watch out Gainward, Albatrons in the house!

[walldow]

just got the card little over an hour ago!
http://www.allstarshop.com/shop/product.asp?pid=8186&sid=AAT2J1DB2HL58GSQ7DVU7GSBE5MKAH46#

took second in 2003 with my first full overclocked run!
http://service.futuremark.com/compare?2k3=1675824

the card auto detected 534/1020. and runs 620/1050 with stock cooler and stock voltage! my system wasn't tweeked as you can see i was only running 2300mhz on my cpu and the fsb is low. wasn't planning on submitting the score i was just wanting to see what the card could do, and bam 2nd spot. hell the thing didn't even break 45c!

however the ram runs extremly hot i touched the stock ram sinks and ouch but they are aluminum. i have some all copper BGA ram sinks. i think with propper cooling the ram can hit close to 1100.

i'm just speculating but with core temps like this, with a volt mod and water cooler i'm looking at close to 700mhz! ya that sounds crazy but i'm dead seriouse!

i plan on playing with the card for a couple of days and finding it's limits. i'll have time to do some seriouse overclocks come fri or sat.

veary impressed with Albatron.

Walldow- OC 3DMark Team

[Pollux]

awesome. are u going to watercool it when u do the voltmod?

[walldow]

Originally posted by Pollux
awesome. are u going to watercool it when u do the voltmod?

ya, but i'll use the stock cooler after the volt mod just to see how the core handles heat and find the core threshhold. giveing me a little better insite on how mutch voltage i can stick too it. on my 5600 core threshold was around 109c anything over that and it started to get boggy!

the 5700 ultra is almost identical to the 5900. as far as size and layout it's pretty mutch the same. the IC (voltage controller) is also same as the 5900. i guess because of the 256bit core archetecture.

as soon as i do the mod i'll update my "volt mod youre FX" thread for the 5700!

[Sir Barton]

you have a hell of a card their walldow. the highest i can get mine so far is 560/1040. when you get to voltmoddin' spread the knowledge on this way :)

[NewbiePerson]

just wondering but does ur abit board support over 12.5 multipler without the pin mod?

[walldow]

Originally posted by NewbiePerson
just wondering but does ur abit board support over 12.5 multipler without the pin mod?


to be honest i really don't know i use the higher fsb speeds so i don't run any multies over 11!i think it depends on the cpu you are running most of the new mobos can unlock the new athlons so all multipliers are open and useable. but next time i reboot i'll give it a try and let you know.




update on my card guys! i couldn't get anything over 625/1050 with stock cooler and voltage.!i managed to get 1080 on the ram after i put on my BGA copper ram sinks. i was only able to hit 4800 in 3d mark 2003 @625/1080 with stock voltage.

BUT! the core has alot of headroom i done the volt mod on the card an reached 700mhz core clock with ease! i'll post a screen shot later.

took first in 2003 non WHQL drivers 53.03. i haven't run the 52.16's yet but plan on doing so tomarrow.

http://service.futuremark.com/compare?2k3=1690361
:D

screeny
http://www.theforumisdown.com/uploadfiles/users/hayes/files/700mhz core clock.JPG

[NewbiePerson]

wow thats sweet. How mch did that card cost again?

[walldow]

$195 free shipping


i wonder how the ATI 9600XT's are faring up? does snyone know?:p

[micamica1217]

Originally posted by walldow
$195 free shipping


i wonder how the ATI 9600XT's are faring up? does snyone know?:p

I have to say that, that is a fine OC you got there.....nice scores too.
let us know what you get with the 52.16 drivers, OK?

btw, I don't think you could push a 9600xt to such mem and core speeds, but a score of 45xx is not out of the question.

mica

[walldow]

Originally posted by micamica1217


I have to say that, that is a fine OC you got there.....nice scores too.
let us know what you get with the 52.16 drivers, OK?

btw, I don't think you could push a 9600xt to such mem and core speeds, but a score of 45xx is not out of the question.

mica

thanks!

the core runs alot cooler, witch is usually the main factor when atempting a good OC. i really had to struggle to get my 5600 ultra to reach 530mhz. the thing ran super hot. the fx5700 ultra is taking it in stide!

major thumbs up to nvidia on this card!

