AM3 seems to be DDR3 only.Quote:
Originally Posted by VulgarHandle
Also its extremely unlikely for any AM2+ and AM3 mix. Since AM2+ got parallel VIDs and AM3 serial VIDs. So unless they drop serial VIDs it will basicly be impossible.
Printable View
AM3 seems to be DDR3 only.Quote:
Originally Posted by VulgarHandle
Also its extremely unlikely for any AM2+ and AM3 mix. Since AM2+ got parallel VIDs and AM3 serial VIDs. So unless they drop serial VIDs it will basicly be impossible.
I think it was not exactly:Quote:
Originally Posted by Shintai
Board AM3 supports DDR3
Boards AM2, AM2+ support DDR2.
Processors AM2, AM2+ have DDR2 controller
Processor AM3 has BOTH DDR2 & DDR3 controllers.
So, there is picture explaining everything:
Is there actually any confirmed AM3 information for DDR2+DDR3 since jedec made sure DDR2 and DDR3 aint compatible for the same pins?Quote:
Originally Posted by SEA
Sofar the AM3+DDR2 seems to be a old old rumour.
By the time we get AM3, DDR3 will be quite comon. I see a 939->AM2 situation again.
...you forgot that AM2+ combined with K10 will support channel interleaving. Two indepenedent memory controllers will allow for better command efficiency, lower loaded latency, and concurrent reads and writes. See post # 99.Quote:
Originally Posted by VulgarHandle
Just because the memory slots aren't pin compatible, doesn't mean the IMC cannot support both architectures. The lack of pin compatibility only means the supported memory standard would be determined by the motherboard.Quote:
Originally Posted by Shintai
But to answer your question, it no longer seems like AM3 CPUs will be backward compatible:
http://www.dailytech.com/AMD+45nm+DD...rticle7132.htmQuote:
Previous AMD documentation indicated that AM2 and AM3 would be forward/backward compatible -- yet AMD engineers claim the AM3 alluded to in 2006 is not the same AM3 referenced in the 2008 launch schedule.
"At the time AM3 was the likely candidate to become AM2+," claimed one field application engineer familiar with AMD's socket migration. "[AMD] wanted to keep the socket name associated with DDR2 memory and backwards compatibility, but AM3 emphasizes DDR3 support."
The INQ is confused because first K10, Barcelona, does not have HTT3.0. So if sockett AM2+ is meant for K10 and if K10 dosnt have HTT3.0 then maybe sockett AM2+ is not HTT3.0 only sockett. But as we all know INQ is easly confused. Barcelona is server chip and as the first K10 chip to help backwards compatablity its comes with HTT1.0, plus save time and money for AMD. I dont think i need to explain AM2+ sockett again.Quote:
Originally Posted by VulgarHandle
Also Sockett G or should i call it Sockett S1G1 or just Sockett S1 (The Turion X2 sockett) is just just the mobile version of AM2+ not a replacement of S1207.
http://www.amd.com/us-en/assets/cont...docs/41407.pdf
AM3 sockett will be physically differnt then AM2+ so that someone dosnt put AM2+ CPU into it. DDR3 has different pins to DDR2 so the motherboard will only support one memory type, seems pointless then for AMD to waste silicon on DDR2 support when it really is a waste.Quote:
Originally Posted by _damien_
No such thing as AM3+...it's AM3.Quote:
Originally Posted by Syn.
I don't think we're too far off from dual ddr2/ddr3 solutions...but enthusiast platforms have almost always used one type of memory, which we'll see here.
Prolly a value board from Asrock will have both.
Perkam
"agena" quad core at 125W? is this right? if it is dat's gonna git HOT.
early AMD X2 and dual core opterons operated at what..? 110. Not too far off. Some of the lower clocked chips are in 2 digits, no?
Ryan
I heard somewhere that AMD are going to use Intel's measurements for TDP. If so, that explains the power draw from 120W to 98W.
That makes no sense, since CD, C2D and C2Q is way below their TDP. Intel TDP thing is a rampart fanboy thing from the P4 days. Before and after the P4 its a whole different game.Quote:
Originally Posted by gOJDO
depends on the load state and application used. They can easily surpass their TDP in certain conditions while they can aswell dissipate less than their TDP, both under full load.
