[Artikbot - Radeon 9200 SE @ 375/220MHz - 7190 marks 3DMark2001 SE]

Vcore modded via normal opamp mod, vmem modded with external power supply.

[Artikbot - Radeon X1600 Pro 512mb DDR2 @ 681/427.5MHz - 27858 marks 3DMark2001 SE]

I know, I know. During this holidays I'll re-rig the GPU to my system and run the benchmark again.

[Split some graphics cards depending on bus]

Thanks a lot!

 

By the way, excellent results you got there on the AGP cards! I look at them with a telescope of how high they are

[Split some graphics cards depending on bus]

Good evening,

 

 

I've been working lately with some old AGP graphics cards, to see how well do I fare in a completely CPU bottleneck free platform, and where the processor clocks aren't as important as in other platforms.

 

Well, to my surprise, when submitting scores some of the cards I work with don't discrimine between 64 or 128bit bus widths.

 

On newer PCI Express cards this doesn't pose that much of a threat, but on AGP that memory bus width can (and does) affect performance quite a lot.

 

For instance, take this FX5200 submission I made:

 

http://www.hwbot.org/submission/2264760_artikbot_3dmark2001_se_geforce_fx_5200_7228_marks

 

I ran the card to the clocks that gave me the best performance (305/500 - 250 DDR).

 

Now look at, for instance, this submission here:

 

http://www.hwbot.org/submission/886061_student_3dmark2001_se_geforce_fx_5200_9711_marks

 

He gets more than 2000 points more, while running at remarkably lower clocks.

 

 

The only difference between our cards is the bus. He has a 128bit-wide bus while I run a 64-bit narrow bus.

 

What I mean is, there is no contest for 64bit card users if there are 128 bit versions, because with little effort, a 128 bit card can easily score a 30% more points than a 64bit one overclocked to hell and beyond.

 

 

 

Thanks for your understanding!

[Artikbot - Radeon 7000 DDR ~ Radeon VE DDR @ 300/216MHz - 2134 marks 3DMark2001 SE]

On the bench picture I'm not using the Radeon 7000, although in the second picture it can be clearly seen that I use the same watercooling in the 7000 that in the 1st pic (I was using a FX6200 there)

[The official Temessi's 3DMark06 global challenge - Jan 22, 2012 until Jan 29, 2012 thread.]

I also want to know this, because if there's no GPU limitation I won't enter. Would be cool if we'd be limited to IGP only.

[jonny391@OCQQ - Phenom II X6 1055T @ 4965.8MHz - 4965.84 mhz CPU-Z]

Jizz in ma pants. Air cooled on those volts, did the chip blow right after? xDD

Nah seriously, amazing chip you have there, and amazing job you did

[bennyluk - Phenom II X6 1055T @ 4830.1MHz - 4830.09 mhz CPU-Z]

Amazing mate, how did you achieve this on air cooling ? Is that a 95W or a 125W chip?

 

Great job!