[Teh]

Cool, a 5700 Ultra approaching Radeon 9800 3DMarks.

[IR1]

Originally posted by Teh
Cool, a 5700 Ultra approaching Radeon 9800 3DMarks.

4800 isnt 9800 type 3dmarks... I get 5800 with my 9700pro oc'd. So I know the 9800's must do better.


that being said thats an awesome overclock. Keep it up and let us know what you get after volt mod.

[Overclocker550]

man amazing overclock! hows she do in 2001 also?

[walldow]

Originally posted by IR1


4800 isnt 9800 type 3dmarks... I get 5800 with my 9700pro oc'd. So I know the 9800's must do better.


4800? the score was 5044!

anyway, i don't know what the 9800's-9700's are doing but i have to say this is one of the badest 4 piped cards i've ever come across.

also you have to remember the 9800 has 8 pipes!:eek:


as fore 3dmark 2001. my system is to slow. 2001 is more cpu and fsb oriented. even thow my card is clocked faster the guys with the p4's and AMD 64's blow me away. i'm only hitting 16000! but Zakelwe from extremsystems is running a p4 @3.6gig on a canterwood 800mhz FSB and he's getting 20000's. 3dmark 2003 gives me a little leveler playing field.it's more card oriented. after x-mas i plan on going AMD 64. that should help the 2001 scores out a bunch!

[L337 M33P]

O_o nice overclock - My Radeon 9600XT is arriving tomorrow or Tuesday - let's hope my dad lets me use it before christmas :)

Apparently the R360 core overclocks extremely well too... let the games begin :D

[fafnir]

walldow, save ur money for an 940/939fx system when it comes out

try ur best to wait it out, cause i believe that it'll be worth it

http://forum.oc-forums.com/vb/showthread.php?s=&threadid=251653

just got enough horsepower to blow any p4 short of the extreme edition outta any mass of water minus the bathtub

though i'll tell you now i got one of em' as a gift just as i ordered my own, which means i just wasted another 400 bucks replacing a system which really didn't need replacing, but oh well, so much for the barton system anyways

but seriously, good hunting keeping the 5700 top spot on 3Dmark03 man!

[zakelwe]

Nice scores an clocks walldow ! :toast:

:)

Regards

Andy

[Pollux]

Originally posted by walldow
after x-mas i plan on going AMD 64. that should help the 2001 scores out a bunch!

i dont think increasing already over-inflated synthetic benchmark scores is worth it. have u noticed any type of noteworthy improvement in your gaming?

[ajrettke]

pollux, people do things for different reasons. Some may not seem "practicle" but none the less we try for the best. Why do you spend a lot of cash on your corsair C2, you could get cheaper stuff to run same FSB at lesser timings...you wouldn't notice the difference either...but it matters to your most likely....same for your XT, even with a pro you may not have noticed the difference.

Just gotta push the envelope, nice OC walldow!

[micamica1217]

Originally posted by ajrettke
pollux, people do things for different reasons. Some may not seem "practicle" but none the less we try for the best. Why do you spend a lot of cash on your corsair C2, you could get cheaper stuff to run same FSB at lesser timings...you wouldn't notice the difference either...but it matters to your most likely....same for your XT, even with a pro you may not have noticed the difference.

Just gotta push the envelope, nice OC walldow!

what pollux is mostly talking about is his slightly inflated scores, while using the 53.03 drivers......
once he trys the 52.16 he'll see his GT4 and vertex shader scores go down by 25%
his GT2 and GT3 will go down by 13%....
and his total score will drop about 15%
notice that in real world gaming....his FPS will not drop at all.
(or only 1% or 2% max)

mica

[Pollux]

Originally posted by ajrettke
pollux, people do things for different reasons. Some may not seem "practicle" but none the less we try for the best. Why do you spend a lot of cash on your corsair C2, you could get cheaper stuff to run same FSB at lesser timings...you wouldn't notice the difference either...but it matters to your most likely....same for your XT, even with a pro you may not have noticed the difference.

