It never ends
You are correct in that the CPU may have a problem working at -269C...however just because the coolant being used can achieve such temperatures, does not mean that the CPU itself will be running at it.Originally Posted by Shintai
Formerly XIP, now just P.
The temp. of the liq. helium was -269C,not the CPU itself...
The problem are the mobo components too.
And just in case anyone has any doubts about this guys credibility, and i'm sure there are some, here's what Michael Schuette has to say about the chap:
Baldy's been a good personal friend for over 10 years even though I have only met him once in person, some 6 years ago. He doesn't work for anybody other for himself (master carpenter and obsessed speed skater in his early 60s who still manages to beat most of the <20 years old competitors). The same mentality that keeps his speedskating going also pushes his overclocking, there is simply nothing that is sacred for him
If there is anybody I am not worried about with respect to objectivity, it is him because he doesn't really care about either side, he's a Baldeagle and known by the tattoo on his calf
M_S
He is still relying on AMD's accuracy that the 30 CPU's themselves weren't cherry picked.
He has no way of knowing that.
great that he made it clear today, but the post in question (yesterday) wasn't so clear after all. Hence all the confusion regarding this here and in the other thread.
@Practicability
AMD and Intel have equipment that cant test far more efficent and faster then anyone could do with hand. (just a example)
AMD and Intel have the means to do it, if they do it is another question.
Say what now?
oooh, exciting.
It's not the normal testing procedure.You can't do it any other way but by hand(LN2 testing).Since it takes a lot of time,like it was posted in previous post by bldegle2,it's highly impractical.They might have done what you suggest for the "normal" chips they showed(air,phase,wc-but we know already most will do ~3.6-4Ghz on air).But for LN2,even testing 5 of them would be impractical(they do have regular jobs there after all@Practicability
AMD and Intel have equipment that cant test far more efficent and faster then anyone could do with hand. (just a example)
AMD and Intel have the means to do it, if they do it is another question.,the ones that have nothing to do with LN2 testing all day long ).
Last edited by informal; 12-07-2008 at 08:50 AM.
Ok Chad, but what exactly are you relying on? At least he's got somebodies word, you've got, well, what have you got? You've got a postition that you can argue right up until these chips are released into the wild. It's a convenient position to have, but it has really started to break. Really, I think there are only a few left who believe these are ultra cherrypicked parts. But keep up the good fight!
I don't need to have anything, least of all blind faith.
I am just saying that this is not a settled issue because of the reason I stated.
Even if this is a cherry picked part, so what?
The biggest issue in PhII's success or otherwise will be pricing, IPC and Clock Speed at stock release.
Yes testing all on LN2 would be stupid, but you just said what was my point. When the 30 cpus where binned better then the regualar stuff (lets say 4ghz at a given voltage of 1.4V)* the chances are a lot higher to get a 6ghz chip form this batch of 30 then form a batch of 1000 of regualr chips.It's not the normal testing procedure.You can't do it any other way but by hand(LN2 testing).Since it takes a lot of time,like it was posted in previous post by bldegle2,it's highly impractical.They might have done what you suggest for the "normal" chips they showed(air,phase,wc-but we know already most will do ~3.6-4Ghz on air).But for LN2,even testing 5 of them would be impractical(they do have regular jobs there after all
That was my point from the beginning. I dont doubt that 5 and even 6ghz chips make its way into the channle and to users that can put that potetntial to use.
From my view point amd suggested with that demo, that you can get 4ghz with pretty much every chip out there. Thats why i want results from retail chips form regular users. If it turns out to be true more power to them, else shame on them.
*note this numbers are all pure fiction
Ouch
Well soon we will all know if their claims are true or not.AMD stated that with good air you can get up to 4Ghz;with wc 4+Ghz.For 5+ghz you would need phase and for the 6+Ghz runs LN2.Of course ,6+Ghz will be harder to reach,but we don't know how much harder.It seems that Denebs are made to withstand the low temps and scale with higher volts and the design was additionally tweaked for this(low temps).If they fail to deliver retail chips that can do what was claimed,they would lose a lot in enthusiast community.Yes testing all on LN2 would be stupid, but you just said what was my point. When the 30 cpus where binned better then the regualar stuff (lets say 4ghz at a given voltage of 1.4V)* the chances are a lot higher to get a 6ghz chip form this batch of 30 then form a batch of 1000 of regualr chips.