I'm eagerily waiting for these CPU's to pop out, as i heard they'll get a huge, incredible, boost in rendering apps over current top CPUs. Finger crossed :)
We don't know that for sure, but ATM I agree that incompatibility seems to be the most likely scenario. But yeah, if the CPUs will be incompatible, they'll almost certainly make the sockets physically different as well.Quote:
Originally Posted by Syn.
Not necessarily. Once again, if the IMC of AM3 CPUs were to support both memory standards, a possibility which we cannot rule out, the rest would only depend on the motherboard's DIMM slots. So provided AM3 CPUs will be compatible with both socket AM2 and AM3, support for both memory standards wouldn't be a waste, but a must.Quote:
Originally Posted by Syn.
http://dailytech.com/AMD+Claims+Quad...rticle7232.htm
Late Quarter 4
Quote:
Originally Posted by XeRo
so, the dual-cores will really be quad-cores with 2 cores disabled? hehe, anyone thinking what i'm thinking?...connect a bridge here, cut one there....Quote:
Originally Posted by DailyTech
No bridges. It's all hidden.Quote:
Originally Posted by VulgarHandle
The cores will be laser cut like the pixel/vertex shaders on nVidia G7x GPUs.Quote:
Originally Posted by VulgarHandle
ok, fine, pin mod then :D lol
More pretty pictures of mobos used in the demo:
http://babelfish.altavista.com/babel...e%2F37143.html
http://www.hardspell.com/pic/2007/5/...2c44501200.jpg
http://www.hardspell.com/pic/2007/5/...dda4501200.jpg
http://www.hardspell.com/pic/2007/5/...b854501200.jpg
http://www.hardspell.com/pic/2007/5/...a6c4501200.jpg
http://www.hardspell.com/pic/2007/5/...50c4501200.jpg
http://www.hardspell.com/pic/2007/5/...e634501200.jpg
http://www.hardspell.com/pic/2007/5/...66c4501200.jpg
http://www.hardspell.com/pic/2007/5/...9414501200.jpg
http://www.hardspell.com/pic/2007/5/...9bc4501200.jpg
http://www.hardspell.com/pic/2007/5/...0214501200.jpg
http://www.hardspell.com/pic/2007/5/...9414501200.jpg
http://www.hardspell.com/pic/2007/5/...d494501200.jpg
45nm wafer!^^
Uhmm..just like almost all other chipsets? (Cept nVidia perhaps).Quote:
Originally Posted by LOE
However if you wanna OC or the board is made for OC. Then you gonna get your 1kg copper anyway.
mmm those boards look pretty nice!
Except the dual socket one: chipset is in the worst place possible if you wanna use 1st pcie16x!
Quote:
Originally Posted by XS Janus
hope E-ATX one here :D
Wow, those are some tiny chipset coolers! I really love the power efficiency of the ATI/AMD chipsets, I wouldn't be surprised if the AMD quadfather chpiset draws more than 100w less power than the stupid Nvidia Asus one (it uses two northbridges, no?)
i think the asus one is better mobo ever donne if you want use your 2kwatts PSU over 500W :DQuote:
Originally Posted by Frodin
But: if Asus would come out with a board with that very same chipset, it would have 1 KG of copper to keep cool.Quote:
Originally Posted by LOE
When I saw the northbridge I was like :eek:
AMD knows how to make a cool chipset then :D Hopefully it'll stay cool while OCing too :D
I did some digging in the AMD 10h Programmer's Manual and Intel's Programmer Manual.
Barcelona shuffle latency (PSHUFHW): 2 cycles latency
Conroe shuffle latency (PSHUFHW): 2 cycle latency
Penryn shuffle latency: 1 cycle (from super shuffle engine)
Barcelona 128bit division latency (DIV*): 16-20 cycles latency
Yonah 128bit division latency (DIVPD, DIVSD): 63 cycles latency (not a typo!)
Conroe 128bit division latency (DIVPD, DIVSD): 18 cycles latency
Penryn 128bit division latency (DIVPD, DIVSD): ~5 cycles latency (estimated from Radix-16 divider)
Barcelona 128bit square root latency (SQRT*): 21-27 cycles
Yonah 128bit square root latency (SQRT*): 118 cycles (not a typo!)
