Radeon 9600 PRO Overclocking

[Hellbinder]

Thanks Ichy..

Btw Dave. What other Reviewer? have they posted anything online i can read about?

 

[Morris Ital]

My guess is this card will be over the top for DX7 games and no good for Doom III when it comes out and therfore it is pointless overclocking it.

How can you say that? I think this will be a very capable board for Doom III, it might even do better than a R9500pro due to the stencil buffer optimizations. The leaked Doom III alpha however is utterly useless for performance estimations of the game.

Doom III is still better than 3dmark03 though as it is not as intensive and will only get better, while 3dmark03 gives 20+fsp on Dave's tests ... I bet Doom III alpha gives better than that already. Like I said, lets see how a overclocked 9600 does on Doom III alpha and current DX7 games, If it gives 30+ fps on Doom III alpha then caveat emptor. Better get a 9800 Pro or nv35

In this scenario an overclocked 9600 becomes an oddity, good for benchmarking.

It reminds me of my GF4 MX440, and this is what I meant to say to Dave, the GF4 MX440 was a great overclocker, because it was more simplistic than the Ti series it overclocked out of this world ..mine does 405 / 680 ..check that against the default. Buit it still does 10 000 against 15000 for the ti series at 320/650 in 3dmark 2001.


You can polish a turd, but it's still a turd.

 

[nelg]

I think some people might owe TSMC an apology.

 

[mczak]

Ok, since you've found the limit. . . let's try M-E-M-O-R-Y!

and what about that volt-mod

 

[gokickrocks]

speaking of vmods...the vcore and vmem regulators of the 9600 pro look the same as the 9700 pro, just located in different areas *nudge nudge*

 

[Tagrineth]

Doom III is still better than 3dmark03 though as it is not as intensive and will only get better, while 3dmark03 gives 20+fsp on Dave's tests ... I bet Doom III alpha gives better than that already. Like I said, lets see how a overclocked 9600 does on Doom III alpha and current DX7 games, If it gives 30+ fps on Doom III alpha then caveat emptor. Better get a 9800 Pro or nv35

The DOOM3 alpha is also hardly indicative of what final performance will be like.

What do you think Carmack is doing right now? Well, aside from fixing the otherwise-horrendous NV3x performance of course

 

[Ante P]

Doom III is still better than 3dmark03 though as it is not as intensive and will only get better, while 3dmark03 gives 20+fsp on Dave's tests ... I bet Doom III alpha gives better than that already. Like I said, lets see how a overclocked 9600 does on Doom III alpha and current DX7 games, If it gives 30+ fps on Doom III alpha then caveat emptor. Better get a 9800 Pro or nv35

benchmarking an alpha that doesn't even run the path that the final game will utilize seems pretty damn pointless from just about any viewpoint I can imagine
once the game comes out it will most probably perform a hell of a lot better than this alpha does, especially on ATis cards

 

[David G.]

I think some people might owe TSMC an apology.

Here's something I've red recently @ rage3d forums :

Originally posted by East17
Told you !

I doesn't matter if TSMC makes the chip as it's not manufacturer's blame if the VPU doesn't scales high .

Here's what I've posted more than 2 weeks ago :

Originally posted by East17

Originally posted by demonic
That would be awesome if ATi switched to Intel for producing the R400.. TSMC really should not have been used, after the fiasco with NV3X.

As some of you know , the fact that the R300 design was so succesfully on 0.15 micron technology has also to do with the close relationship between Intel and ATi ... some know what I'm talkiing about ...

Anyway , the important fact is how you design your product and it's generaly your fault if it's not working properly not the manufacturer's .

Look @ the RV350 ... it's doing 400 Mhz in mobile version too . Or let's put it in another way : if the R300 was doing 400 Mhz and even more easily ... do you really think that the RV350 can only do 400 Mhz ? I think that RV350 can even go beyond 500 Mhz !

The design is wat counts ... take the Thoroughberd A and B issue for another example .

As for nVIDIA ... I would be happy if they will get their act together as that would mean smaller prices for the ATi cards and the same 6 month cycle and not a 9 months one .

Have a nice day everybody .

Link

Makes me look very sma*t errrr .... I mean intuistic , doesn't it ?

Hehe ... just joking ...

There's a point .

This has nothing to do with TSMC .

 

[Mark]

Considering this is ATIs .13u "test" platform, great overclocking like this bodes very well for the .13u version of the R350 (if such a beast exists)... a 500MHz or so R350 core matched with 500MHz DDR2 memory would be a decent card I think

 

[Ante P]

Considering this is ATIs .13u "test" platform, great overclocking like this bodes very well for the .13u version of the R350 (if such a beast exists)... a 500MHz or so R350 core matched with 500MHz DDR2 memory would be a decent card I think

bah, I want a 1 GHz core to match my 1 Ghz DDR-II (with a 256 bit bus mind you)

in any case, if the 9800 Pro 256 MB is truly clocked at 400/920 while still being on 0.15
only asking for 500/1000 when stepping to 0.13 seems pretty pessimistic IMHO

 

[Dave Baumann]

Well, looks like 540/340 might be the upper limit. I've been handed some test suggestions for stress testing so it may scale back from that a little more.

http://service.futuremark.com/compare?2k3=612457

 

[Ante P]

Well, looks like 540/340 might be the upper limit. I've been handed some test suggestions for stress testing so it may scale back from that a little more.

340 for the memory is ok I guess
so 540/680 it is

could you do some AA/AF tests with these settings?

 

[David G.]

Hey hey ...

Over 1050 views per hour on a single theread ?!?!

 

[David G.]

WTF is this :

@537/300 gets 3816 3DMarks in 3DMark03

@537/340 gets 3860 3DMarks in 3DMArk03 ?!!??!

Only 44 3DMarks with almost 40% overclock ?

Why ??? :?

 

[Freak'n Big Panda]

@537/300 gets 3816 3DMarks in 3DMark03

@537/340 gets 3860 3DMarks in 3DMArk03 ?!!??!

Only 44 3DMarks with almost 40% overclock ?

Yeah I don't get that either Dave wanna explain? Also 3860 sounds a little low to me considering I get 42XX with a 9500 pro @ 385/317.

Heres a link: http://service.futuremark.com/compare?2k3=8998

But still 537 is just amazing considering that size of that fan....

 

[Ante P]

WTF is this :

@537/300 gets 3816 3DMarks in 3DMark03

@537/340 gets 3860 3DMarks in 3DMArk03 ?!!??!

Only 44 3DMarks with almost 40% overclock ?

Why ??? :?

he only ran with 437/430 right?

 

[Ichneumon]

Core clock makes much more difference than memory clocks on most cards in 3DMark03.

Its shader limited for the most part.

 

[YfOrU]

pretty easy your are not able 2 set mem clock
had the same problem on my r9000mobile @ first
you need a new version of your oc tool with r9600
support.

 

[David G.]

he only ran with 437/430 right?

It says 537 in the Futuremark page towards he linked ...

 

[David G.]

Its shader limited for the most part.

Please , give some details on that .