Any idea how these might perform on WCG?
Printable View
Any idea how these might perform on WCG?
whats that @ 6+ mean? the ghz?Quote:
Originally Posted by chew*
i think that would be a special occasion since the cold air from LN2 might be good to keep the board cold too. im thinking water cooling fans might have to buy water blocks for the PMW, were talking about 40% more power draw from having 50% more cores.
im just saying we might see a thread in the AMD section listing what the limits are for boards. i already saw one set fire myself, lol
Quote:
Originally Posted by Manicdan
I believe any boards supporting 140W TDP has no problem allow Phenom II x6 to run. The problem is that IF motherboard manufacturer going to update the BIOS to support the chip or not.
i know that any 140W mobo can handle a 125W hexi with just a bios update, but its how much power do these things draw when OCed. if a quad hits 180W at 1.5v 4ghz, then how much do u think a 1.5v 5ghz hexi will draw? and how many boards were built to handle such cases? id bet air coolers will not see this as an issue since the power draw turns to heat which turns to temperature which turns to instability, so that leaves water cooling, and if theres bad airflow on the PWM heatsink, id be more than willing to bet alot of people are going to be seeing issues.Quote:
Originally Posted by haylui
but it seems everyone thinks im just being paranoid or dont understand the question.
source; http://www.xbitlabs.com/news/cpu/dis...rocessors.htmlQuote:
More details have also emerged about AMD’s Turbo Core technology. Depending on the actual model and its specifications, AMD Phenom II X6 1000T-series chips will be able to boost their clock-speeds by 400MHz or 500MHz when only half of available cores are active, e.g., microprocessors work in triple-core mode. That said, the model Phenom II X6 1090T at 3.60GHz will most likely outperform the currently available Phenom II X4 965 Black Edition in typical cases.
i thought it was going to be only a single core that would be turboed, i think this is great news to hear it will be 3 cores not 1. AMD has really matured 45nm.
AFAIK it works in four core mode too(4 active cores/threads). It's one bin over stock in that case though.Quote:
Originally Posted by Monkeywoman
From what I've read, they switched the 1090 at 3ghz for the 1095 3.2 ghz... So the 1090 will still be out, but in Q3 instead of Q2. And the 1095 will be out at Q2... anyone can confirm?
It's been posted and discussed before,please read the thread...
Quote:
Phenom II X6 1090T : Thuban X6 9MB 125W 3.20GHz/3.60GHz AM3 Q2 2010
Both work at 3.2 GHz. It's not clear yet what's the difference.Quote:
Originally Posted by vitchilo
1095 will have a little beat more aggressive turbo mode.Quote:
Originally Posted by dess
I see the problem, but its easy to solve, watercool the pwms:yepp:Quote:
Originally Posted by Manicdan
1095T has been mentioned through this thread but doesn't appear on the official road map...http://forums.vr-zone.com/photopost/..._Roadmap_1.jpg
Fitting 6 cores under 125W is pretty darn impressive job if we consider the 3.2/T3.6Ghz and 45nm SOI process!
Now these will surely go under LN2 sooner or later,there should be no cold bug as with the deneb cores.
I want 4ghz 24/7 w/am2+ mobo.:D
I forgot why there was 125w and 95w 1055T coming out at the same time?
What was the diff again?
That should be an equivalent to standard(125W) and HE(highly efficient,95W?) Opteron models.Quote:
Originally Posted by ajaidev
Exactly. Few weeks ago I said that a 3GHz/140TDP modell could be the top until Q4. So this 3.2GHz/125W at the start is a surprise. GF has made the home work. It was very important for the AMD beacuse same dies will be under the Opteron 6100 series CPUs cap.Quote:
Originally Posted by informal
Anyway I'm expecting at least ~6.5GHz bench stable results under LN2.:D
Indeed,those Magny C. should do great in server arena ;). Thuban/Zosma with current starting power bands and clocks are giving AMD a chance to even fight in the upper perf. desktop segment (just below Gulftown). And with the aggressive pricing we have seen and socket compatibily these cores are destined to succeed!Quote:
Originally Posted by Oliverda
motherboard with less than 4+1 phase power supply circuit on board are at the lower end, thus this is not recommended for overclocking. While those overclocking friendly board would have large piece of heatsink on PWM units. So with good casing air flow I guess this is not a problem.Quote:
Originally Posted by Manicdan
30WQuote:
Originally Posted by ajaidev
SCNR ;)
Well, as informal said, very likely simply two models with different TDP at different price points.
