New 3DMark06 WR with Phenom II X4

[Sampsa]

You can claim about Intel hand pick es cpu to beat all retail cpus, but hasn't claim AMD ES to beat all retail cpus. Why? because AMD already sent you the best ES, but Intel didn't. Or you cannot find any intel good cpu to beat some guys i7 965s? fair enough~

yes, fair enough. I know I can't find as good i7 965 to beat some guys i7 965s. I also know retail Phenom II X4s can't beat our ES.

Show me retail Core i7-965 which can beat +5,265 GHz in Vantage and +5,3 GHz in 06 and prove me wrong.

Hehe, talk about thread detrailing. From temporary AMD 06 WR to my quest for good i7 965 chip

[ghost_recon88]

What the heck is with all the Xtreme 3D Team threads getting derailed by jealous punks so fast First the i7 975 one then this

[AndreYang]

yes, fair enough. I know I can't find as good i7 965 to beat some guys i7 965s. I also know retail Phenom II X4s can't beat our ES.

Show me retail Core i7-965 which can beat +5,265 GHz in Vantage and +5,3 GHz in 06 and prove me wrong.

Hehe, talk about thread detrailing. From temporary AMD 06 WR to my quest for good i7 965 chip

Ok, I will show some retail 965 numbers next week. You let me have energy to bench 3D again. I cannot say my retail 965 can beat all WRs, but I promise they are very close.

I hope you can show some "Good Retail AM3 numbers". I didn't see any "Retail" AM3/AM2 3D results close to your AMD ES so far.

[AndreYang]

What the heck is with all the Xtreme 3D Team threads getting derailed by jealous punks so fast First the i7 975 one then this

AMD and Intel Samples marketing war.

[Oliver]

respect Vince...

beg up

[k|ngp|n]

Black holdown was kickass for Ln2, but alot guys were popin them with acetone MM.
40k Vantage has been done on a retail chip Sampsa. Honestly until that is done....there is nothing really to say there about ES this or that.

Lets not derail again same as fuggers/mike guave's thread, I understand what you mean Andre as it seems like those guys took alot more heat than in this thread, which is basically same thing. This is what you mean right?
All of this whats an "official" and what isn't a wor** reco** stuff is getting old . How do the scores get in the HOF? If they dont let video cards on the orb before launch, I don't get why the same doesn't apply to pre-release cpu's, especially when they are game changers This goes for all hardware. Futuremark is really dissapointing sometimes, but I guess if it's their benchmark and they set the precedent by allowing this stuff on the HOF, that's how it goes.


Titles aside lol, the scores are impressive and it looks like AMD could have a winner here, at least in 2k6. With the low card clocks the score is quite good. I just got my ASUS M4A79 and I can't wait to test it now after seeing your results guys. So does the cpu load the container up in 06' cpu test similar to york?

[SF3D]

F1 is holding the load well. Pot full of LN2 is helping a lot compared to york's. Now there is no limit, so benching is so different in every way.
Our idle temp was -188 an in full load it was -185 in worst case. (Of course multimeter and sensor can be out of the range)

[Sampsa]

40k Vantage has been done on a retail chip Sampsa. Honestly until that is done....there is nothing really to say there about ES this or that.

Lets not derail again same as fuggers/mike guave's thread, I understand what you mean Andre as it seems like those guys took alot more heat than in this thread, which is basically same thing. This is what you mean right?
All of this whats an "official" and what isn't a wor** reco** stuff is getting old . How do the scores get in the HOF? If they dont let video cards on the orb before launch, I don't get why the same doesn't apply to pre-release cpu's, especially when they are game changers This goes for all hardware. Futuremark is really dissapointing sometimes, but I guess if it's their benchmark and they set the precedent by allowing this stuff on the HOF, that's how it goes.

The game has definitely changed. Big companies have entered this scene and there is no way back. There is no overclocking federation or anything to make "official rules" regarding equipment being used like in professional sports to make competition fair for everyone (for example Formula 1 etc).

In case of 3DMark if Futuremark's ORB accepts the score and it is being displayed in Hall of Fame.. the score is legit. No matter if CPU is engineering sample or not even released yet. With unreleased graphics cards the system seems to work better mostly because drivers need to be validated.

Nowadays it is nearly impossible to make WR in latest 3DMarks (05,06&Vantage) with retail chip since CPU is playing such a big role in these benchmarks. You need to have access to engineering samples or test several retail chips to find a good one. I think we all agree on this.

Just my 2 cents. I won't continue on this topic anymore and I'll continue to focus getting higher 3DMark results with all platforms AMD/Intel/ATI/NVIDIA.

[ilkkahy]

So does the cpu load the container up in 06' cpu test similar to york?

Maybe now some of those too cold container prototypes might work nicely when there is no coldbug? I remember you had some round container (in the times of evo 2.3 or something) with very much surface area and you told it responded way too fast to and become too cold.

Just out of curiosity, do you have some unspoken or written deal with intel that you cant do any amd-overclocking? It would be intresting to see what happened.

[Gautam]

F1 is holding the load well. Pot full of LN2 is helping a lot compared to york's. Now there is no limit, so benching is so different in every way.
Our idle temp was -188 an in full load it was -185 in worst case. (Of course multimeter and sensor can be out of the range)

Weird, when I was watching macci benching in Boston I recall the container staying at something like -177 to -178 and never going colder.

illkahy, look at his post again, he just said he's about to test PII...and remind me again who had all the highest scores on the FX57 and the FX60??

