5970 testing

Quote:

---

Originally Posted by ..::G80::..

With MSI afterburner I overclocked my 5970 @ 1000MHz core and mem 1200MHz with no issue. ;) On air.
But I want to go over CCC limits. :D:D

BTW nice scores Macci & SF3D :up:

---

Hehe! You did it to me G80 :up:

Well yes, we are talking about 1200MHz and -100C° numbers here.

Quote:

---

Originally Posted by SF3D

Yes, there is some issues with cards under LN2. Card is confused about 2D / 3D voltages and it is doing thermal throttling at the same time. So, it need to be fixed with new bios again.

Now, I need some people to tell me, that they can clock their cards without any issues :rofl:

---

sorry i meant to say CPU on LN2 (AMD platform)
cards on air

Quote:

---

Originally Posted by SF3D

Hehe! You did it to me G80 :up:

Well yes, we are talking about 1200MHz and -100C° numbers here.

---

Holy shiiit! :D:D
How do you past CCC limits? :)
1200MHz core is awesome. :up:

nice scores man

Are you useing moded bios?

Quote:

---

Originally Posted by ..::G80::..

Holy shiiit! :D:D
How do you past CCC limits? :)
1200MHz core is awesome. :up:

---

With AMD GPU clock tool i can oc my 4870 beyond the CCC limit. ;)

Quote:

---

Originally Posted by Zeus

With AMD GPU clock tool i can oc my 4870 beyond the CCC limit. ;)

---

I just tried AMD GPU tool, I can olny overclock one core and 1025MHz works...
SF3D show some nice scores. :up:

Love to see that score are come from Dragon power :up:

Is 35k `06 is a lot for this setup and monster 5970 ?
I can`t understand one thing: My E8400 @ 500x9 and 3870 512Mb did 14000 at`06.
Why 5970 can not do 50k ?

Quote:

---

Originally Posted by ..::G80::..

Are you useing moded bios?

---

He probably is. :yepp:
Come on Petri, show us some krazeh numbers :D :up:

Things will come out, when time is ready..

I tested some Vantage X preset today. Got some interesting results in dual card CFX configuration. CPU speed will give quite a lot of boost already, so this last real 3D test is not like that anymore. Too bad..

Single 5970 card in X preset.
http://www.aijaa.com/img/b/00016/5383209.jpg

4GHz vs. 5.4GHz
http://www.aijaa.com/img/b/00714/5383218.jpg
http://www.aijaa.com/img/b/00886/5383222.jpg

I had just a few liters of LN2 left, so I did not have a chance to test 6GHz+ speeds.

GT2's still GPU-bound, though :-)

Not bad huh Petri?

these Hemlocks seems to kick quite strong along, and 1200Mhz GPU :up:

ORB damage incomming :D

Nice going S... :up:

Quote:

---

Originally Posted by massman

GT2's still GPU-bound, though :-)

---

GT2 is NOT GPU bound....With high CPU you gain a LOT out of it..... ;)

Sorry, George, I was confused by the 0.88FPS increase going from 4 to 5.4G. :p::D

I suppose you're talking about GT2 in P preset?

Quote:

---

Originally Posted by massman

Sorry, George, I was confused by the 0.88FPS increase going from 4 to 5.4G. :p::D

I suppose you're talking about GT2 in P preset?

---

Yes.... :)

finally got my home OC rig to run at the proper speeds. for the longest time I had issues getting even 6G 3D stable. turns out that the issue was the suboptimal thermal contact with LN2 pot. recently I have been mounting the pot without springs. bad idea. :D (thx ryba for solving this!)

Used a fresh 965BE C3 and tried something new to improve the cooling (in addition to utilizing the springs). see the last photo. I haven't tried it with "the usual stuff" so hard to tell if its any better / worse this way. We'll figure it out during the weekend.

ok, the following rig:

965BE C3 week 47
Crosshair III
GSkill Perfect Storm (thx hiwa!)
one and two 5970s (crappy pre-production cards)
SF3D LHe / LN2 pot (produced by ryba)
single 1200W Thermaltake PSU

http://img.techpowerup.org/091214/rig_s.jpg
(the temps are not accurate - sensor 1 is mounted to CPU heatspreader, its a K-type sensor but at LN2 temps the reading is closer to a T-type sensor, its about 5-8C too high, sensor 2 is on the base of the pot, its a proper K-type sensor but now reads too low as the meter is set to read T-type)

and following scores:

single card 05:
http://img.techpowerup.org/091214/48256.png
CompareURL

single card 06:
http://img.techpowerup.org/091214/36263.png
CompareURL

dual card 05:
http://img.techpowerup.org/091214/49720.png
CompareURL

dual card 06:
http://img.techpowerup.org/091214/38052.png
CompareURL

everest showing the bandwidth:
http://img.techpowerup.org/091214/Ev_6384.png

and pot base:
http://img.techpowerup.org/091214/pot_base_s.jpg
very easy to mount and unmount. zero mess with paste. optimal? no idea but so far so good :D

Very nice. :up: Canyon flight seems especially strong in both 05/06 with this GPU configuration. :)

impressive clocking :)
what exactly did you use there at the base?

Awesome! Almost 4.8 CPU-NB is unreal, that sure boosts the scores.

Is that pure silver at the base of the pot?

Looks to be some sort of thermal pad but not thick like the normal ones. Would like to hear more info on these. Also that maybe the sickest debug card I've ever seen :)

Thanks :)

its a sheet of indium

05 scores would improve slightly with GBT mobo (especially GT1), need to try that during the weekend as well.

What makes the GBT board to have a slight advantage? Better mem-nb subsystem or pcie lay-out improvement?

siiick 3d speeds and sick nb speed :O what voltage was it running at?

Quote:

---

Originally Posted by macci

Thanks :)

its a sheet of indium

05 scores would improve slightly with GBT mobo (especially GT1), need to try that during the weekend as well.

---

This is a bit confusing to me. Indium's thermal conductivity is rather poor :confused:. Could it be that you were using too much paste and not enough pressure on your IHS pehaps?

http://environmentalchemistry.com/yo...c/thermal.html