Conroe 128bit square root latency (SQRT*): 29 cycles
Penryn 128bit square root latency (SQRT*): ~9 cycles latency (estimated from Radix-16 divider)
Conclusion: Shuffle speed improvement for Penryn shouldn't make too big of a difference... 2 cycles latency was already very low. Division/sqrt isn't that common (but is common in F@H) and performance will improve accordingly.
Do you have numbers for K8 whatever latency since you have Yonah and Conroe numbers?
If someone can find the K8 programming manual, sure.
Very interesting, thanks Shadowmage. Wolfdale and Yorkfield should be Distributed Computing monsters.
For A64 latency perhaps this document could help:
http://www.agner.org/optimize/instruction_tables.pdf
AMD Phenoms will introduce to the market this Monday - 14th, with HD2xxx cards!
Kudos to Shadowmage..now thats good work :) :clap:
Are you sure AMD heard about that :confused:Quote:
Originally Posted by OBR
I'm pretty sure it will just be an announcement saying "Our next-gen Processor will be called Phenom...."Quote:
Originally Posted by OBR
Yes i´ve seen... i have some slides under NDA ... and this NDA ends 14th ...
Guys I have conspiracy theory in addition to what (BTW well informed) OBR said.
Anyone with interest in F1 racing knows that tomorrow in Barcelona/Spain is next event...
By accident code name of new architecture form AMD is Barcelona!
If I would be AMD then guess what date I would choose for releasing first benchmarks to the public???
Kombatant(working for AMD as sort of PR/adviser) was hinting several times on Beyond3D forum that on 14th of May we will get very nice surprise besides R600XT... (it wont be RV630 coz, as rumor is saying, it's delayed).
On several older roadmaps Barcelona was slated for 1H launch.
Working quads was demonstrated in Monterey, California on 9th of May (AgenaFX with non-registered DDR2 DIMMS ;) 1 socket and 2 sockets configs)
THE END!!
:toast: if it's true ...
PS. I found original words from Kombatant:
Quote:
Originally Posted by Kombatant View Post
How's this: a decision was made regarding the R600 release, that will be a very pleasant surprise to many people.
Interesting/funny/controversial/stupid enough? ;)
Tnx Lightman:)
Now could you make the "Get all 3" concept fit into your conspiration theory please :DQuote:
Originally Posted by Lightman
Quote:
Originally Posted by Lightman
Not to mention Fellipe Massa won for Ferrari whom AMD sponsors BIG TIME !!! Now for that surprise :rolleyes:
nice
14th is NDA lifting date. Products launch is still in 2H
Full review launch or just preview ala Conroe?
The 14th is x2900 series card. No K10 anywhere. Not sure how people mixed it in with it. K10 is something there will be in summer/after summer.
NDA on K10 finish on 14.
Sweet. So how about availability? They've been flip flopping on that a lot, summer, Christmas..
Well, then we should see it tomorrow....or not.Quote:
Originally Posted by madcho
Are you sure? according to anandtech article (11 may)Quote:
Originally Posted by madcho
http://www.anandtech.com/cpuchipsets...oc.aspx?i=2986
"Barcelona Demos and Motherboards
Much to our dismay and definitely against our recommendations, AMD will not follow in Intel's footsteps and let us do a performance preview of Agena or Barcelona. In fact, AMD wouldn't even let us know what clock speeds its demo systems were running at. While we cautioned AMD that a lack of information disclosure at this point would only reinforce this idea that AMD is lagging far behind Intel, AMD's counterpoint does have some validity. AMD's reasoning for not disclosing more information today has to do with not wanting to show all of its cards up front, and to give Intel the opportunity to react. We still don't believe it's the right decision, and we can't help but believe that the reason for not disclosing performance today is because performance isn't where it needs to be, but only AMD knows for sure at this point."
So if even anandtech which was invited to the 10 may demo hasn't had the right to bench, it's curious that nda ends 14?
NDA finish the 14, but don't dream to see Barcelona on stock :nono:
a french web site :
translate by google ( seems good )Quote:
Officialisés dès lundi, ces nouveaux bébés devraient être en démonstration intensive lors du prochain Computex, mais ne seront pas disponible dans leur version grand public avant l'été voire la rentrée, pour les modèles Quad Core.