Hm, and 1Q 2011 Bulldozer :-D. Jupiiiii :).
PS: yes, TDP 125W for 3.2 GHz hexacore, 45nm and SOI process is brliliant !
I remember people *cough* claiming that the six-cores would all run <2.6Ghz :rofl: (the usual intel suspects btw)
As for the 1055T 95W vs 125W, it's probably that:
@ Base clock with full usage power load is <95W, but 125W gives more possibilities for turbo (if they are calculating by TDP/Power usage).
Pricing difference shouldn't be great, it's just that you potentially lose single threaded perf with the 95W version.
But the power consumption in stock clock and turbo clock modes (P-0 and P-Max) is strongly related (as I've calculated on my blog lately). So if the 1055T is supposed to run 6 cores at 2.8GHz at <95W, then it will do that also with 3 cores at 3.3GHz and the others idle.Quote:
Originally Posted by Macadamia
We all know them ,we just don't mention their names now(it would be even more embarrassing :D ).Quote:
Originally Posted by Macadamia
125W vs 95W looks like binning job to me :)
think there will be a nice way to modify the pstates so we can set our own turbos? im a fan of overclocking, but also power saving, and its been a real pain to keep CnQ on while OCing.
now i know, what is x6 1095T. Think, this point is at turbo schema..
1035T +500Mhz
1055T +500Mhz
1075T +500MHz
1090T +400MHz
think, 1095T will with 500MHz with turbo
i support this theory, sounds logical to meQuote:
Originally Posted by FlanK3r
That's ridiculous. You mean AMD is actually going to rate processors based on the "gimmick" called Core (Turbo) Boosting? :p
It also looks like SFF or HTPC sex-core CPU :p:Quote:
Originally Posted by informal
@FlanK3r
Just look at my post several pages ago ;)
http://www.xtremesystems.org/forums/...8&postcount=50
Quote:
I think is has something to do with 'only' 400MHz turbo compared to 500MHz for 1035-1055-1075 models maybe?Quote:
Originally Posted by ajaidev View Post
I dont understand the 1090T codename??
1035-1055-1075 and then 1090 ?? why shouldn't it have been 1095T
Just a guess ...
@Ohnoes!
For what gimmick are you blabbering? That Core (Turbo) Boosting will boost performance of their processors (running at specs). Thanks to that feature Nehalem is so better than Phenom II and Core2 Quad at "same" clock.
Quote:
Originally Posted by OhNoes!
yes, and as u watch, im not alone (im not noob at CPU AMD man -some years ago interesting for it ;) )
Lightman
Heh, thanks, now we are 2 mans here with the same speculation :up:
Since the irony is lost on you guys, allow me to clarify. It's quite funny how some condemned Turbo Boost as an Intel "gimmick" designed to win benchmarks, but when AMD implemented it, everybody is drooling all over it and are actually going to cough up for cash to settle for the higher turbo on the 1095T. This is all assuming that the ONLY difference between the 1090T and the 1095T is the 100mhz extra boost when the feature kicks, since they're both supposed to be clocked at 3.2GHZ iirc?
I like the price of these processors a lot, but if the above is true, then AMD is setting a new precedent here (cough, cough) - I don't want Intel getting any ideas :p.
ill explain. a gimmick is something used to look good on benchmarks without proper representation of actual frequency and is misleading. we are happy that amd is pushing cores when some are idling, which has NOTHING do to with benchmarks. if its going to be 10$ more, then it might sell. but also keep in mind that people will probably not have a clue thats the case since sites like newegg or microcenter, probably dont list the turbo speeds, just stock speeds and model numbers.Quote:
Originally Posted by OhNoes!
and like u just said, were guessing, its fun to do and its not going to cause any harm, lol
OhNoes! When I red your post I thought you were serious. But now, i saw the smiley :p at the end of the post, which should have been :p:.
