Powercolor X800XT PE Adventures

eva2000 · 14th November 2004, 1:42 AM

Thanks to Coolmodz.com.au OCAU sponsor special I managed to pick up a rare Powercolor X800XT PE AGP card

It's stock speed is 520/560 according ATI Tool and idle and load GPU temps from ATI Tool = 36-39C idle and 64-68C load.

Front side has Samsung 419 GC16 chips, Backside has Samsung 422 GC16 chips heh

I used initially Omega 8.07 drivers then Cat 4.11 official. The Omega still was ~100pts faster in 3dmark2003 than Cat 4.11 for my card and faster in 3dmark2001 too. But here's a full set of 3dmarks using Cat 4.11 drivers at stock cpu speeds with my gaming OS image.

AMD64 3700+
DFI NF3 250Gb 05/10/04 4v bios
2x 512MB KHX3200/512 BH-5 @ 200mhz 2-2-2-6 1T
120GB WD 1200JB
LiteOn LDW-411s burner
580W Topower Adjustable PSU

Win2k SP4 fully loaded gaming OS image

Powercolor X800XT PE connected with OCZ VGA Enhanced power lead/molex

3dmark2001 = 29,147 vs Omega 8.07 score of 29,243
3dmark2003 = 14,033
3dmark05 = 6,704

[ r2 ] · 14th November 2004, 1:59 AM

Great results Eva, better performance over your Asus 6800Ultra that you owned.

Be interesting to see results with your cpu being clocked.

600/606 is the max you achieved?

Edit: Have you tried Cat 4.12 Beta?

eva2000 · 14th November 2004, 2:03 AM

Well I didn't have an AMD64 3700+ cpu when I had my Asus 6800U so results can't be totally comparable. 606mhz on mem is about it's max without artifacts. As to GPU core haven't gone much higher than 600mhz for now plus the past 2 days in Brissy have been bloody hot 32-34C!

But the OCZ VGA Enhanced power lead did help some http://forums.overclockers.com.au/sh...d.php?t=326071 you read and decide if it's worth it hehe

Cat 4.12betas don't work with Win2k, only WinXP. The bugs are WinXP specific and doesn't affect Win2K OS users so Cat 4.11 is what Win2K users will have to use

[ r2 ] · 14th November 2004, 2:09 AM

Quote:

Originally Posted by eva2000

Well I didn't have an AMD64 3700+ cpu when I had my Asus 6800U so results can't be totally comparable. 606mhz on mem is about it's max without artifacts. As to GPU core haven't gone much higher than 600mhz for now plus the past 2 days in Brissy have been bloody hot 32-34C!

But the OCZ VGA Enhanced power lead did help some http://forums.overclockers.com.au/sh...d.php?t=326071 you read and decide if it's worth it hehe

Cat 4.12betas don't work with Win2k, only WinXP. The bugs are WinXP specific and doesn't affect Win2K OS users so Cat 4.11 is what Win2K users will have to use

Fair enough, yeah i had a look at that thread before and ive seen some reports in other forums that it does boost a small amount of performance. I can see how there might be slight performance boost for VGAs but for HDDs?
Didnt know you were running Win2k.
Well still top results, so when are the benchies coming with cpu clocked?

eva2000 · 14th November 2004, 5:34 AM

Well here's some results with cpu oc'd

Room temps dropped 5C, so did GPU idle temps dropped from lowest 36C to 34C!

AMD64 3700+ @ 11x 246HTT = 2706mhz
2x 512MB KHX3200/512 BH-5 @ 246mhz 2-2-2-6 1T

Powercolor X800XT PE
@ 520/560 = 13,336
@ 602/605 = 14,563

3Dmark05
@ 602/605 = 6,766

3Dmark2001
@ 602/605 = 32,470

Aquamark3
@ 602/605 = 82,000

Okay screenies

34C idle

eva2000 · 14th November 2004, 5:45 PM

Okay some FarCry v1.3 patched benchmarks

AMD64 3700+ 12x 200 = 2411mhz stock
DFI NF3 250Gb 05/10/04 4v bios
2x 512MB KHX3200/512 BH-5 @ 200mhz 2-2-2-6 1T
120GB WD 1200JB
LiteOn LDW-411s burner
580W Topower Adjustable PSU

Win2k SP4 fully loaded gaming OS image

Powercolor X800XT PE connected with OCZ VGA Enhanced power lead/molex

Using HardwareOC Farcry Bench v1.4 all benchmarks done with 4xAA and 4xAF

@520/560
HardwareOC Steam Demo results
HardwareOC Volcano Demo results
TechReport Regulator Demo results

