Can you guess what Im thinking?

[youngpro]

Today I do the vmods to a 4890 to see what sort of clocks we can get out of it on LN2.. PROPER clocks..

SAPPHIRE Vapor-X 4890 (vgpu, vddr, vddq)
GIGABYTE P55 UD6
Intel 870 @ 3200
GSKILL 16000 C9
GSKILL Falcon SSD 64GB

http://i200.photobucket.com/albums/aa50/preeeezy/1325_1200_3d01nature.jpg

http://i200.photobucket.com/albums/aa50/preeeezy/4890_front.jpg

http://i200.photobucket.com/albums/aa50/preeeezy/4890_vmods.jpg

http://i200.photobucket.com/albums/aa50/preeeezy/IMG_0678.jpg

[TMWH]

Nice Vmods there pro!

[subaruwrc]

I think you are thinking about, vmodding 3 other cards and put them into a classified and do HWBOT/ORB damage :D
With only 1.3v its insane....

[dinos22]

stop thinking
do it
do it :D

[K404]

ok... I know nothing about ATI benching but are you getting full performance there? 636FPS in nature is.....crap.....

[Vivi]

wicked sick :D
did u use vaseline


k404 its win7 and only 3200 cpu ;)

[youngpro]

forget about the frames mate, its windows 7, theres no driver optimization, theres no lod.. you should know me better than that, my test runs are always the worst performance product on earth ;)

let me think how to word this properly.. i dont want to be in the same position as you roodt... i used dielectric grease to insulate my graphics card... that seems safe to me :)

[Oliver]

yam yam, thats tasty !

[Vivi]

lol james, im a bad spot ya

die-electric grease sounds usable, i should get that rahter

[carpo93]

nice mod

[proxis]

two 4890s + e8600 ?
four 4890s + i7 ?

[N3RO]

Thinking about >130k on 01 with ATI :)

[SF3D]

I can guess, but don't spend your time with 4890 anymore :D

I have seen the light ;)

[zalbard]

Did you use the modded BIOS for 4890?
And nice clocks! :D

I have seen the light ;)
Do tell! :D

[hiwa]

Nice mod james
what's your plan with them? :up:

[stummerwinter]

I can guess, but don't spend your time with 4890 anymore :D

I have seen the light ;)

:eek:

Sounds like 5870 (X2)...:yepp:

[massman]

That core frequency is absolutely fantastic, James! Nice work.

@ SF3D: LOL :D

[youngpro]

two 4890s + e8600 ?


you got it bro

[dinos22]

Petri u party pooper :p:

[MusicIsMyLife]

Nice mods. Good luck with your plans, pro. :yepp:

[hipro5]

I can guess, but don't spend your time with 4890 anymore :D

I have seen the light ;)

You filthy pron!.... :D
You and your filthy friend macci!... :D :p:

We are "hungry" for power down here in Greece you know!...That's why I haven't moded those 4890s of mine till now..... :) ;)

[lolhalol]

lol@ hipro..

nice core clocks james.. even though the new cards are near.. maybe a pre launch WR then smash that WR with the 5xxx cards again on teh same day? hahahha

[FUGGER]

Where is the rest of the tests? Next to no scaling with Crossfire in 01

Ive been trying to figure out nv/nhm ch bug with no success. Noticed you picked the worst test in 01 Nature. If you could bring up the cpu some so it is not so cpu bound I am sure you would pass the cf setup at 1Ghz.

For compare;

P55 2k1 ATI (http://fugger.netfirms.com/msi012.jpg)

P55 2k1 Nvidia (http://fugger.netfirms.com/msinv2.jpg)

[dduckquack]

you thinking of making pancakes?

[youngpro]

Where is the rest of the tests? Next to no scaling with Crossfire in 01



yeah its kind of sad that p55 sucks so far for 01

didnt bother with the other tests as im planning to use x48 as its still 01 king

[massman]

P55 2k1 Nvidia (http://fugger.netfirms.com/msinv2.jpg)

What's up with those scores? The system tests look awful compared to the relatively untweaked runs I've here.

What B2B are you at?

[FUGGER]

B2B at 1, lets see your results.

[SF3D]

What's up with those scores?


Hehe! This reminds me about some statement made in Fooz Bar :rofl:

I will break ... ALL ... WR's with P55 :D

Waiting to see the video about it.

[Esajuppeli]

No-one needs i5/LGA1156. Core 2 Duo/Quad does almost it all, and i7 does it even better.

[massman]

B2B at 1, lets see your results.

Assembling P55/HD4890 as we speak.


Hehe! This reminds me about some statement made in Fooz Bar :rofl:

I will break ... ALL ... WR's with P55 :D

Waiting to see the video about it.

IIRC, I did specify the P55 brand!

:rofl:

[massman]

B2B at 1, lets see your results.

Couldn't be more wrong about the scores apparently; comparing to the X58 results I was referring to yours looked awful but when testing pretty much the same config the number don't look that bad. I'm hitting 82-83k consistently at 4.7k 2000CL7 with a single card. CH should maybe be a bit higher, tho, 372 at 4.7 here

But, apart from that: :up:

[dinos22]

Where is the rest of the tests? Next to no scaling with Crossfire in 01

Ive been trying to figure out nv/nhm ch bug with no success. Noticed you picked the worst test in 01 Nature. If you could bring up the cpu some so it is not so cpu bound I am sure you would pass the cf setup at 1Ghz.

