X1800XT benching with Dothan platform

[Astennu]

read the thread, macci suggests that he feels like some sort of resistor mod may be neccesary to get the higher vcore stable as max bench stable seems to be lower then expected.

I get the same feeling with my XL card. It won't go over 1.2v without artifacting :\

[macci]

Here is an interesting ORB comparison:


Its kingpin's X850 score (currently the highest ATI score in the ORB) w/ 3572MHz Dothan on P4C800, now if we take system scores from kingpin's run and add the X1800XT Nature score we are already at almost 50k (Drago's would be quite a bit higher too at those speeds so at 3572 it would probably be already over 50k). And this is 'only' 3572MHz - we have seen 3.9GHz Dothans in the ORB

[zakelwe]

Yes that is interesting, Drag high would get a good boost as well like you say, I think more like 51k rather than 50k even with 3500-3700

Do you fancy having a go at it macci ? I bet k|ngp|n will !

Regards

Andy

[Astennu]

OMG That nature score. Is that so low because u have a AMD ? and not a Dothan ? jees this is a serious difrence

[G H Z]

Do you fancy having a go at it macci?

I think he's already planning to

I know I'd love to see it. Wow, I forgot where Kingpins Dothan was on that run.

[JacobHansen]

I think he's already planning to

I know I'd love to see it. Wow, I forgot where Kingpins Dothan was on that run.

Now we can really see the power in the dothan compared to the fx.. damn, id give away one of my nuts for a 3.9 dothan..

[G H Z]

id give away one of my nuts for a 3.9 dothan..

I'd give away one of mine too, just not one thats attached

[sky]

has this been the 256mb or the 512mb x1800xt?

[macci]

512MB version.

And here is yet another interesting ORB pic. Shows the difference between i875 and i915

i875 (P4C800) running at 245FSB, i915 (P4GPL-X) running at 262FSB, both w/ fastest mem timings and similar windows tweaks (only the driver is different).
915 chipset needs some 200MHz extra CPU speed and around 20MHz higher FSB speed to match i875 as far as 3dmark2001se system scores go..

I'll install winxp on IDE drive and see if I can get the FSB over 270 w/ this i915 board.

[G H Z]

If we take system scores from kingpin's run and add the X1800XT Nature score we are already at almost 50k (Drago's would be quite a bit higher too at those speeds so at 3572 it would probably be already over 50k). And this is 'only' 3572MHz - we have seen 3.9GHz Dothans in the ORB

Only thing I would add is this gen is not quite as efficient with system power.

Here's my FX running X850 & X1800.



Still, I think 50K is doable

[brandinb]

512MB version.

And here is yet another interesting ORB pic. Shows the difference between i875 and i915

i875 (P4C800) running at 245FSB, i915 (P4GPL-X) running at 262FSB, both w/ fastest mem timings and similar windows tweaks (only the driver is different).
915 chipset needs some 200MHz extra CPU speed and around 20MHz higher FSB speed to match i875 as far as 3dmark2001se system scores go..

I'll install winxp on IDE drive and see if I can get the FSB over 270 w/ this i915 board.

well macci dont you think that the driver effeciency of the x1800 can be different then the x850???
like how a nvidia card gets low system score are you sure the x1800 is as good as x850 as far as system speed goes??? 915 should be faster i would think but im not sure
________
house wives Cam

[macci]

i915 is slower than i875 and that is the case in every CPU/memory dependant benchmark. Just like i815 was slower than 440BX

GHZ, Yet another interesting pic
thats quite a drop in car low.

Yes there is a difference in clock to clock performance between x1800 and x850 in mark2001 scores. Wheter its just the driver (6497 being faster than the new ones) or the x1800 taking more system resources remains to be seen. I have x850 here and can test em side by side w/ the same new driver.

[Astennu]

i915 is slower than i875 and that is the case in every CPU/memory dependant benchmark. Just like i815 was slower than 440BX

GHZ, Yet another interesting pic
thats quite a drop in car low.

