seen: 14 years 2 weeks ago
Joined: Sep 24 2005 - 07:28
Posts: 664
yea that is what i would do a

yea that is what i would do as well, take the 9600 Pro in the MDD and run your games and run benchmarks, then do the same with the GF4 4600 Ti .

i could not find any place on the net that shows good comparison info on those video cards ( as benchmark and game FPS tests)

but i did find out a bit on wiki

the Geforce4 4600 Ti spec's
GeForce4 Ti 4600 Feb 2002
code name: NV25
fab (nm): 150
bus interface: AGP 4x
memory (max) 128
core clock: 300
memory clock: 650
config core: 2:4:8:8 ( Vertex shader : Pixel shader : Texture mapping unit : Render Output unit)
Fill rate: 2400
bandwith (max) GB/s: 10.4
memory bus type: DDR
memory bus width: 128
direct Secret 8.1
openGL: 1.4

Radeon 9600 Pro specs
Radeon 9600 Pro Apr 2003
code name: RV350
fab (nm): 130
bus interface: AGP 8x
Memory (max): 128, 256
core clock: 400
memory clock: 300
config core: 2:4:4:4 (Vertex shader : Pixel shader : Texture mapping unit : Render Output unit)
fill rate: 1600
bandwidth (Max) GB/s: 9.60
memory bus type: DDR
memory bus width: 128
directx: 9.0
openGL: 2.0

its hard to tell what card will perform better just from looking at the stats. what might be the difference is the cards memory clock speed, the Geforce's memory is about double as fast as the 9600's memory, but the 9600 also supports openGL 2.0 as to where the Geforce only supports openGL 1.4, the 9600 does have a 100mhz faster GPU then the Geforce's GPU but most of the times i have seen better performance with faster memory then with a faster GPU.

Jon
Jon's picture
Offline
Last seen: 13 years 6 months ago
Joined: Dec 20 2003 - 10:38
Posts: 2804
It look like it may come down

It look like it may come down to feature support. The 9600 may have some tweaks done to certain functions as it is newer and supports higher OpenGL and DX version, but the GF4 Ti nearly matches it in terms of speed and other hardware specs. The biggest difference might well be what particular software is to be used. Older software isn't going to have stuff tweaked in for the 9600, but it may come down to driver support on top of that anyway.

And what model of the GF4 Ti is in question here anyway? Wink

madmax_2069's picture
Offline
Last seen: 14 years 2 weeks ago
Joined: Sep 24 2005 - 07:28
Posts: 664
its a Geforce4 4600 Ti

its a Geforce4 4600 Ti

alk
alk's picture
Offline
Last seen: 3 years 6 months ago
Joined: Dec 20 2003 - 10:38
Posts: 369
:-P

Alright, Dan...

Here are the results from one set of timedemos in Halo:

With the GeForce4 4600 Titanium:


For further information, please visit the timedemo FAQ at: http://halo.bungie.net/site/halo/features/hpcperformancefaq.html
Date / Time: 12/17/07 9:57:14 (0ms)
1400MHz, 2048MB
Macintosh HD\Applications\Halo\Halo Frames=4700
Total Time=159.26s
Average frame rate=29.51fps
Below 5fps= 8% (time) 0% (frames) (13.589s spent in 11 frames)
Below 10fps= 9% (time) 0% (frames)
Below 15fps= 12% (time) 1% (frames)
Below 20fps= 27% (time) 10% (frames)
Below 25fps= 47% (time) 25% (frames)
Below 30fps= 56% (time) 34% (frames)
Below 40fps= 73% (time) 53% (frames)
Below 50fps= 88% (time) 76% (frames)
Below 60fps= 96% (time) 91% (frames)
###Sound Options###
Hardware Acceleration= No
Sound Quality= Low
Environmental Sound= No
Sound Variety= Medium
###Video Options###
Resolution= 1024 x 768
Refresh rate= 0 Hz
Framerate throttle= No Vsync
Specular= No
Shadows= No
Decals= No
Particles= Off
Texture Quality= Medium

With the Radeon 9600 Pro (OEM):

For further information, please visit the timedemo FAQ at: http://halo.bungie.net/site/halo/features/hpcperformancefaq.html
Date / Time: 12/17/07 10:18:30 (0ms)
1400MHz, 2048MB
Macintosh HD\Applications\Halo\Halo Frames=4700
Total Time=226.41s
Average frame rate=20.76fps
Below 5fps= 6% (time) 0% (frames) (15.794s spent in 12 frames)
Below 10fps= 16% (time) 4% (frames)
Below 15fps= 25% (time) 9% (frames)
Below 20fps= 45% (time) 26% (frames)
Below 25fps= 69% (time) 52% (frames)
Below 30fps= 87% (time) 76% (frames)
Below 40fps= 95% (time) 88% (frames)
Below 50fps= 97% (time) 94% (frames)
Below 60fps= 99% (time) 97% (frames)
###Sound Options###
Hardware Acceleration= No
Sound Quality= Low
Environmental Sound= No
Sound Variety= Medium
###Video Options###
Resolution= 1024 x 768
Refresh rate= 0 Hz
Framerate throttle= No Vsync
Specular= No
Shadows= No
Decals= No
Particles= Off
Texture Quality= Medium

The system specs are as follows:
Dual 1.42 GHz G4
2 GB of PC2700U-25330 DDR DIMMs
180 GB IBM Deskstar-180GXP (7200 RPM, 8 MB cache, ATA/133)

So it looks like the GeForce wins! And although the two cards look about even, the GeForce proved to be almost 1/3 faster. Given the other reviews I've seen that put the 9600 and the GeForce4 4600 almost neck-and-neck, I'm thinking this is merely a fluke. Or it could be that the 9600 I've got (the Pro OEM version from a G5) is actually less capable than the retail version. Apple has been known to offer lower-speced cards in their BTO options in the past. I haven't really done my homework here, but that would be my hunch.