That was my point from the beginning. I dont doubt that 5 and even 6ghz chips make its way into the channle and to users that can put that potetntial to use.
From my view point amd suggested with that demo, that you can get 4ghz with pretty much every chip out there. Thats why i want results from retail chips form regular users. If it turns out to be true more power to them, else shame on them.
*note this numbers are all pure fiction
Last edited by informal; 12-07-2008 at 09:17 AM.
Well, precisely, and that's why it's such a convenient arguement to make.
If that were the case we'd be seeing Phenom's sell equaly as well as C2Q as they are absolutely price competive, regardless of IPC and clock speed. The thing that pushed C2Q over the top, was it's overclocking ability. That arguement never held water 2 years ago, so it sure won't now.The biggest issue in PhII's success or otherwise will be pricing, IPC and Clock Speed at stock release.
I think you are overstating Phenom's competiveness, possibly on the basis of gaming benchmarks where being GPU limited made the Phenoms look closer than they really were.
I suspect most people buy a CPU wanting a bit of performance headroom for future games/applications, not just settling for what is good enough right now.
Whilst I disagree that is a reasonable summation of the various reasons for C2Q's success, I suspect that Dual Core C2's have hurt AMD far more than the Quads.The thing that pushed C2Q over the top, was it's overclocking ability.
I don't wanna see this thread pushed to the same place as the others in the past.But for the CPU bound game like GTA4,Phenom looks quite competitive:
http://www.pcgameshardware.com/aid,6...essors/?page=2
("The Phenom X4 9950 easily passes the Q6600 and theoretically is on the same level as the Q6700 (not included in the benchmark)." )
It was shown before that Phenom trails C2Q(65nm) by ~5% on average.The main selling point of the C2Q and Conroe was their great OCing potential.
Now can we get back to PhII topic now?
That is not reflective of reviews that have been out in the previous 12 months(which would have influenced the buying public) and of course C2Q got a nice boost when Penryn came out.
The ~5% calculation was done by Hornet331.You can search for it,don't take my word for it. I didn't say Penryn in my post.But yeah,overall Penryn brought around 5-7% and improved OCing potential.The same things Deneb will bring us,more IPC and great OCing headroom(a breakthrough compared to Phenom I).
Ive yet to see any personal PR 'puppets' for a CPU brand to be honest. At least it's better to act like one then to be one and talk rubbish
What did they lie about AM2? Where did they lie about 4x4? Where did they lie about native quad core?
And how can you call it lying for shipping 45nm later than expected because they show the balls to actually face the B2up and fix that first? I mean, Intel only managed to get a new CPU available in time in like years, so it's not like AMD did a major scam there.
The only thing you could eventually call a lie is how AMD said K10 would perform, but yet that's questionable pretty much. Although they did make it show like it was going to be an epic performer, they didnt say that though.
Donnie, I found something cool today by the way. AFAIK, green blood doesnt excist, but blue blood does excist pretty much in a lot of animals
Well, not working is something else than 'z0mfg 10Ghz+ OC's'. Im not like a specialist in science on such grounds, but I think it's possible to have molecules and electrons move since it's at -273C when they wont move anymore. Although they'll be slow as
Although I dont know how exactly low temps work for such components unlike human/animal tissues, prolly such low temps ain't going to make any CPU really happy by defrosting/warming up certain parts, just like in body's, will warm up faster and might destroy eventual... things
On the other hand I just cant be sure anyhow, -200, although a CPU will be a bit warmer than that, is already very low. How would the slowing molecules/electrons react then? In the end you'd think the lower the temperature the slower they move, but at LN2 you actually get the records thus far, so why would HE be such a problem all of a suddenAs said, I dont know, you've a good point there but where's the border and those aprts in a CPU start to become slower instead of having a potential WR speed?