I agree with you on that point. AMD fanboys will BS about every Intel feature which is not present in AMD CPUs. But once AMD implements it, they will start talking about it like it's the best invention ever implemented.
Hey, I have a question (sorry if it is a stupid one): from where do you know that a '1095T' is coming? I'm asking because there is no '1095T' in the leaked AMD roadmap. Anyway, I'm really impressed with the 1090T specs: 6 cores at 3.2 GHz @125W.:)Quote:
Originally Posted by FlanK3r
http://www.pcgameshardware.de/screen..._Roadmap_4.jpg
It was first mention here.Quote:
Originally Posted by Karolis
Source
Quote:
AMD will launch several six-core Phenom II processors in the second quarter including X6 1035T (2.6GHz/95W), X6 1055T (2.8GHz/95W) and X6 1075T (3.0GHz/125W), according to the sources. In the third quarter, AMD will add a few more six-core processors including X6 1095T.
I wanted to ask the same...Quote:
Originally Posted by Karolis
(edit)
I would think the 'source' mistook 1095T for 1090T, or AMD changed it's mind in the meantime (I mean, pehaps 3.7 GHz were too much of a hassle already).Quote:
Originally Posted by msgclb
Some AMD fanboys, maybe. Just like some Intel fanboys, about the IMC, then x64 (originally named AMD64)... Anyway, I, for one, as an AMD user since a decade, never thought the Intel Turbo Boost was a bad idea. Indeed, that - for your information - the Barcelona/Agena already has the ability to clock the cores differently, and there was a little application (AMD Overdrive) with which one could set the clocks as liked. There is perhaps even a utility that can do it automatically. It's just not done in HW, as with the new CPU's.Quote:
Originally Posted by gOJDO
AMD did not invent both of those technologies either, though. AMD was the first to bring those technologies to the desktop, however.Quote:
Originally Posted by dess
Having said that, I like where AMD is going with their market/price targeting. It should be reflected in their market share in the next 12 months.
I know, regarding the IMC.Quote:
Originally Posted by OhNoes!
64 bit, as a way to handle values, is not their idea/invention, of course. But, x64/AMD64, as a specific and substantial extension of the x86 architecture and ISA, is AMD's own invention.
oh noes....I am afraid that AMD fans will BS about IMC and the ditch of FSB since K8 era and native quad core design as Intel has follow suit.Quote:
Originally Posted by gOJDO
I'm pretty aware of AMD Overdrive options. But the point is that AMD Overdrive is not for average Joe and does not come installed with the OS by default. Also, using AMD overdrive for OC-ing, practically voids your warranty. Unlike Phenom and Phenom II C2/C3, Thuban will have that feature managed by the CPU and it won't void warranty.Quote:
Originally Posted by dess
BTW, AMD Overdrive and "Turbo" "Boost" are for noobz and are useless for me:
http://valid.canardpc.com/cache/banner/827650.png
http://valid.canardpc.com/cache/banner/1035783.png
;)
Please read:Quote:
Originally Posted by gOJDO
http://www.gamers-crib.com/forum/sho...1&postcount=15
I hate turbo. It's retarded and it's just made for people who can't overclock.
If AMD can't run all 6 cores at a given speed it's a flop. I don't care how anyone explains it. This "turbo boost" tech from both companies is stupid. The average user will not even notice the speed increase, and whoever is doing something that will take advantage of it AND NOTICE will probably be technically inclined to the point in which they already know how to overclock the damn thing.
3.2 Ghz 125w Black Edition 6 core for $250 without turbo would be good.
overclockers dont need it, but the rest of the world would love a 20% increase without having to do anythingQuote:
Originally Posted by BeepBeep2
But in multi-threaded apps it doesn't do anything.Quote:
Originally Posted by Manicdan
What kind of normal user needs a 20% increase in a single threaded app? Encoding is usually multi-threaded, and gaming will probably stress the CPU to where turbo doesn't even "engage" on the single core...