Just gotta push the envelope, nice OC walldow!

i made my comment to learn from his experience with the card, not to be critical. my video card decision basically came down to the 5700U or the 9600xt. i was aware of all the driver cheats, but i knew the 5700U OC'ed like crazy. so basically i wanted to know if an OCed 5700U would outperform a 9600xt. from what everyone here said, i came to the conclusion that it would not. about the athlon 64, i dont consider it worth it if one of the primary reasons to get is an increase in an inaccurate and inflated benchmark score. but everyone has their hobbies.

[NewbiePerson]

One thing is for sure on the texture part of the 3dmark test i'm sure you beat any current non oc'd card on the market. You should get fraps up and running and run some benchies.
at 1280*1024 with max al and af I get an average of 91fps on max payne 2 (first couple levels)with the same settings I get 103fps on call of duty. (the trench level where you blow up the anti-tank guns, forget exact name...)
I'd like to know what you ati fans are getting ;)
Though it's not a true dx 9 measurement those games do utilize some dx 9 features.
On my 5800 on default (400/800) 5300marks and something. thats with me using the 53.03 driver as well but my cpu is only around 2.2ghz thats why I was asking about that abit board of yours.
I'm afraid to oc my video card ram that much just touching the heatsink on the ram is exceptionally warm. You were talking about the gpu 45c being good temps. mine is running at 41-46 c so before i oc it what would be a bad gpu temp? say 55 c or higher?

[walldow]

well there is a 13% decrees in overall scores from the 53.03 to the 52.16.


and as fare as i'm concerned as long as the end product (actual redered picture) has not been physicly altered than the actual rendering process should be left up to the card manufacturer. no matter how it's rendered as long as it's redered correctly, than it shouldn't matter! as fare as i know there are no phisycal differances in the end result (actual picture) between the 52.16's and the 53.03 in 3dmark 2003. unless someone has a link stating otherwise? i might be wrong? havn't been keeping up with the latest he said she said stuff!


anyway as fare as telling the differance in games with the higher clock speeds? yes i can tell a differance. i play final fantasy online and running the higher clock speeds 700/1050 the trees are more lifelike and panning is alot smoother. as for other games i can't say because i spend all my time iether being a father, working, playing ff XI, or benching. haven't really had time to play anything else.

anyone know how the 5700's are doing in AQUAMARK? might download it tonight and give it a spin!

[Teh]

Originally posted by IR1


4800 isnt 9800 type 3dmarks... I get 5800 with my 9700pro oc'd. So I know the 9800's must do better.


that being said thats an awesome overclock. Keep it up and let us know what you get after volt mod.

Average scores are about 5200, keep in mind I'm talking about the NP version. And 9700 PROs w/ 2.8NS RAM@ STOCK usually get ~4800.

[micamica1217]

Originally posted by Teh


Average scores are about 5200, keep in mind I'm talking about the NP version. And 9700 PROs w/ 2.8NS RAM@ STOCK usually get ~4800.

not even close.......

my 9700pro AIW (never OCed gets 5800 in 3dmark03.....with sliders turned on high, yet no AA and no AF, I get about 5400(I may have even had vsync on too.))
I can't see my system being that much better then most here in the forums.....I'm even using a 2.4b not a 2.4c

no way that most systems with a 9700pro will get 1000 points less then me.

mica

[Sir Barton]

Originally posted by walldow
anyone know how the 5700's are doing in AQUAMARK? might download it tonight and give it a spin!

my score of 35,679 got the #3 spot in aquamark3 for the 5700U. check my sig.