Quote:
Officialized as of Monday, these new babies should be in intensive demonstration at the time of next Computex, but will not be available in their version general public before the summer even the re-entry, for the models Quad Core.
Is it going to be hardware review site reviews or a bunch of amd.com marketing baloney?
It's not the problem...Quote:
Originally Posted by madcho
NDA: non disclosure agreement... Why a NDA, if nobody has info? Why anandtech (and every others sites) says they can't bench the 10 may? Do you really think that people who were invited by amd the 10 may, will not be some of the first to have the infos? (except amd staff)
Perhaps there is a 14 may NDA for some kind of amd product (R600, I guess, since board are around) but Barcelona? We will see, but i would be surprised if you're information is true (barcelona nda 14 may)
see my edit ;) :fact:Quote:
Originally Posted by nemrod
What should I see? They speak of computex so in June... And desktop version for this summer??? Very strange...Quote:
Officialized as of Monday, these new babies should be in intensive demonstration at the time of next Computex, but will not be available in their version general public before the summer even the re-entry, for the models Quad Core.
I have find your source:
http://www.pcinpact.com/actu/news/36...Huit-cores.htm
And it looks like they report what anandtech and other have seen but adding this date and computex? Curious?
Their source:
another french site: http://www.hardware.fr/news/lire/10-05-2007/
but no information about nda or computex on this one...
But It doesn't matter, if you're right we will see tomorrow, if not, we will wait few months again as we do since long time...
14... of what month?Quote:
Originally Posted by madcho
lol
there are several posts telling that NDA lift on K10 is 14th of May 2007 and people still asking the date :D
It will be soft launch meaning that actual products will be delivered to resellers only in second half of this year. Can't tell you more cos i don't know the exact data myself.
Tommorow AMD will publish a Press Release.
AMD Barcelona K10 B0 pictured
http://www.fudzilla.com/index.php?op...=959&Itemid=35Quote:
Exclusive: A new revision kicking well
Some of you might remember our Agena screenshot roughly a month ago and we have some news for you. Today we got a snap of a Barcelona sample, up and running. You can see the Agena here.
Barcelona on the other hand works at 2.4 GHz in Socket F 1207 motherboard. It needs 1.144V and CPU-Z recognizes it as AMD Processor model unknown.
The tested sample was model 8, stepping 1 famous revision B0. We already mentioned this chip here.
The CPU specification is even more interesting. This is an Engineering sample 00 (ES) processor. The sample works at 2.4 GHZ with a buss speed of 266 MHz and a multiplier of 9.
The HT works at 1333 MHz. The CPU Z 1.4 recognizes all four cores, 4x64 L1 data and 4x64 KB code cache and the whole 4x512 KB L2 memory but it can not read the L3 memory at least not from this core.
The best part of it is that there is a more to come, just give us some time. So this is how a new Opteron looks like.
So who wants to come over tonight for the countdown party?:toast: :banana:
1333Mhz makes no sense at all for HT. HTT1.0 has a max of 800Mhz, HT2.0 is 1.4GHz max, and HT3 has a max of 2.6GHz. No where is a 266/1333 bus mentioned. I call BS.Quote:
Originally Posted by informal
fudzilla or inq?? which one should i trust more
No! today is not endings NDA for K10! But only Phenom name of CPU!
266x5=1333Quote:
Originally Posted by [XC] Lead Head
I can do that on my DFI board right now. What's so strange about that?
CPU-Z cannot recognize muliplier and HT rightQuote:
Originally Posted by informal
http://www.fudzilla.com/images/stori...na_tested1.jpg
K10 cpu we see has clockspeed = 2400, bus = 200, multi = 12x and HT=3600 MT/sec
AMD Names Athlon's Successor: Phenom
http://www.anandtech.com/cpuchipsets...oc.aspx?i=2987
AMD Announces Phenom Processor Branding
http://www.dailytech.com/article.aspx?newsid=7263
You're probably right...Quote:
Originally Posted by OBR
http://www.amd.com/us-en/Corporate/V...117412,00.html
Fudzilla and CPU-Z screenshots dont mix after all those fakes.