Actually, yes it does.... +1 bin for 3 or 4 cores engaged, +2 for 1-2 cores engaged (with respect to Intel cores). The Lynnfield offer even more in the single and dual threaded scenarios. AMD's first implementation will run nominal with all cores engaged, and bump with 1-3 cores engaged. Either solution is absolutely a step in the right direction.Quote:
Originally Posted by BeepBeep2
The move to multicore is simply a different way to take advantage of Moore's law, the physics, however, limits the clock speed not by Fmax of the design/process, but the practical limits of power and cooling. To fit a quad core, for example, within a reasonable thermal envelop, the processor is volted and clocked lower than what would otherwise be capable.
This is fine, since performance is extracted via thread level parallelism, in fact efficiency can dramatically increase if the task can be well threaded. However, Amdahl's law kicks in or some algorithms just cannot be made parallel at the thread level, as such these situations suffer with respect to the potential because any one core is clocked for the thermals.
Essentially, which would you rather have for a single/dual threaded app, a 3.4 GHz dual core or a 3.0 GHz quad core? Naturally, the best situation would be the higher clock.
So designers face a challenge, how to bring more and more cores to the masses but still yield performance on legacy single threaded applications or provide support until software can catch up.... the very natural thought is, if all but one or two cores are idle, why not just kick up the clock speed of the active cores... hence turbo modes.
It is a great idea, works as it should and gives you the best of both worlds, great performance in highly threaded environments, and a boost in performance over what could be had in lightly threaded environments.
If only the windows task scheduler was smart enough to use that potential... :p:Quote:
Originally Posted by JumpingJack
Ahhhhhh, I am ahead of you there ..... I have looked into this some, actually ...Quote:
Originally Posted by Hornet331
Here is the upshot...
a) Vista -- dump it.
b) Win 7 -- get it.
It is very clear that Win 7 scheduler is doing a much better job, not perfect, but much better. Haven't looked at XP.
Say you buy a 1090T, that means you're pretty much guaranteed a 3.6GHz overclock if you know what you're doing right ?
Yeah, without question. I am particularly holding back an AM3 board for this processor, I have a X4 965 (140W unfortunately) and wall at 3.8 Ghz in a 64 bit OS, but around 4.1 GHz in a 32 bit OS. I haven't played with the newer stepping to see if this wall still exists (and I have really paid much attention to the forums to see what people are getting), but given that my (and everyone else's) X4 965 can reach 3.8 easily (at a minimum), then it is a good assumption to expect that the X6's are not Fmax bound at the very minimum, up to that point.Quote:
Originally Posted by Frontl1ne
Jack
According to some slides, it elevates half of the cores' clocks (while putting the rest to sleep), so 3 of them, in case of Thuban. I don't know about an application that runs only 2 or 3 heavyweight threads, and never more, though.Quote:
Originally Posted by BeepBeep2
20% is not much for an overclocker, but for the rest of people, 20% increase in the clock frequency has its costs in money, until now (considering single or "lightly" threaded apps). And in CPU-bound situations in games, it can bring even that much of increase in fps, which is "a lot"...Quote:
What kind of normal user needs a 20% increase in a single threaded app? Encoding is usually multi-threaded, and gaming will probably stress the CPU to where turbo doesn't even "engage" on the single core...
They can't clock all the cores high, because they have to take care of the power-consumption, too - which is definitely a thing to consider, for many people. Otherways they would have to sell these CPU's with a TDP of one or two classes higher (125W -> 140W, 95W -> 125W).
Or, they could leave the clocks in the lows untouched, but then they could only sell them for rendering, videoediting, etc. and to overclockers.