[micamica1217]

Originally posted by NewbiePerson
One thing is for sure on the texture part of the 3dmark test i'm sure you beat any current non oc'd card on the market.

sorry, but he didn't beat any of my scores, and my card is not OCed.

You should get fraps up and running and run some benchies.
at 1280*1024 with max al and af I get an average of 91fps on max payne 2 (first couple levels)with the same settings I get 103fps on call of duty. (the trench level where you blow up the anti-tank guns, forget exact name...)
I'd like to know what you ati fans are getting ;)
Though it's not a true dx 9 measurement those games do utilize some dx 9 features.

max payne 2 is mostly a DX8 game.....it does have 5.1 surround sound, and some true mapping, but no PS2.0 or VS2.0 support.
COD's highest shader support is 1.1 for PS and VS.....that sounds more like a OpenGL game on the same lvl as Q3TA or a representitive of DX8.0, not even DX8.1

On my 5800 on default (400/800) 5300marks and something. thats with me using the 53.03 driver as well but my cpu is only around 2.2ghz thats why I was asking about that abit board of yours.
I'm afraid to oc my video card ram that much just touching the heatsink on the ram is exceptionally warm. You were talking about the gpu 45c being good temps. mine is running at 41-46 c so before i oc it what would be a bad gpu temp? say 55 c or higher?

mica

[micamica1217]

Originally posted by walldow
well there is a 13% decrees in overall scores from the 53.03 to the 52.16.


and as fare as i'm concerned as long as the end product (actual redered picture) has not been physicly altered than the actual rendering process should be left up to the card manufacturer. no matter how it's rendered as long as it's redered correctly, than it shouldn't matter! as fare as i know there are no phisycal differances in the end result (actual picture) between the 52.16's and the 53.03 in 3dmark 2003. unless someone has a link stating otherwise? i might be wrong? havn't been keeping up with the latest he said she said stuff!


as you can see my predictions on your scores with the 52.16 drivers were farly acurate.
I know, because I've tested a 5900u, and can't see your scores being much differant.
as you know, the end product was not the same, with the 52.16 drivers, durring testing, using the old as well as the new build of 3dmark03.
if you read my post, you'll notice that the VS and GT4 were the biggest changes, with GT2 and GT3 following.
almost nothing could realy be noticed if not compaied to the refrance images....even then some were real close and nit picky.
with every nVidia PR rep saying that they will just redo a new driver that basicly does the same thing as 52.16 pre 340 build,
the 53.03 was born.

please understand that I will not talk about IQ and the 53.03 drivers beyond what I've said....it's just too off topic.

btw, again, realy nice OC on your card.
good luck with aquamark

mica

[Teh]

Originally posted by micamica1217


not even close.......

my 9700pro AIW (never OCed gets 5800 in 3dmark03.....with sliders turned on high, yet no AA and no AF, I get about 5400(I may have even had vsync on too.))
I can't see my system being that much better then most here in the forums.....I'm even using a 2.4b not a 2.4c

no way that most systems with a 9700pro will get 1000 points less then me.

mica
I don't know if I believe that, multiple threads report brand new, unoverclocked 9700 Pros get about 4800-5000 on average.

[OC Noob]

Sweet O/C!

[micamica1217]

Originally posted by Teh

I don't know if I believe that, multiple threads report brand new, unoverclocked 9700 Pros get about 4800-5000 on average.

I have no reason to lie.....here (http://service.futuremark.com/compare?2k3=791646) is a realy old test when I just got my card.

get better info then you are putting out.
I'm the second person to state 58xx scores with a 9700pro within this thread.

mica

[NewbiePerson]

thats why i said neither game was a true measurement of dx 9
I think it also uses a special lighting effect from dx 9
and please post ur 3dmark link

[NewbiePerson]

as for the texture part I just compared a 9800xt multi texture fill rate to my card the 9800xt was 3269.6