For compare;

P55 2k1 ATI (http://fugger.netfirms.com/msi012.jpg)

P55 2k1 Nvidia (http://fugger.netfirms.com/msinv2.jpg)
i messed around with 01 a little while back just for kicks
higher CPU clocks, nasty RAM clocks heheh
lower GPU clocks, different card (i think it was 950MHz core, 930MHz memory on this try)

http://img9.imageshack.us/img9/4784/3dmark01105k9502000154v.png

[dinos22]

Assembling P55/HD4890 as we speak.



IIRC, I did specify the P55 brand!

:rofl:

nehalem is gaaaaaaaaaaaaaaaay :p:

http://img34.imageshack.us/img34/2715/93389377.png

[Vivi]

i agree with dino's :). nehalem needs to take a trip to the you know where bar.


I believe 2001 dualcard might have a chance on p55, but its one of those, "only james can do it" things. just like when he got 01 record with that DFI board.

[tiborrr]

Impressive clocks on that HD4890, James! :eek: I could bench '06 at 1090/1165MHz watercooled (30°C water), with a bit of subcooling it could do 1115-1120MHz. Too bad this card wasn't mine. My HD4890 stops at 1020/1160MHz, 1.42V. Gotta give it some auto-cascade love :up:

Best Regards,
N.

[TheKarmakazi]

wow nice pro!! :clap: Cant wait to see what kinda damage this does with a nice 8600 :up:

[Gautam]

Hey I got a great idea, lets "legalize" mipmap for Nehalem. That way the playing field will become more even and massman might get his wish :D

[massman]

I have called upon thousands of overclockers to prove me right.

It's either that or have an extreme amount of examples that prove me wrong :D

[massman]

B2B at 1, lets see your results.

Finally had some time to digg into the 01 this evening.

http://i154.photobucket.com/albums/s261/Massman-/89190.png

What are you doing with:

- that extra 300MHz cpu?
- that massive memory bandwidth?
- that extra bit of uncore?
- that cascade?
- that second card?

Haven't even been tweaking that much AND, quite frankly, I'm not even close to the best ATI tweaker around.

[youngpro]

your running N first to get the most out of nature since its ati weakest?

have you tried something like

LH-LL-N-CH-DH-DL-CL?

i guess this approach is trying to get the most out of the ati strongest bench...

[Gautam]

Finally had some time to digg into the 01 this evening.

http://i154.photobucket.com/albums/s261/Massman-/89190.png

What are you doing with:

- that extra 300MHz cpu?
- that massive memory bandwidth?
- that extra bit of uncore?
- that cascade?
- that second card?

Haven't even been tweaking that much AND, quite frankly, I'm not even close to the best ATI tweaker around.

Yo Pieter-Jan, I thought ATi was supposed to be stronger than nV in lobbies, what gives? :p:

http://www.hwbot.org/result.do?resultId=897181

[massman]

your running N first to get the most out of nature since its ati weakest?

have you tried something like

LH-LL-N-CH-DH-DL-CL?

i guess this approach is trying to get the most out of the ati strongest bench...

I ran: N-CH-LH-DH-LL-DL-CL, because:

N: Ran at 1055/1205, which is very unstable. If I don't run it as first I can't get those frequencies stable. In addition, this is the test where you get most from higher clocks; 25MHz more on the core is like 10FPS.
CH: Damage control: the longer I post-poned CH, the worse the score got.
LH: Doesn't vary that much, tho, more than DH, so before DH
DH: Didn't vary at all, could run it as last, but prefer to have the xH benches over as quickly as possible

same story for LL and DL. As for CL I needed to re-run a couple of times (4-5) before I got the good one; earlier the night I kept re-running for 25+ times, but never got the good one, so I'm guessing that the first good run is the one you should be holding on to.

Maybe I should've run CH first as sometimes I got a 370+ run instead of 360. But, if I did, the N wouldn't be stable anymore.


Yo Pieter-Jan, I thought ATi was supposed to be stronger than nV in lobbies, what gives? :p:

http://www.hwbot.org/result.do?resultId=897181

Might have something to do with the uber-fast 1400CL6 memory I'm running :p:.

Have you checked with Everest what MEM/L3 bandwidth you're producing? I presume 90k is not that bad, but I have very limited experience with i7 and GX2.

[dinos22]

lol you just opened up a can of whoopass massman :D


..... i'd start hitting F5 soon :D

[EDIT]:cheer::cheer: Go James Go :cheer::cheer:

[youngpro]

http://i200.photobucket.com/albums/aa50/preeeezy/lynnfield_91980_4500mhz_105.jpg

whats wrong with your nature and dragos pieter?

[Foamy]

http://i200.photobucket.com/albums/aa50/preeeezy/lynnfield_91980_4500mhz_105.jpg

whats wrong with your nature and dragos pieter?

9.4 vs 9.6 catalyst? :)

[youngpro]

9.4 vs 9.6 catalyst? :)

might be :rolleyes:

[Foamy]

might be :rolleyes:

:D

(The message you have entered is too short. Please lengthen your message to at least 3 characters.)

[massman]

I'm already impressed that you've got the 9.6's to work decent for Nature, but I do wonder why you cover up the application name in the task bar and in the application itself :D

[Gautam]

Ohhh I wonder why the PCI-E lane is showing up at 8x

Say james can you show us what else is in the drop down list in GPU-Z? :p:

:wasntme:

[massman]

GX2 purchased.