4890 OC limit fixed!

[ceemic]

T.Rex, real Vgpu max with soft is ~1,49V (due to the IC). You can move the slider forward, but Vgpu doesn't rise

[T.Rex]

T.Rex, real Vgpu max with soft is ~1,49V (due to the IC). You can move the slider forward, but Vgpu doesn't rise

oh,i dont knew that,but at this 1.49v i think 1325mhz is nice
but maybe its the reason to the bad performance
like it need more voltage ?

[dinos22]

your canyon looks weak
but i dont know AMD CPUs very well
it would be nice if you had Intel tests as well
1.4v is the limit actually not 1.49v

[W1zzard]

oh,i dont knew that,but at this 1.49v i think 1325mhz is nice
but maybe its the reason to the bad performance
like it need more voltage ?

can you measure to confirm if the voltage applies or not?

[Martin.v.r]

rivatuner report ? of 1325MHZ
why (ur/watch) stop all performance,at (MAX MHZ SET) in bios,(ur/watch)/gpu has a limit.
i can clock high to,but gives no more in 3d,but more voltset help
within the max limit of (ur/watch).

next time you hit 1325MHZ ,get a rivatuner report to

you ned a (ur/watch)which goes up to 1400/1500 MHz to the gpu max set, so I will show you a boost you get in performance
so max must be in 1100 with this (ur/watch) here, http://forums.techpowerup.com/attach...8&d=1245187251

had the same problem with the 3870x2 and (ur/watch) http://www.hwbot.org/result.do?resultId=805978 oc proceeded to watch the service was to fall, even with more volts .

using XP 32BIT GPUTOOL ver:0.0.0 to a MSI 4890 OC

[ceemic]

dinos22, I just rechecked Vgpu on my Sapphire and it reads ~1,488V under load. I use GPUTool to change Vgpu...

[W1zzard]

1.4v is the limit actually not 1.49v

using which software?

[T.Rex]

rivatuner report ? of 1325MHZ
why (ur/watch) stop all performance,at (MAX MHZ SET) in bios,(ur/watch)/gpu has a limit.
i can clock high to,but gives no more in 3d,but more voltset help
within the max limit of (ur/watch).

next time you hit 1325MHZ ,get a rivatuner report to

you ned a (ur/watch)which goes up to 1400/1500 MHz to the gpu max set, so I will show you a boost you get in performance
so max must be in 1100 with this (ur/watch) here, http://forums.techpowerup.com/attach...8&d=1245187251

had the same problem with the 3870x2 and (ur/watch) http://www.hwbot.org/result.do?resultId=805978 oc proceeded to watch the service was to fall, even with more volts .

using XP 32BIT GPUTOOL ver:0.0.0 to a MSI 4890 OC

i just got my new asus re,and test it later,then i use this 4890 and see what happens but 1.4v is nothing norm is about 1.31v,but martin i call you when im ready and then we can have a talk about it

[Martin.v.r]

ok T.REX
you can get tlf:nr on octeamdenmark/msn

[chew*]

your canyon looks weak
but i dont know AMD CPUs very well
it would be nice if you had Intel tests as well
1.4v is the limit actually not 1.49v

And you would be 100% correct except for voltage under normal circumstances.

Phase ( mach1 ) 720 BE card on air AC cooled room.



What SF3D has is the missing piece to my puzzle and bios mod that I have been working on and saying for a while and getting told NO NO NO.......software works fine.

Glad this is coming from him and macci.

[dinos22]

ATI cards can be funny when pushed too hard or at the wrong clocks

but he very well may have a card that is capable of benching with stable 3D numbers past 1250 who knows heh

to be honest 1.4v was something i was told recently. i benched 4890s and have returned them since without checking with DMM but if it says 1.488v that should be pretty good for high LN2 clocks

[chew*]

ATI cards can be funny when pushed too hard or at the wrong clocks

You nailed that on the head.

[ownage]

your canyon looks weak
but i dont know AMD CPUs very well
it would be nice if you had Intel tests as well
1.4v is the limit actually not 1.49v

As others said, not true

What's the actual limit? Did go up to 1.53v so far.

[ceemic]

Did go up to 1.53v so far.

Measured with DMM?

[ownage]

Measured with DMM?

Nope, but the program also reports the current voltage is 1.53v. I know it doesn't have to be reliable. Will do more testing next week, with DMM.

[runmc]

Brandy- What was the temperature and model of the card? Stock bios without mods and what was the core voltage?

This thread is not about these rare LN2 clocks. If 100/100 cards would do 1200MHz on LN2 out of the box, then this thread would be worthless

When this bios will be in public use, we will see some sick overclokcs for sure.

Do you have any news for us?? Updates please!!

[massman]

Petri and me tested the bios thursday-night, but it's not all positive news. I'm not sure if I can talk about this ... so you'll have to wait for SF3D to update.

[chew*]

Petri and me tested the bios thursday-night, but it's not all positive news. I'm not sure if I can talk about this ... so you'll have to wait for SF3D to update.

Hmm even more interesting....maybe I should just release my bios once it's served its purpose.

[runmc]

Hmm even more interesting....maybe I should just release my bios once it's served its purpose.

What seems to be the hold-up??

[chew*]

What seems to be the hold-up??

I have some 3d to tackle....how screwed would that be if someone beat me before I got a chance to fully utilyze it myself

[runmc]

I have some 3d to tackle....how screwed would that be if someone beat me before I got a chance to fully utilyze it myself

LOL - not to cool I don't blame you chew*. Milk that baby for all it's worth.

[massman]

Hmm even more interesting....maybe I should just release my bios once it's served its purpose.

Well, the issues we ran into were most definitly interesting.

Would the only reason not to make your bios public be that you don't want others to beat you before you can make any results? Because ... that's not the reason why this one isn't made public yet

[chew*]

Well, the issues we ran into were most definitly interesting.

Would the only reason not to make your bios public be that you don't want others to beat you before you can make any results? Because ... that's not the reason why this one isn't made public yet

Yah no problems with this bios toasting anything not that that bios your on did at least I hope it didn't.

I to have tested on a stack of cards.....other than the weird screen over 1065 core....no issues. Of course I have not played cold on them but I doubt it will make a diff.....I know there is a bug just haven't found it yet. The screen issue is funky to.....03 gives gray/darker gray lines......vantage is hilarious a colorful reddish/orange/pink screen with an AMD logo.

Software clock the cards easily higher.......however according to results well USE whats fast......not what you think is fast. Canyon points to the fact that software is not working.
Vantage screen stutters like crazy when loading using software.....scores lower......

This bios was done solely by me no outside help, It also only works on ref cards...

[[XC] gomeler]

These secret BIOS' hardly matter considering how much power there is in 4x HD4890 at 1GHz, let alone stock. This is like HD4870x2 CFX, you are already CPU bottlenecked at 850/950 and see scaling in only 03. Plus chew*, you need LHe to be competitive on AMD

What we need is another Vantage to even things out, Vantage sort of keeps scaling, especially High/Extreme, so where is 3DMark2010?

[chew*]

ehh scaling dies at 6.2 with 2 cards...till you increase clocks over 1000/1150....then they scale again.

As for Lhe...eh 6390 3d stable on ln2 nothing to cry about.....see If i can push the bar higher in the future.....