Sign in with
Sign up | Sign in
Your question

AMDs Financial Day tomorrow, some BD info coming

Tags:
Last response: in CPUs
Share
November 11, 2009 1:51:59 AM

Go here for infos
http://blogs.amd.com/nigeldessau/2009/11/10/the-next-ge...
Starts at 8:30 AM PT
Or, Im sure therell be writeups on it

More about : amds financial day tomorrow info coming

November 11, 2009 2:02:30 AM

I got this thread bookmarked. I'm curious to see what AMD has coming in terms of product in the near future, including their 800-series.'

Here's hoping the words 'postponed' or 'delayed' are not anywhere in their statements.
m
0
l
a c 127 à CPUs
November 11, 2009 2:58:38 AM

Hope is all people have. They have delayed and changed their roadmap so much in the past few years its pretty crazy.

Maybe they are on the right path though since they did get Phenom II out when they planned it.

Well if they even planned it or it was a jump for 45nm anyways.....
m
0
l
November 11, 2009 3:48:36 AM

Istanbul came out 6 months early.
m
0
l
November 11, 2009 2:42:34 PM

The stocks went up 11 cents, maybe they have good news
m
0
l
November 11, 2009 2:59:36 PM

Hope so. Im thinking theyll be stories out and the first link may, may be followed up by Fruhe, if hes not too worn out
m
0
l
November 11, 2009 4:34:14 PM

I didn't see this thread when I made mine. I'll just copy the relevant bits here.

The thing I almost creamed myself over was the info about the upcoming Bulldozer in 2011. It's going to be an AM3 part. Untill now I hadn't read anything other than speculation and rumors. I have no idea what socket Intel's Sandy Bridge is going to be but if it needs a new platform AMD might have quite an advantage when they come out. Unless Bulldozer is a huge failure against Sandy Bridge I don't see myself switching to Intel in 2011.

Edit:
source http://phx.corporate-ir.net/External.File?item=UGFyZW50... page 26
m
0
l
November 11, 2009 8:14:48 PM

Wow i just read the whole slideshow and i must say that AMD is making a huge comeback. And a major plus- most technologies are.... ahead of schedule!!! :bounce: 
m
0
l
a c 87 à CPUs
November 11, 2009 11:11:54 PM

I'm not impressed. From the slide linked (I don't have time right now to read the links.) thats their priorities for 2010? As in all their goals for the year? Lets see, 1) winning platform could mean nearly anything. 2) Increase LAPTOP performance and battery life by 25%. Thats good, assuming your running a laptop. What do us desktop users get? 3) Expand OpenCL. Neat, maybe some products will come out in 2011? 4) Win DX11 NOTEBOOK transition. That should be easy, what does nvidia have out that's DX11? 5) Launch 12 core Opteron CPU. Great, a server chip, and even more cores that might not scale with consumer programs.

So for those of us gamers, we get what exactly? What new CPU thing is AMD planning on releasing soon to get us excited? If that's the most exciting slide showing whats coming, I'm not happy.
m
0
l
November 11, 2009 11:27:15 PM

I think everyone agrees with you, the sooner the better.
And unfortunately, so far, no big surprises on release dates/timing
But, you have to be encouraged with the 1 TB ability at the low end, means better games down the road, thats IF LRB can keep up, and leaves nVidia cold tho for low end
m
0
l
November 11, 2009 11:30:51 PM

Think of a 4850 for low end, sound good?
m
0
l
November 11, 2009 11:46:40 PM

JAYDEEJOHN said:
Think of a 4850 for low end, sound good?



No, that puts me below low end. Not cool.

Though, Bulldozer looks sweet.
m
0
l
a b à CPUs
November 12, 2009 2:58:09 AM

4745454b said:
So for those of us gamers, we get what exactly? What new CPU thing is AMD planning on releasing soon to get us excited?

What exactly do you want? The only thing AMD can really do to potentially boost gaming performance is higher clock speeds and higher IPC. It's not like a consumer 12-core CPU is going to be useful to you. CPUs are for work, GPUs are for games, HPCGPcGPUs are for lots of work.
m
0
l
a c 127 à CPUs
November 12, 2009 3:32:39 AM

4745454b said:
I'm not impressed. From the slide linked (I don't have time right now to read the links.) thats their priorities for 2010? As in all their goals for the year? Lets see, 1) winning platform could mean nearly anything. 2) Increase LAPTOP performance and battery life by 25%. Thats good, assuming your running a laptop. What do us desktop users get? 3) Expand OpenCL. Neat, maybe some products will come out in 2011? 4) Win DX11 NOTEBOOK transition. That should be easy, what does nvidia have out that's DX11? 5) Launch 12 core Opteron CPU. Great, a server chip, and even more cores that might not scale with consumer programs.