@590/600
HardwareOC Steam Demo results
HardwareOC Volcano Demo results
TechReport Regulator Demo results

jamesrhadley · 14th November 2004, 5:56 PM

Just a quick one why are you running 2 2 2 6 I though 10-12 was better for amd 64's? Or was that only tru for xp's?

eva2000 · 14th November 2004, 5:59 PM

Quote:

Originally Posted by jamesrhadley

Just a quick one why are you running 2 2 2 6 I though 10-12 was better for amd 64's? Or was that only tru for xp's?

tRAS of 10-12 for AMD64 only gives more sandra bandwidth, tRAS of 6 still gives real gains in gaming etc

groovetek · 14th November 2004, 6:16 PM

hey eva, do you bench at whatever feasible clocks for a particular benchmark suite, or do you use utilities like ATITOOL to REALLY find the maximum clock? It seems you're always using difference speeds for different benchmarks... that must mean the ones run at higher speeds are ones that can pass relatively reliably , but not necessarily 100% stable?

edit: does this also apply to the rest of your system? I see that you memtest/prime and all that, but when it comes to ripping out big numbers, do you compromise stability at all?

malonem · 14th November 2004, 6:25 PM

eva how the heck u get 600mhz core on stock... even with my peltier @ -10c I could only get it to 640mhz (vmod too)

eva2000 · 14th November 2004, 7:50 PM

Yeah partially reason i'm using different speeds is the card is less than 2 days old, still figuring it out for each app

As to ripping out the big numbers, they may not be 100% stable but for me they need to at least be superpi 16/32M, pifast, and memtest clean ... usually the cpu is the component that ain't 100% stable for me.

ATI Tool no way accurate i.e. find max core differs from time to time... as with all ATI cards room/case ambient temps are are a factor - on a cool night you can probably get an extra 5-10mhz off the core compared to the hottest most humid day you can experience

malonem... lucky i guess but why complain 640mhz on core is nothing to sneeze at

TazBazz · 15th November 2004, 12:10 AM

eva are you still using the stock cooler on the XT?

eva2000 · 15th November 2004, 12:25 AM

Quote:

Originally Posted by TazBazz

eva are you still using the stock cooler on the XT?

yup indeed I am

wastegate · 15th November 2004, 6:10 PM

Interesting Eva, wary interesting.

I have the same card but it came in a different box.

What other cooling did you do to the card?

eva2000 · 15th November 2004, 6:16 PM

ah different packaging... the powercolor web site shows the packaging I have so either yours is newer or older packaging hehe

only things i have done now are

1. add 2x 80x25mm ADDA case fans from my Lian Li - zip tied them together then placed them to blow over X800XT PE stock heatsink - there's a 2.5 cm gap between fans and heatsink

2. zalman fan bracket + 120x38mm panaflo 103cfm over video card and chipset of my motherboard = zalman fan bracket does wonders for every video card I've owned

3. oh yeah and using ATI Tool 0.22 i fixed fan speed at 100% instead default 67% - drops GPU idle temps by 4-10C!

And it allowed me to totally destroyed my old high score of 32,963!

Powercolor X800XT PE Stock heatsink @ 605/605 = 33,453

DX8.1 with Omega 8.07
Win2K SP4 latest updates

Full system config in compre url description field

Anyway thought I'd smash the 34k barrier

Powercolor X800XT PE stock heatsink @ 606/604 = 34,008

The is where I am now. seems to be the max of the GPU, tried 612mhz but froze in car tests.

Powercolor X800XT PE stock heatsink @ 610/605 = 34,056 (car low bug in effect as normally i can get 472fps at 606/604)

Aquamark3 rank #8 for X800 (R420) class

Powercolor X800XT PE @612/606 = 83,709
http://arc.aquamark3.com/arc/arc_vie...run=1558837733

3dmark05

Powercolor X800XT PE stock heatsink @ 609/606 = 6,831

14th November 2004, 1:42 AM	#1
eva2000 DDR1/DDR2/DDR3 Addict Join Date: Jun 2001 Location: Brisbane, Australia Posts: 21,887	Powercolor X800XT PE Adventures Thanks to Coolmodz.com.au OCAU sponsor special I managed to pick up a rare Powercolor X800XT PE AGP card It's stock speed is 520/560 according ATI Tool and idle and load GPU temps from ATI Tool = 36-39C idle and 64-68C load. Front side has Samsung 419 GC16 chips, Backside has Samsung 422 GC16 chips heh I used initially Omega 8.07 drivers then Cat 4.11 official. The Omega still was ~100pts faster in 3dmark2003 than Cat 4.11 for my card and faster in 3dmark2001 too. But here's a full set of 3dmarks using Cat 4.11 drivers at stock cpu speeds with my gaming OS image. AMD64 3700+ DFI NF3 250Gb 05/10/04 4v bios 2x 512MB KHX3200/512 BH-5 @ 200mhz 2-2-2-6 1T 120GB WD 1200JB LiteOn LDW-411s burner 580W Topower Adjustable PSU Win2k SP4 fully loaded gaming OS image Powercolor X800XT PE connected with OCZ VGA Enhanced power lead/molex 3dmark2001 = 29,147 vs Omega 8.07 score of 29,243 3dmark2003 = 14,033 3dmark05 = 6,704 __________________ * Questions about memory? No PMs Ask on forums * i7 980X Club @4844mhz H20 * Asus P6T6 WS Revo - W3540 @4821 H20 \| W3570 @4997Mhz H20 \| DFI UT X58-T3EH8 USB3.0 Last edited by eva2000; 14th November 2004 at 2:07 AM.