I know but they could easily fit a 3.4 Ghz or even 3.6 Ghz native 6-core in if they raised the TDP envelope to 140w. 3.2 @ 125 is just like the 955 and 965.Quote:
Originally Posted by dess
would it be possible to disable the turbo feature in the bios? would it then run at 3.6 all 6 cores? Also temperatures wouldn't be that bad, a Noctua NH-D14 would work nicely.Quote:
Originally Posted by BeepBeep2
That's what I would call OC so yes. I think it's a good feature, it boosts where possible. Also, I believe the reason for pro-AMD people to have been talking down the turbo boost on Intel CPUs is because it gives the impression that their CPUs perform much better per clock than an AMD CPU even though the difference isn't that big.Quote:
Originally Posted by Monkeywoman
It is likely you can disable the feature -- it won't likely work unless C&Q is enabled anyway. However, it will default to the base clock, 3.6 would likely exceed the thermals. Then, of course, you could just over clock it to 3.6 anyway (black editions are unlocked), which is what most enthusiast will do anyway.Quote:
Originally Posted by Monkeywoman
Turbo for an OCer is almost worthless ... most just turn it off and crank all up as high as they can get it.
its not right Beep, watch at example i5 750...This 2.66 Quad working normally at 2.8 GHz (4 threads activ) or more (for single, dual operations), if u have turbo enabled. For common users is turbo good point!Quote:
Originally Posted by BeepBeep2
http://www.tomshardware.com/news/PAX...-X6,10004.htmlQuote:
AMD Brings Phenom II X6, Eyefinity6 to PAX
9:10 PM - March 27, 2010 by Devin Connors - source: Tom's Hardware US
Intel got its six core CPU to market first, but AMD's Phenom II X6 offerings are just around the corner.
While Nvidia might be stealing the show so far at PAX East, AMD is here with some new hardware of their own. While a lot of what was revealed is still under embargo, here's some clarification on what's coming in the next several months.
http://media.bestofmicro.com/,6-F-242583-3.jpg
http://media.bestofmicro.com/,6-D-242581-3.jpg
http://media.bestofmicro.com/,6-E-242582-3.jpg
:)
http://media.bestofmicro.com/,6-F-242583-3.jpg
^
one of the reasons why you wont see much turbo action in most apps... :p:
The task manager shows around 50% CPU utilization or 3 cores,which lines up with AMD's Turbo Core specification that states the Turbo kicks in to 3.6Ghz if <=half the cores are used,meaning 3 cores or 50% of X6 :). You get 3.6Ghz for 3 or less cores used,mind you.And you get a very good 3.2Ghz for 4,5 or 6 cores used which guarantees a big speed up over X4 in multithread and/or multitask scenarios.Quote:
Originally Posted by Hornet331
Yes if only 3 cores would be used, but windows shuffels the threads between the different cores. so basically it utilizes all 6 cores, and not just 3.Quote:
Originally Posted by informal
Its the same for intels turbo, thats why you hardly see any situation where you get more then +1x multi.
The only time when i see that happen is, when i assign the task manually to a physical cpu and there isn't any other app running, that utilizes another core more then 1-2%.
I heard that win7 handels this a bit better, but haven't had a chance to test this, since im still on vista.
In the end we'll have to wait and see how it is done, but it doubt it much different from how nehalem handels it aka how windows handels it. Meaning you get a nice standard boost in speed but don't count on seeing the highest multi under normal user behaviour.
According to news
die size of here :
phenom II x6 around ~ 356 mm2 (??)
phenom II x4 around ~ 258 mm2
athlon x4 around ~ 169 mm2
athlon II x2 around 129 mm2
different size of athlon II x2 and x4 is (169-129) = 30 mm2
now it's possibly to make athlon x8 for die 356 mm2 (if we insert two die of athlon x4 in one packet ,, 169 * 2 = 338 mm2 )
http://www.anandtech.com/cpuchipsets...oc.aspx?i=3638
so athlon x6 should be able to handle 90% of performance a Phenom II X6.Quote:
At the same clock speed the Athlon II X4 should offer roughly 90% of the performance of a Phenom II X4.
Hope that Athlon octa core should defeat core i7 !!!
lol, why? 2H 2010 are samples Bulldozer and 1Q2011 maybe a launch. Need not K10.5 8-cores...