[NewbiePerson]

eerrr ignore that last post
my mistake i thought the multi texture part was more graphics processor intensive. however i checked the scores of a 9700 and he's very close to the fastest 9700 (which isn't using certified drivers) and it's total score is 5074 running 335(285) I dont think the mem is running at pro but the gpu definately is.
I think he would be extremly close to a default non'cd 9700pro.
The highest score a 9700pro gets is like 6800 thats heavily oc'd and with a super fast processor.
my multi texture fill rate beat the fastest 9700pro
which used omega and was oc'd to 466gpu and 369mem

[micamica1217]

Originally posted by NewbiePerson
thats why i said neither game was a true measurement of dx 9
I think it also uses a special lighting effect from dx 9
and please post ur 3dmark link

my link to 3dmark03 is listed at the top of this page.

mica

[NewbiePerson]

Do you have an update ver?

[micamica1217]

Originally posted by NewbiePerson
Do you have an update ver?

no, but I haven't benchmarked since I've realy put this last CPU and mobo in my rig.

btw, if your wondering.... no, the numbers or scores will not realy change at all with any driver or build using an ATI card.

I'll be right back with a linky....look. (http://www.nordichardware.se/artiklar/Grafik/2003/3DMark03_340/)

see how with the ATI 9800, you have only 4 tests that have a change of more then 2%....yet all changes are less then 3%
7 tests are unchanged......thats out of 14 tests, plus the final score.
all well within the margin of error.

mica

[Sir Barton]

hey walldow, did you run aquamark yet?

[NewbiePerson]

that was versus the 3 build to the 4 build
you are using the 2 build.

[micamica1217]

Originally posted by NewbiePerson
that was versus the 3 build to the 4 build
you are using the 2 build.

and????

as stated before, with ATI the scores basicly remain the same...
no matter what the build, or driver.

I think you are confusing yourself with nVidia....
were the scores can change drasticly from driver to driver, and from build to build.

what would be the point, to retest and list it, if the scores would be about the same?

mica

[Sir Barton]

simply put, mica doesnt need to rerun his benchie because ATI doesnt need to put out new drivers so their customers are happy with their benchies.

[NewbiePerson]

meh

[walldow]

Originally posted by Sir Barton
hey walldow, did you run aquamark yet?

not yet i'm running a 56k motem and it t will take all night to down load. i'll start down loading tonight and it might be done by somtime tomarrow! as soon as i do i'll let you know!

[walldow]

well guys just broke 720mhz and first to brake 5100 in 5700 ultras.


screenshot and link to ORB in this post:
http://forum.oc-forums.com/vb/showthread.php?s=&threadid=251566

[NewbiePerson]

your gaining on my 5800. Very gj
just wondering did nvidia improve their vertex shaders with the 5700 and 5950? you beat me quite well in that area.
http://service.futuremark.com/compare?2k3=1690361

and what's with my extremly low audigy sound score?
i'm almost losing to onboard via 97 sound. or is that particular benchmark no good for true sound measurement?

[walldow]

Originally posted by NewbiePerson
your gaining on my 5800. Very gj
just wondering did nvidia improve their vertex shaders with the 5700 and 5950? you beat me quite well in that area.
http://service.futuremark.com/compare?2k3=1690361

and what's with my extremly low audigy sound score?
i'm almost losing to onboard via 97 sound. or is that particular benchmark no good for true sound measurement?

actualy the 5700 uses the 256bit core desighn like the 5900 and 5950. were as the 5800, 5600 ,and 5200 used 128bit.

i think thats rite?