Yes there is a difference in clock to clock performance between x1800 and x850 in mark2001 scores. Wheter its just the driver (6497 being faster than the new ones) or the x1800 taking more system resources remains to be seen. I have x850 here and can test em side by side w/ the same new driver.

Its a bit stange. Because nature is not a real system limmeted test. Its all about the GFX card. In 01 its not totally GFX indipendent. But that kind op drop is strange. The X800 was also faster in 01 nature then X850 ? I have heared that a couple of times. May be its a driver thing. Drivers are far from optimal ATM.

[macci]

Nature??
The huge NAture FPS difference in all the pics is because its X1800 vs. X850/X800 comparison and obviously the X1800 is miles ahead in that one

[crotale]

i915 is slower than i875 and that is the case in every CPU/memory dependant benchmark. Just like i815 was slower than 440BX

GHZ, Yet another interesting pic
thats quite a drop in car low.

Yes there is a difference in clock to clock performance between x1800 and x850 in mark2001 scores. Wheter its just the driver (6497 being faster than the new ones) or the x1800 taking more system resources remains to be seen. I have x850 here and can test em side by side w/ the same new driver.

Intel don't seem to get their iX15 chipsets performing well

That comparison between X850 and X1800 with the same driver will be very interesting.

[macci]

Anyone know how far these i915 ASUS mobos have been taken with 2-2-2-5 timings?

[hipro5]

Anyone know how far these i915 ASUS mobos have been taken with 2-2-2-5 timings?

NICE work.....

IF I remember well, I've seen about 275MHz 1:1 benchable....

[macci]

ok that sounds about right then, 275 is where I'm at too (SPi 1M)

[Astennu]

Nature??
The huge NAture FPS difference in all the pics is because its X1800 vs. X850/X800 comparison and obviously the X1800 is miles ahead in that one

ARf my bad looked wrong :S. Its early in the morning here

[hipro5]

Lucking out of PAT, 915 chipset is OK........

ALSO for the benchers of the ASUS P4GD1 and P4GPL-X mobos with the "lock" PCI-E/PCI frequency, I have to tell that when someone is ABOVE 266MHz fsb the PCI-E/PCI are 100/33 MHz locked........

To be a little "more" correct on this, when you pick: 100 or 133 or 166 or 200 or 267 or 333 or 400 MHz fsb, the PCI-E/PCI START from 100/33 MHz........Depents of the bits that ASUS has given into bios for controlling the PLL chip....OR 270MHz and up for 101/33 PCI-E/PCI.....

[mrlobber]

To be a little "more" correct on this, when you pick: 100 or 133 or 166 or 200 or 267 or 333 or 400 MHz fsb, the PCI-E/PCI START from 100/33 MHz........Depents of the bits that ASUS has given into bios for controlling the PLL chip....OR 270MHz and up for 101/33 PCI-E/PCI.....

Hipro, did I understand you right, if I set 200fsb/400 mhz DDR in bios, and leave 100/33 PCI-E/PCI frequencies, they will remain locked at those 100/33 values if I later play with Clockgen in Windows?

So far, it has been for me that I cannot get over 250 fsb (immediate lockup) unless I up the PCI-E frequency to 120 Mhz in bios before (and onboard SATA craps out at ~117 PCI-E, board is unmodded though). With 120 PCI-E, the board does 265+ fsb.

[hipro5]

No......When you boot @ 267MHz, then the PCI-E/PCI are locked at 100/33MHz....

EDIT: Locked PCI-E/PCI are when you boot @ 100, 133,166,200,267,333,400MHz fsb....

[mrlobber]

Edit: nevermind, just read your edited post

[hipro5]

Maybe IF you boot @ 200MHz and go into windows lower the multi via eist and then jump from 200MHz to 267+MHz via clockgen, the PCI-E/PCI are locked again @ 100MHz/33MHz.....ONLY thing is that your rams must handle over 267MHz.....

[kaze]

Just insane great work neighbour