Without cache coherency 4-> 6 cores would be rather useless, starting to diminish. You need a bit of cache at least. Even Llano's staying at 4 cores.Quote:
Originally Posted by behrouz
This small difference in die area is because AII X2 has 1MB L2 per core, and AII X4/X3 has 512kB per core.Quote:
Originally Posted by behrouz
according to w0mbat, he could access to the parts that resemble BullDozerQuote:
Originally Posted by FlanK3r
source: http://www.xtremesystems.org/forums/...&postcount=333
yes BullDozer will come out in 2011 , but i don't want to see the dead of AM3.
You won't, bulldozer will be AM3 compatible.Quote:
Originally Posted by behrouz
No one is sure about that its rumored that BD will use AM3 r2 it could be possible that its compatible with AM3 or maybe its not who knows.Quote:
Originally Posted by RaV[666]
Hopefully it is tough :D
I read few reports about it being fully compatible.And even JF wrote that bulldozer server parts would drop in to c32/g34 infrastructure.So they have to be electrically comaptible with phenoms.Quote:
Originally Posted by ajaidev
And then there are multiple AMD roadmaps as this:
http://icrontic.com/uploads/2009/11/...op_roadmap.png
We can be pretty sure that AM3 is bulldozer ready ;-).
X6 listed at giezhals with stepping E0 (Istanbul is D0), no prices yet.
Here's the 960T, showing stepping C2 which must be wrong.
Asus users, here you go
http://i39.tinypic.com/261oq2s.jpg
Gigabyte users:
http://i39.tinypic.com/30lfh2b.png
Hopefully it will be AM+ compatible so i dont need to upgrade mobo. :DQuote:
Originally Posted by ajaidev
:D
The problem is that of update frequency and timing, if the windows scheduler switches context over the 6 cores faster than the turbo update can occur, then the turbo algorithm will assume 6 cores are used.Quote:
Originally Posted by informal
To put it another way, turbo mode does not force cores idle, it watches for idle cores. This is based on the power state of the cores as specified by the OS, if the OS does not let cores or enough cores go into the lower power states, turbo mode does not engage.
Thread scheduling is a component of how well or effective turbo mode will operate as well as the polling/update rate for the method employed. If you recall, Penyrn (mobile) implemented a rudimentary form turbo mode which simply defined new power states for up clocking the cores, the problem was it did not turbo very often and this was because Windows kept throwing out spurious threads that kept the cores awake. There was a big whoopla about this during the Win 7 roll out as I recall, I have done some simple testing myself and frankly Win 7 is the OS to run with this CPU, Vista will perform worse.
Now, having said that, there is no doubt that the X6 is going to be a nice leap over the X4 965 and will be a serious CPU in consideration for anyone building a rig.
i couldnt agree more with that.and considering the price range of these cpu's,they will make for a great centerpiece to build a very powerful system.it will be my next setup,thats for sure,since i cant afford an intel x6 setup:up:Quote:
Now, having said that, there is no doubt that the X6 is going to be a nice leap over the X4 965 and will be a serious CPU in consideration for anyone building a rig.
yep, I am holding open an AM3 board waiting for this processor so I can play with it.Quote:
Originally Posted by radaja
So when will we be able to buy it at Newegg? First week of May? Because I'm trying to find the Opteron 6128 and impossible to find it anywhere.
im waiting on funds right now,but as soon as i have the cash "i'll be having some 6 core fun" :DQuote:
Originally Posted by JumpingJack
Still no AM3 boards with SLI support, is there?
It has been here for ages.Quote:
Originally Posted by Anemic
http://www.newegg.com/Product/Produc...82E16813131383
nForce 980a (pretty much a 780a).
Ahh, yeah. Forgot to mention DDR3 support =( But that is an option, going to look a bit deeper into it.
I like the look of this board! Asus Crosshair IV Extreme with Lucid Hydra. To bad the hydra chip seems to suck...
http://www.slashgear.com/asus-rog-cr...tured-0784711/