If I find more time, I'll do some more bench marking and research and get back with those results.

Peace,
Drew

madmax_2069's picture
Offline
Last seen: 14 years 2 weeks ago
Joined: Sep 24 2005 - 07:28
Posts: 664
so having faster memory, was

so then having faster memory must have been the ball breaker. like you said Apple is known to hamper the video cards ( underclock them).

i would love to see what the core and memory speeds is set to on your 9600 pro vs the info in my post. there is a few app's that will tell you the memory and core speeds when you launch them.

I assume that the game had identical settings selected for both cards during tests ( in the pop up config window before the game runs)

Jon
Jon's picture
Offline
Last seen: 13 years 6 months ago
Joined: Dec 20 2003 - 10:38
Posts: 2804
Maybe I should be reading top

Maybe I should be reading topic titles closer... Sad Heh.

madmax_2069's picture
Offline
Last seen: 14 years 2 weeks ago
Joined: Sep 24 2005 - 07:28
Posts: 664
Re: Maybe I should be reading top

Maybe I should be reading topic titles closer... Sad Heh.

at first i almost did the same thing, but i did a double take.

dankephoto's picture
Offline
Last seen: 1 year 5 months ago
Joined: Dec 20 2003 - 10:38
Posts: 1899
Howza 'bout trying the G4 version of Q3A timedemo?

Q3A is an OpenGL app, is Halo as well? (I forget . . . ) As a datapoint, on my dual 800 QS/GfTi4600 running 10.3.x, in Q3A G4 with timedemo I got ~74fps (res=1280x, everything maxed.)

Also be curious to see what sort of speed hit you get when you bump up the settings to a higher level on Halo. Some cards take to higher settings better than others.

Still this is a pretty good indicator, 29fps vs. 20 fps is a big difference. This is under X.x I assume (your version of MDD is X-only?), so I'm still curious how much difference one might see under classic MOS.

dan k

madmax_2069's picture
Offline
Last seen: 14 years 2 weeks ago
Joined: Sep 24 2005 - 07:28
Posts: 664
yea i think halo is the harde

yea i think halo is the hardest to run that he has in his list, i play halo on my DA 533 with a Geforce2 MX (everything is set to minimum) only then will it play well. some servers in halo use modded maps which can bog the game to a unplayable rate, but if the map is stock then it plays fine

soon (when its incomtax time) i will be getting a dual 7455 1.4ghz upgrade and a 6800 GT.

then i bet halo will play allot better.

also you should check out UT 2k4 ( it ran good on my system, with medium settings) it should be nice and smooth on your MDD.

alk
alk's picture
Offline
Last seen: 3 years 6 months ago
Joined: Dec 20 2003 - 10:38
Posts: 369
Another Data Point

I got a chance to play around with it some more (yep, the wife isn't home ;)).

I played back a timedemo in Quake 3. Keep in mind that other than the video card, the settings are "max" and identical between the two. The results are startling. I set up Quake 3 to play at full screen native res on my SGI 1600sw at 1600x1024. And the results?
- Radeon 9600 Pro w/ 128 MB VRAM: 86.7 fps
- GeForce4 4600 Ti w/ 128 MB VRAM: 130.5 fps

Holy smokes! The Gf4Ti spanks the Radeon!

Unfortunately, during this benchmarking, I've discovered that the fan on the GeForce has seized up (the previous owner let this thing collect dust until it expired from heat exhaustion, which is fine by me because I got the "dead" G4 for free). I had to replace the PSU to get it to boot, and now apparently, I have to replace the GPU heatsink or at least the fan. As an aside, I tried to run Halo at 1600x1024, but it kept locking up after a few seconds while I still had the Gf4Ti installed. Now I know why...

Anyway, it looks like the GeForce card is where it's at.

Now, does anybody have any tips or recommendations for replacement GPU coolers?

Peace,
Drew

madmax_2069's picture
Offline
Last seen: 14 years 2 weeks ago
Joined: Sep 24 2005 - 07:28
Posts: 664
it depends on how much you wa

it depends on how much you want to spend on that video card,

this is a good replacement
http://www.xoxide.com/vantec-iceberq-5.html

but that is a bit expensive. and might not fit some models of the GF4 4600 Ti

here is one that might work, depends if has that bolt pattern.
http://www.directron.com/ec4510m12sb.html

i have seen references that the HS is glued on to the GPU, if it is you can only replace the fan (which should be held onto the HS with screws) you can find replacement fans that you can hook up to a molex from the PSU (which would be allot cheaper). i only hope you haven't caused any damage to the card with running it without proper cooling and overheating it.

or you can just jurry rig a fan on the HS (a 45mm fan somehow hooked to the HS to provide airflow) cause really that is all you need.

or you can get lucky and score a faster GF for a cheap price on ebay ( and flash it to Mac). i seen a 6800 GT sell for $25 shipped. i do think the MDD has issues with some cards cause of the case and mobo design.

Log in or register to post comments