So for those of us gamers, we get what exactly? What new CPU thing is AMD planning on releasing soon to get us excited? If that's the most exciting slide showing whats coming, I'm not happy.


AMDs main focus is going to be in the server market. Remember this: A Phenom II is the same CPU as a Opteron equivalent just priced for consumers. Server chips normally start at $1K and go up to $3K sometimes. They make much more money there than they do on the DT.

If anything I would expect the 5870X2 to be the main on in 2010 then the 5890 and possibly 5890X2. Probably wont see much of a gaming boost from AMD until R900 which might not be out till 2011 since nVidias G300 is not set till next year and all ATI will have to do is a refresh of the 5800 to stay in the game.

randomizer said:
What exactly do you want? The only thing AMD can really do to potentially boost gaming performance is higher clock speeds and higher IPC. It's not like a consumer 12-core CPU is going to be useful to you. CPUs are for work, GPUs are for games, HPCGPcGPUs are for lots of work.


Lol....
m
0
l
November 12, 2009 4:58:42 AM

So, Thuban will be a server chip?
m
0
l
a c 87 à CPUs
November 12, 2009 5:39:30 AM

Thanks guys, but I was asking you to tell me something I don't know. I know that a 12core chip isn't going to do much for me, I pointed that out earlier. If they had a slide saying something about a new process that will allow higher clocks, or a really new Arch coming that will increase IPC then I would be happy. But saying things like 12core chip and increase battery life doesn't "speak" to me.

Actually Anand had a slide on their site that did get my heart rate up 5 beats.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=367...

Like hyper threading only better due to more hardware? So they to will be able to release a 4 core/8 threaded chip like the i7. I assume this will help them compete in threaded apps, and be cheaper to make then a real 6 core and definately the 12 core monster. Still won't help in gaming much. We'd need their version of Overdrive for that.

I wonder what other improvments Bulldozer has?
m
0
l
a b à CPUs
November 12, 2009 5:51:38 AM

Meh, I want more FP performance for the CPU, not Integer.
m
0
l
November 12, 2009 5:53:37 AM

Itll be IPC mucho improved, itll have higher clocks. Cant find the refernces, looking now. Ill find em.
From a gaming perspective, itll have a higher connect rate, eliminating completely any Intel advantage, but we will se this before BD anyways. Its the main reason for the multi card perf differences, which show higher than IPC in some cases
m
0
l
a c 127 à CPUs
November 12, 2009 5:58:41 AM

JAYDEEJOHN said:
Itll be IPC mucho improved, itll have higher clocks. Cant find the refernces, looking now. Ill find em.
From a gaming perspective, itll have a higher connect rate, eliminating completely any Intel advantage, but we will se this before BD anyways. Its the main reason for the multi card perf differences, which show higher than IPC in some cases


Still like everything to be taken with a grain of salt. There is no 100% until its out to the testers to show us.

AMD can tout like Intel does all they want and we can swoon to it but the truth of the matter is that everything they say cannot be taken as 100%.
m
0
l
November 12, 2009 6:05:32 AM

Certainly, but 2 things.
First Intel
Intels image has to be seen as the leader, without question, no slipups, bring the stability to IT industry yadda yadda, so, deviation is somewhat out the window
AMD. Under the gun. No more behind, underperforming, HAS to perform for its very credibility, for future existence/investors etc, so, deviating is out the window.
Having said that, everything changes down the road, so it happens.
This isnt the old AMD, everyone recognizes this is more like the older AMD, where they domintaed in perf etc yadda yadda.
Theyve shown signs of it lately, maybe this is just more of that.
m
0
l
a b à CPUs
November 12, 2009 6:17:42 AM

Barcelona outperforms Clovertown by 40%.
m
0
l
November 12, 2009 6:28:38 AM

Exactly, the old AMD, we need the older AMD, and lately, starting from where they were, and coming thru alot more, its lookig that way
m
0
l
November 12, 2009 7:28:40 AM

randomizer said:
Barcelona outperforms Clovertown by 40%.


That was true* (the actual statement was 40% clock-for-clock IIRC).


*In very specific applications.



The quote got twisted out of all context and was presented as across the board.
m
0
l
November 12, 2009 7:43:38 AM

I remember trying to pont that out back in the day, as thats what I read too, but that too went the way of HKMG
m
0
l
a b à CPUs
November 12, 2009 8:33:09 AM

Amiga500 said:
*In very specific applications.

The ENIAC could probably beat a pocket calculator in very specific applications too. There's no point advertising a number that nobody will reach in reality (although it is marketing...).
m
0
l
November 12, 2009 11:03:26 AM

randomizer said:
There's no point advertising a number that nobody will reach in reality (although it is marketing...).


No, well...

Certain non-linear FEA (LS-DYNA) and all CFD simulations would invariably show Barce as over 40% quicker than clovertown clock-for-clock.