[violineb]

Nice O/C for sure! I don't mean to make noobish comments but wouldn't it be easier just to get something like a XFX of eVGA 5900 non u and clock it to 5900 ultra speeds? Or are these clocks already better than a 5900 ultra? Just wondering ;)

[NewbiePerson]

i thought 5800 had 256core 128mem it has ddr2 though
and it has 8 pipelines (4*2)
i think the 5700 has some kind of shadow boost onboard and ddr2 other than that its the same as the 5600 256core 128mem 4 pipelines, or so I thought. It is quite a boost over the 5600 though. I looked at 9600 3dmark scores the highest I saw was 4700 or so.
So you are doing very well
the highest 9700np is like 5077

[zakelwe]

Originally posted by NewbiePerson
your gaining on my 5800. Very gj
just wondering did nvidia improve their vertex shaders with the 5700 and 5950? you beat me quite well in that area.
http://service.futuremark.com/compare?2k3=1690361

and what's with my extremly low audigy sound score?
i'm almost losing to onboard via 97 sound. or is that particular benchmark no good for true sound measurement?

It has 3 times the vertex processing power as the 5600U so you are quite right.

The only other differences are, as you say DDR2 and the IBM core, which runs a lot cooler than the 5600U and overclocks a lot higher. The top 5600U overclock is 540Mhz and the top 5700U overclock is 720Mhz at present.

The DDR2 memory also overclocks very well also, it is 2.2ns so rated at 900Mhz effective, but 1050+ is attainable.

If IBM are making the nv40 it is looking very good if the 5700U is their practice shot.

Regards

Andy

[f00t]

Originally posted by violineb
Nice O/C for sure! I don't mean to make noobish comments but wouldn't it be easier just to get something like a XFX of eVGA 5900 non u and clock it to 5900 ultra speeds?

that's the whole point of budget silicon. you get it, oc the hell out of it, and laught at people who bought much more expensive card cause you are almost on the same performance level as they are.

gotta love budget vid cards. i know i do

[walldow]

Originally posted by Sir Barton
hey walldow, did you run aquamark yet?


droppped it this morning! hit 43565. beat out the top spot by 6021 marks. running 700/1050@2360mhz cpu:

http://arc.aquamark3.com/arc/arc_view.php?run=206355581

some fine tweeking and 44-45000 maybe?

[walldow]

Originally posted by f00t


that's the whole point of budget silicon. you get it, oc the hell out of it, and laught at people who bought much more expensive card cause you are almost on the same performance level as they are.

gotta love budget vid cards. i know i do


true that!

i bought the card as a toy just to see how fast i could make it go! it's acually turning out to be a sweet card(gaming wise)! 100% better than the 5600. the 5600 ran way hot. the 5700 is cold as ice!alot more headroom.

[Fuzion]

droppped it this morning! hit 43565. beat out the top spot by 6021 marks. running 700/1050@2360mhz cpu:

You should be able to squeeze some more out of it. My 9800XT is hitting 50K with very little tweaking, I think I can eventually hit 60K with this setup, we'll see.

Good Job and Good O/C Man, keep up the good work.

[zakelwe]

got your AMD64 yet walldow, looking forward to the results.

Regards

Andy

PS Tried to volt mod the 5700U again and again it failed .. Gainward do not have the standard component :rolleyes:

[walldow]

no budy me and the old lady had a fallen out, and i moved out!:cool: all my money is tied up on bills and such! no extra for system upgrade anytime soon.if i had one i'd show ya 6K in 2003 though (on a 4 piped card. HEHE!) and thats no lie! but i am working on a new mod for the 5700 ultra has to do with the triggering of the soft start. it happens when voltage between the uper and lower mosfets getts greater then 25% (i think thats what the tech sheet states) but anyway i'm thinking if i can change the mosfets to handle a bigger power load than this soft start bull will be out the window. i truly think the card can handle it! i'm pulling 780mhz @1.71 volts and the thing wont get over 5c on my icewater settup. the only thing holding me back is those darn mosfets! i'm thinking if my calculations are correct the thing should be able to hit 900mhz at around 1.9-2.0 volts the only down fall is i dont know if the memory bandwith at 1200mhz will keep up with it? after i get the full mod done i plan on writing an artical on it. thats the only reson i havn't done so yet is because im not done modding!:cool: sorry to here youre modd didn't turn out! take care ad let me know if you need anything!

Walldow- OC 3DMark Team