[ilkkahy]

Oh, it was am3 motherboard. I kinda assumed it was intel. Will we see the results online too?

[macci]

Weird, when I was watching macci benching in Boston I recall the container staying at something like -177 to -178 and never going colder.

Temp probe was not properly attached

the other system / Pot (Dragon Evo) had a properly installed probe - but I managed to rip it off at some point during our roadshow. so it wasnt there anymore when we arrived to Boston.

edit: usually its about -188...-190C idle and -186...-187C loaded

[massman]

Nowadays it is nearly impossible to make WR in latest 3DMarks (05,06&Vantage) with retail chip since CPU is playing such a big role in these benchmarks. You need to have access to engineering samples or test several retail chips to find a good one. I think we all agree on this.

The impact of engineering samples on the overclocking scene is so big at the moment because both Intel and AMD have been launching new technologies these last few months, not because the scene is evolving to a more professional stage. When the new technology matures and retail hardware becomes as decent as the engineering samples (or at least available to the big public), we'll see that world records will be broken with retail hardware again, much like what we see with the E8600 / DDR3 platform, which is dominated by retail chips and not at all ES samples.

I believe the biggest issue we're currently facing is making a good distinction between new technologies being put their limits to show what this technology is capable of and using non-retail material to claim (even implicitly) world records. As long as there's no clear line between those two ways of using engineering samples, there will be discussions between people regarding the validity of a world record and regarding the political correctness of using ES samples to compete against retail hardware.

Like I mentioned in Charles' thread already: there's no point in making up rules that determine whether a certain score is in fact a valid world record or not, because in the end only the numbers tell the story. 48k is lower than 49k, therefor 49k will be regarded as record, no matter what the author wants it to be. If someone shows an 58k, that's the new record, again regardless of what the author wants it to be. There are just a very limited of possible solutions to this issue and one of them is to blacken out certain parts of the screenshots and thus making sure that people don't know how high the score exactly is. Not making FM validated links is mandatory if you don't want to claim a 3DM world record. There's another decent way to solve this issue, but that's not up for discussion yet ...

OH, just for the record, none of the technologies we've been seeing these last few weeks overclock themselves to these extreme heights; in both Intel and AMD's case, there are people who do the actual overclocking, who put time in testing these technologies. Regardless of your opinion about world records, it's important to acknowledge the work people have put into these sessions.

[Gautam]

Temp probe was not properly attached

the other system / Pot (Dragon Evo) had a properly installed probe - but I managed to rip it off at some point during our roadshow. so it wasnt there anymore when we arrived to Boston.

edit: usually its about -188...-190C idle and -186...-187C loaded

Oh good, looks like I got someone to speak.

[kiwi]

Well said Sampsa Even though it is a rather sad fact

Still nice to see big numbers from AMD

[macci]

Oh good, looks like I got someone to speak.

Hehe, but honestly: the temp probe is not needed with these parts. Easy to tell when the Dragon pot is "in the zone". Everything runs perfectly until you run out of LN2.

see, no temp meter (actually, there is one, the white one, but its for the ambient temp, was +20C):


I was on my way to a higher score but ran out of Liquid. poured in the last drop halfway thru Canyon flight and the pot was empty at the end of that test. Still made it halfway thru the final Game test with empty container.

So I only got this for now (first and final run of the day):

[Gautam]

Hehe, but honestly: the temp probe is not needed with these parts. Easy to tell when the Dragon pot is "in the zone". Everything runs perfectly until you run out of LN2.

see, no temp meter (actually, there is one, the white one, but its for the ambient temp, was +20C):

I was on my way to a higher score but ran out of Liquid. poured in the last drop halfway thru Canyon flight and the pot was empty at the end of that test. Still made it halfway thru the final Game test with empty container.

So I only got this for now (first and final run of the day):


URL = http://service.futuremark.com/compare?3dm06=10023035

Yep, for sure. These containers work awesome for all temperature levels and heat loads.

With you being at the helm of the container I'm sure you'll dock the wr in short order. That canyon flight is really something. O.o

[macci]

yup, Canyon seems to be pretty good, almost 4FPS better than on the 35273 run even though its with lower GFX clocks. the rest of the tests were about the same.

btw, none of the components I used in that run were the same we used at Petri's place yesterday. Different ES CPU, different ASUS mobo (we used Sampsa's mobo, which seemed to have a bit more droop under load), different GFX and DIMMs and PSU. Also, this time the CPU was cooled with Dragon F1 "macci edition". Thermal paste was also white instead of grey

[LardArse]

Go macci!!

[cdawall]

i have 4 ES chips hiding with one of my buds if anyone wants C1 DDR3 chip :roll: i have already seen 5ghz on LN2 with them

[GTSRboy]

Whats faster Dual 295's or 4870 x2 CF?

[stangracin2]

macci, what is that card between the videocards

[macci]

macci, what is that card between the videocards

its a PCI POST card.

[Zeus]

Nice one macci! Good to see DDR3 allows 1t command rate, or is it a bug?

[nickshih]

I didn't see any 06 WR here , wait for backup of backups