Synaptic Overflow
CPU:
-Intel Core i7 920 3841A522
--CPU: 4200Mhz| Vcore: +120mV| Uncore: 3200Mhz| VTT: +100mV| Turbo: On| HT: Off
---CPU block: EK Supreme Acetal| Radiator: TCF X-Changer 480mm
Motherboard:
-Foxconn Bloodrage P06
--Blck: 200Mhz| QPI: 3600Mhz
Graphics:
-Sapphire Radeon HD 4870X2
--GPU: 750Mhz| GDDR: 900Mhz
RAM:
-3x 2GB Mushkin XP3-12800
--Mhz: 800Mhz| Vdimm: 1.65V| Timings: 7-8-7-20-1T
Storage:
-3Ware 9650SE-2LP RAID controller
--2x Western Digital 74GB Raptor RAID 0
PSU:
-Enermax Revolution 85+ 1250W
OS:
-Windows Vista Business x64
ORDERED: Sapphire HD 5970 OC
LOOKING FOR: 2x G.Skill Falcon II 128GB SSD, Windows 7
If you are aware that IBM OC is a more of a transistor and not a CPU... totally zero value here... and that was in 2006 or such, so if they continued at that rate why the hell are we still looking for 4GHZ on air ?
Question : Why do some overclockers switch into d*ckmode when money is involved
Remark : They call me Pro Asus
Oh emoWell, not working is something else than 'z0mfg 10Ghz+ OC's'. Im not like a specialist in science on such grounds, but I think it's possible to have molecules and electrons move since it's at -273C when they wont move anymore. Although they'll be slow as
Although I dont know how exactly low temps work for such components unlike human/animal tissues, prolly such low temps ain't going to make any CPU really happy by defrosting/warming up certain parts, just like in body's, will warm up faster and might destroy eventual... things
On the other hand I just cant be sure anyhow, -200, although a CPU will be a bit warmer than that, is already very low. How would the slowing molecules/electrons react then? In the end you'd think the lower the temperature the slower they move, but at LN2 you actually get the records thus far, so why would HE be such a problem all of a sudden
I think you miss the point of kinetic energy. Plus the high voltage that increases it. And then with the highly reduced flexibility at a colder temperature. It wont be slower. It just degrade very very fast as the transitors brake under pressure due to mechanical stress. Its basicly eletromigration and "hot electrons" at an extremely accelerated pace.
And before you say superconductor. They basicly have the completely opposite goal than CPUs in this matter.
Last edited by Shintai; 12-07-2008 at 11:00 AM.
Crunching for Comrades and the Common good of the People.
im not quite familiar with vantage, but how much does P score scale with better cpu?
It should be like this:
The reason for the weighting is simple, there is not just one 3DMark Vantage test, there are four. From bottom to top they are Entry, Performance, High and Extreme. Each has a different weight for the scoring, with the low end giving CPU a higher weight and high end emphasizing the GPU more. If you want the exact numbers, they are below
Kinda complicated,more at the l'inq.
Entry and Perf. settings obviously rely on the CPU score more than the High and the Extreme.
edit:
I just realized it was 3Ghz vs 3Ghz,not stock 9950 vs 940BE.Of this is true,then PhII looks a lot faster in Vantage than same clocked Phenom I,which is good news.I doubt overall score scales perfectly with CPU score(like seen from the link) so CPU score of 940BE might be a bit higher than 16% advantage over 9950@3Ghz we see in overall score.
Last edited by informal; 12-07-2008 at 11:36 AM.
I don't think that's real, for one, 500ghz is ridiculous (simply because not even intel has bothered to try for that when they had a goal of 10ghz stock cpus, it would require way too much power to run, you'd be better off going for efficiency, and for another, just that it is some 490ghz above the world record for the p4 makes it even less plausible when the p4 was designed for ghz), for another, it talks about sony waiting to use the 500ghz cpus for the PS3 which we know for a fact did not happen (the article is from 2006), probably just some joke
500Ghz is easy as such...for a few transistors. But making 1 billion of them go at 500ghz is..a whole other story.
Crunching for Comrades and the Common good of the People.
Posting Permissions
Bookmarks