Yes, because i KNOW it!Quote:
Originally Posted by nemrod
Quote:
Originally Posted by frankR
Because why would AMD chose to not even use the HT2.0 bus to its full potential, not to mention HT3 has vastly higher speeds then that. And that there no possible way to get an even 1800 or 2600MHz when the base bus speed is 266MHz.
The HT should have been 1.8Ghz or so. (3.6GT/s). So smells so much of a fake.
Barcelona is HT2.0 not HT3.0 ...
HT 3.0 is chipset related feature NOT primary CPU ... and Barcelona "Agena" is HT 3.0 ready!
Is there any AM2+ boards coming out soon(ish), or will they be released with K10 in Q4? I would of loved to pick up a X2 3600+ at ~50€ and an AM2+ board, later dropping a Kuma in it.
Boards will hit retail around June/July(at least that was the plan)
Is the actual K10 NDA in place until server parts launch in July/Aug, or is it going to expire in the near future?Quote:
Originally Posted by OBR
Actually it's a bus connecting CPU with chipset.Quote:
Originally Posted by OBR
Official date is 2H. But I wouldn't expect good variety until autumn. Two advantages of AM2+ are HT 3.0 support and multy lane power for CPU and memory controller which should be better in terms of overclocking.Quote:
Originally Posted by fyseni
Barcelona B0 2.4 GHz tested
:banana:
lol just more screenshots and we r bout 6months away from launch
The odds that I win twice a row in the lottery is higher. Photoshop and Fudzilla goes hand in hand.Quote:
Originally Posted by w0mbat
Yes, of cause :stick:Quote:
Originally Posted by Shintai
If its something good from AMD it must be a fake :slapass:
A poster in this thread from a Swedish forum claims that he created them:
http://www.sweclockers.com/forum/sho...hreadid=680996
Can u or someone read swedish? Is he just showing the score or is he telling that he faked this one?Quote:
Originally Posted by accord99
He's claiming to have faked it. If you scroll down, he's showing his original cpu-z, Sandra, etc. elements and the full quality fake screen shot that Fuad apparently used (and dirtied up).
Ok, that sucks very hard.
And you have been shown its a fake now. :stick:Quote:
Originally Posted by w0mbat
Yes, but im a positiv person so i always think the best :toast:Quote:
Originally Posted by Shintai
But u where right, ok, i lost :fact:
dam
He do know it :slapass:
http://www.imagebanana.com/img/wmraaour/Unbenannt.JPG
You have to understand that sweclockers is our countrys lamest forum ever.
If you think THG was bad well then you havent seen this one..
If Fuad knew how many that does that with him...h0h0h0 :rofl:
Well for all being said recently about Fudzilla and what they showed us so far. I really feel sad for what it is trying to do, putting seeds inside some minds is not going to help anyone but themselves.
Likewise... but I'd rather have a Phenom X4 :p:Quote:
Originally Posted by fyseni
You sure?! That would be cool :DQuote:
Originally Posted by informal
How retarded do FUD have to be not to realise it was probably another fake? (since they've posted about 5 allready)
He is only trying to make some money. He is looking after rumors and he tries to write attractive stories with a hope that some of them might be true. He is using the_INQ style of spreading BS. Although Fudo is writing very poor without any argument, fact, scientific data or logical explanation, the average mind without IT education consumes the BS without doubt and will visit FUDzilla again. How much Fudo can earn depends of the traffic to his site, so he'll keep writing BS.
Amd "they" better unlock those cpus otherwise no one will buy locked crap(again).
Good old, XP-Bartons!! Imagine how much cpus amd sell.!!
Overclocker can tell and judge wich cpu is good for use. Not a regular noob user!
This year- winner is intel with 6600
This thread is getting nowhere fast, but if i took it out, I'd have to take out the Penryn thread as well.
So for now it stays. But, next person to post a CPUZ screenshot that hasn't been verified gets a nice pm.
Perkam
gamespot review
http://www.gamespot.com/features/617...om_clk=topslot
(also a techdemo team fortress2 with a X2900 )
regards
Wrong thread?
I believe the 6600 is lock, isn't?Quote:
Originally Posted by Ego