nice to see full AMD system and that score :)
Printable View
so guys, you gained +170mhz with liquid He over ln2..hııımmm :)
very cool that the 3dmark 05 WR is on Finnish possession WOOHOO you guys made us Finnish clockers proud..:toast:
and awesome to see amd clocking like mad. competition is always good
So hows the LHe benching experience? Does it feel like LN2 or is it more dangerous or volatile in any way? Does it need to be poured way more often?
Just surreal to see an AMD of all things doing this. It all sounded like a bad rumor. :p:
And congratulations to you guys again.
Liquid helium is unbelievable cold as you can imagine. I saw it only once in liquid form after one and half our benching. That really small drop was in the bottom of F1.
LHe is very easy to use with AMD stuff. Just put the hose in to pot and let it run. LHe evaporates to gas immediately, so you only see huge cloud of steam around system. We had to wait 5 minutes to get CPU pot as cold as possible and then we just started benching.
It was crazy to install drivers in front of people and run the system around -230 to -240c :rofl: I have no words...
AMD sure does have some tricks on their sleeves.. future is looking promising on their side too. This is very good for us.. consumers are the winners.
You are right. 500 liters of liquid helium might be fatal for example in some building. If someone is going to try this, remember to have a really good ventilation.
We were in the desert and in the ventilated tent, so we didn't have funny voices or any other symptoms. Except both of us got some cramps to our legs. I think we got those from standing three days in a row.
Oppainter put it on words: "these f***ing Finns are f***ing crazy.. they stand and bench, stand and eat, stand and drink.. then they sleep two f***ing hours and stand the whole f****ing day again" :rofl:
you guys use ln2 to max out the pot before hitting it with LH right?
Id hope so as if you didnt must take a year to get the pot to temp with LH alone as the LH would evap before it touched the copper as LH has no thermal heat removal propertys hence reaseon LH is always used as a insulator to either another LH barier or LN2 in cryrogenic cooling machines :)
awsome results tho, but im sure you where pouring LH into the pot at a constant rate even at -200c right?
Man this sounds absolutely incredible. And congrats on the ORB WR for 05. Great stuff.
Man I can't wait for my 940 to arrive. Straight from the States to South Africa. Got some contacts...he he he
sooo did your voices go all thin with the LHe?!
VPII will u have it by 23Jan for the OC event? :D?
un-f***ing-believable huh :D congrats guys, amazing milestone in overclocking, benching at -240, and running 3d at 6.3G on AMD... crazyness :up:
nice results
shame you didnt have cards on subzero to take the W.R. but live event etc its just too much pressure i'm sure/too many variables :up::up::up:
lol at comment OPP made :rofl:
i heard that Liquid Helium costs $300/litre :shocked:
I am absolutely amazed...incredible job guys...keep it up~!
No.. liquid helium is not that expensive in US. In Finland it is around 10-20€/l
If we would have used some hotfix drivers or beta 9.1, the score would have been a lot better. We had that multimonitor setup + some other weird bug while running in public. On Friday we got 45261 score with 6130MHz CPU speed, so if we would have that efficiency on Saturday, score would have been a lot higher :shrug:
Are you guys gonna give it another shot with LHe?
well they did 05 at 6.3ghz so i think so :D
Here is the highest CPU-Z validation we got:
http://valid.canardpc.com/show_oc.php?id=483738
We had +6,5 GHz on screen and CPU-Z but unfortunately validation didn't seem to be saved :(
sweet validation Sampsa and wicked with 6.5g+ thats good to hear
One thing I want to mention about this screenshot: the Memory tab of CPUZ is reading slightly incorrect clocks. the CPU NB was really at 3920MHz and MEmory clocks at 560MHz (280MHz HT ref. clock with x14 CPU NB Multiplier and 1:2 Mem divider).
when the screens were first loaded there was fluctuation w/ the clock readings (one of the CPUZ screens was reading 274HT ref. clock and one was reading the correct 280MHz clock). that screenshot is taken in a hurry and the readings did not stabilize. Thats why the Mem / CPU NB clock reading is slightly off.
a photo was posted in another thread that shows the correct readings:
http://i23.photobucket.com/albums/b3...s/vegas187.jpg
6301MHz CPU / 3920MHz CPU NB / 560MHz RAM
Totally sick NB clocks. I'm sure that really helped the clock/clock efficiency.
what made you guys use the dfi instead of the gigabyte/referance boards you guys used at the preview events?
There was no particular reason for using DFI, but we got our NB clocks and 3D performance up with that mobo. It is a good choice, if you plan to go with AMD.
I found 6.5GHz picture:
Attachment 92970
What a shame, that validation was broken..
DFI and AMD go to together like apple pie and vanilla ice cream. Mmmhhmmmmm...delicious. :D
Actually we had couple board failures with DFI.. also lost one good CPU because in case of mosfet explosion :(
Few questions though. Why were you using ES chip for this OC attempt? Was this chip newer than retail available now?
Any leak how RB-C3 respond to cold Macci?
Most important: is AMD planning similar events in the future. Will they be hosted on other continents as well?
It's amazing to see 6.5GHz on a monolithic 750+ million transistors chip. You should make an entry to Guinness book with that :D