14th November 2004, 1:59 AM	#2
[ r2 ] Member Join Date: Aug 2004 Location: Sydney Posts: 1,022	Great results Eva, better performance over your Asus 6800Ultra that you owned. Be interesting to see results with your cpu being clocked. 600/606 is the max you achieved? Edit: Have you tried Cat 4.12 Beta? Last edited by [ r2 ]; 14th November 2004 at 2:00 AM.

14th November 2004, 2:03 AM	#3
eva2000 DDR1/DDR2/DDR3 Addict Join Date: Jun 2001 Location: Brisbane, Australia Posts: 21,887	Well I didn't have an AMD64 3700+ cpu when I had my Asus 6800U so results can't be totally comparable. 606mhz on mem is about it's max without artifacts. As to GPU core haven't gone much higher than 600mhz for now plus the past 2 days in Brissy have been bloody hot 32-34C! But the OCZ VGA Enhanced power lead did help some http://forums.overclockers.com.au/sh...d.php?t=326071 you read and decide if it's worth it hehe Cat 4.12betas don't work with Win2k, only WinXP. The bugs are WinXP specific and doesn't affect Win2K OS users so Cat 4.11 is what Win2K users will have to use __________________ * Questions about memory? No PMs Ask on forums * i7 980X Club @4844mhz H20 * Asus P6T6 WS Revo - W3540 @4821 H20 \| W3570 @4997Mhz H20 \| DFI UT X58-T3EH8 USB3.0 Last edited by eva2000; 14th November 2004 at 2:05 AM.

14th November 2004, 5:34 AM	#5
eva2000 DDR1/DDR2/DDR3 Addict Join Date: Jun 2001 Location: Brisbane, Australia Posts: 21,887	Well here's some results with cpu oc'd Room temps dropped 5C, so did GPU idle temps dropped from lowest 36C to 34C! AMD64 3700+ @ 11x 246HTT = 2706mhz 2x 512MB KHX3200/512 BH-5 @ 246mhz 2-2-2-6 1T Powercolor X800XT PE @ 520/560 = 13,336 @ 602/605 = 14,563 3Dmark05 @ 602/605 = 6,766 3Dmark2001 @ 602/605 = 32,470 Aquamark3 @ 602/605 = 82,000 Okay screenies 34C idle __________________ * Questions about memory? No PMs Ask on forums * i7 980X Club @4844mhz H20 * Asus P6T6 WS Revo - W3540 @4821 H20 \| W3570 @4997Mhz H20 \| DFI UT X58-T3EH8 USB3.0 Last edited by eva2000; 14th November 2004 at 6:43 AM.

14th November 2004, 5:45 PM	#6
eva2000 DDR1/DDR2/DDR3 Addict Join Date: Jun 2001 Location: Brisbane, Australia Posts: 21,887	Okay some FarCry v1.3 patched benchmarks AMD64 3700+ 12x 200 = 2411mhz stock DFI NF3 250Gb 05/10/04 4v bios 2x 512MB KHX3200/512 BH-5 @ 200mhz 2-2-2-6 1T 120GB WD 1200JB LiteOn LDW-411s burner 580W Topower Adjustable PSU Win2k SP4 fully loaded gaming OS image Powercolor X800XT PE connected with OCZ VGA Enhanced power lead/molex Using HardwareOC Farcry Bench v1.4 all benchmarks done with 4xAA and 4xAF @520/560 HardwareOC Steam Demo results HardwareOC Volcano Demo results TechReport Regulator Demo results @590/600 HardwareOC Steam Demo results HardwareOC Volcano Demo results TechReport Regulator Demo results __________________ * Questions about memory? No PMs Ask on forums * i7 980X Club @4844mhz H20 * Asus P6T6 WS Revo - W3540 @4821 H20 \| W3570 @4997Mhz H20 \| DFI UT X58-T3EH8 USB3.0 Last edited by eva2000; 14th November 2004 at 5:50 PM.

14th November 2004, 5:56 PM	#7
jamesrhadley Member Join Date: Apr 2003 Location: Brisbane Posts: 50	Just a quick one why are you running 2 2 2 6 I though 10-12 was better for amd 64's? Or was that only tru for xp's? __________________ JamesRhadley SOLTEK Nforce2 MRN -Mobile 35watt 2400+ at 2600MHZ BFG 6800oc softmodded 16+6 2x256 Ballisitix PC3200.

14th November 2004, 6:16 PM	#9
groovetek Member Join Date: Oct 2003 Location: SE Melbourne, 3150 Posts: 2,971	hey eva, do you bench at whatever feasible clocks for a particular benchmark suite, or do you use utilities like ATITOOL to REALLY find the maximum clock? It seems you're always using difference speeds for different benchmarks... that must mean the ones run at higher speeds are ones that can pass relatively reliably , but not necessarily 100% stable? edit: does this also apply to the rest of your system? I see that you memtest/prime and all that, but when it comes to ripping out big numbers, do you compromise stability at all? __________________ DNA Tuning (Australia) Last edited by groovetek; 14th November 2004 at 6:17 PM.

14th November 2004, 6:25 PM	#10
malonem Member Join Date: Jun 2004 Location: Brisbane, Queensland Posts: 282	eva how the heck u get 600mhz core on stock... even with my peltier @ -10c I could only get it to 640mhz (vmod too)

14th November 2004, 7:50 PM	#11
eva2000 DDR1/DDR2/DDR3 Addict Join Date: Jun 2001 Location: Brisbane, Australia Posts: 21,887	Yeah partially reason i'm using different speeds is the card is less than 2 days old, still figuring it out for each app As to ripping out the big numbers, they may not be 100% stable but for me they need to at least be superpi 16/32M, pifast, and memtest clean ... usually the cpu is the component that ain't 100% stable for me. ATI Tool no way accurate i.e. find max core differs from time to time... as with all ATI cards room/case ambient temps are are a factor - on a cool night you can probably get an extra 5-10mhz off the core compared to the hottest most humid day you can experience malonem... lucky i guess but why complain 640mhz on core is nothing to sneeze at __________________ * Questions about memory? No PMs Ask on forums * i7 980X Club @4844mhz H20 * Asus P6T6 WS Revo - W3540 @4821 H20 \| W3570 @4997Mhz H20 \| DFI UT X58-T3EH8 USB3.0

15th November 2004, 12:10 AM	#12
TazBazz (Banned or Deleted) Join Date: Oct 2004 Posts: 1,014	eva are you still using the stock cooler on the XT?

15th November 2004, 6:10 PM	#14
wastegate Member Join Date: Sep 2002 Location: North Shore Posts: 99	Interesting Eva, wary interesting. I have the same card but it came in a different box. What other cooling did you do to the card?

15th November 2004, 6:16 PM	#15
eva2000 DDR1/DDR2/DDR3 Addict Join Date: Jun 2001 Location: Brisbane, Australia Posts: 21,887	ah different packaging... the powercolor web site shows the packaging I have so either yours is newer or older packaging hehe only things i have done now are 1. add 2x 80x25mm ADDA case fans from my Lian Li - zip tied them together then placed them to blow over X800XT PE stock heatsink - there's a 2.5 cm gap between fans and heatsink 2. zalman fan bracket + 120x38mm panaflo 103cfm over video card and chipset of my motherboard = zalman fan bracket does wonders for every video card I've owned 3. oh yeah and using ATI Tool 0.22 i fixed fan speed at 100% instead default 67% - drops GPU idle temps by 4-10C! And it allowed me to totally destroyed my old high score of 32,963! Powercolor X800XT PE Stock heatsink @ 605/605 = 33,453 DX8.1 with Omega 8.07 Win2K SP4 latest updates Full system config in compre url description field Anyway thought I'd smash the 34k barrier Powercolor X800XT PE stock heatsink @ 606/604 = 34,008 The is where I am now. seems to be the max of the GPU, tried 612mhz but froze in car tests. Powercolor X800XT PE stock heatsink @ 610/605 = 34,056 (car low bug in effect as normally i can get 472fps at 606/604) Aquamark3 rank #8 for X800 (R420) class Powercolor X800XT PE @612/606 = 83,709 http://arc.aquamark3.com/arc/arc_vie...run=1558837733 3dmark05 Powercolor X800XT PE stock heatsink @ 609/606 = 6,831 __________________ * Questions about memory? No PMs Ask on forums * i7 980X Club @4844mhz H20 * Asus P6T6 WS Revo - W3540 @4821 H20 \| W3570 @4997Mhz H20 \| DFI UT X58-T3EH8 USB3.0 Last edited by eva2000; 17th November 2004 at 4:01 AM.