For an opteron, those are definitely advantages worthy of broadcasting. However, they don't translate to the desktop.
m
0
l
November 12, 2009 12:06:54 PM

I hope AMD can deliver on all this stuff, I'm actually kinda excited about AMD 2010!
m
0
l
a b à CPUs
November 12, 2009 12:09:45 PM

Just about everything is ahead of schedule.

In all probability, Llano has taped out with BD taping out soon if it hasn't already.

Global Foundries are ready to take orders for 32nm, so they are way ahead of schedule too.

I think we will see Llano before August 2010, BD coming November 2010, possibly even earlier if everything goes right first time.
m
0
l
November 12, 2009 12:13:25 PM

A lot different than what was being said, thats for sure.
Cpnfusing 40% in 1 test (although, bringing that up as a marketing point is bad) and confusing 32nm with BD is really bad.
Im looking forwards to Bobcat etc, and next years the year.
Better gaming for all, a possible end to consoles, as every AMD cpu should play games
Meaning, no more lousy porting etc, as devs have a decent foundation to do their thing, and way before consoles get off their lazy butts for their next iteration, expected not til 2012
m
0
l
a b à CPUs
November 12, 2009 12:21:04 PM

According to p24 of the presentation, even Thuban is ahead of schedule...which means its coming Q1 2010 surely. "Dorado" I dont get, that just looks like Deneb with another name to me, probably just deneb with the new chipset.

p41 says they aim to deliver 4 new pc platforms in H1...Thuban and Dorado check...I guess another one will be Athlon based Dorado...mby that means a 6-core Athlon is coming too?

6 cores for < $100 maybe?

m
0
l
a b à CPUs
November 12, 2009 12:23:59 PM

I am excited ... I can hear "Amazing Grace" on the telly too.

Could there be a "religious" moment coming for the true believers ??

I might just have gas though ... or a tumour.

/opens an amdzoner window for a peek at delusion.
m
0
l
November 12, 2009 12:26:56 PM

I can imagine what its like over there, but since I avoid that place, let us know, and, please come back, please
m
0
l
a b à CPUs
November 12, 2009 12:34:09 PM

I'm thinking a 6-core athlon II would be small and cheap to make, and sell by the barrowload.

If you consider the x4 620 is on par with the q8200, what do you reckon a 6 core would be on par with?
m
0
l
November 12, 2009 12:38:51 PM

6 real threads vs 4 plus 4 virtual?. Should do well
m
0
l
a b à CPUs
November 12, 2009 12:50:13 PM

It won't give you any further single threaded performance with current software - so unless the core has significant revision, or they reduce the latency a shedload on the cache, and lets face it it isn't wide enough for a start, then the returns are diminishing from a single socket perspective.

They need a more sophisticated prefetcher and reduced latency on the cache elements.

AMD talking 6 core is like when Intel were pushing the last of the P4's.

Time to do some serious work on the core design.

Speed, cores and cache isn't the answer in simple terms
m
0
l
a b à CPUs
November 12, 2009 12:51:17 PM

A lot of people might be dissing Thuban thinking they don't need 6 cores, but what they really mean is they don't want to pay enthusiast prices for 6 cores.

I do expect AMD to release an Athlon II X6 after the Phenom II X6 has been out a couple of months. Obviously they don't want to hurt sales of Thuban too much.

Sounds crazy I know but Thuban should beat an i7 920 and Athlon II X6 won't be far behind an i7 in performance, costing 450m transistors and probably not much more than $100...while plugging into current AM2+ mobos.
m
0
l
a b à CPUs
November 12, 2009 12:51:48 PM

And while I am at it, it is time to drop the power draw to a couple of watts on idle now ... sustainable power use please !!
m
0
l
a c 87 à CPUs
November 12, 2009 3:32:38 PM

Haven't seen this yet, this just in.

http://tech.yahoo.com/news/ap/20091112/ap_on_hi_te/us_i...

Cases settled, GF good to produce chips, and AMD is 1.25B richer. Not a bad way to wake up if your AMD.

Quote:
AMD talking 6 core is like when Intel were pushing the last of the P4's.

Time to do some serious work on the core design.

Speed, cores and cache isn't the answer in simple terms


Exactly Reynod. This is what I'm hoping they get around to doing. I remember when they increased the size of the L2 cache, then doubled the latency to access it. Sure it had more cache, but the time it took to get the data went up a lot. Net effect? No increase in speed. If BD implements AMDs hyperthreading AND makes it faster internally, then it will be a great chip. But they have to stop just throwing cores at the problem.
m
0
l
November 12, 2009 4:05:32 PM

jennyh said:
A lot of people might be dissing Thuban thinking they don't need 6 cores, but what they really mean is they don't want to pay enthusiast prices for 6 cores.

I do expect AMD to release an Athlon II X6 after the Phenom II X6 has been out a couple of months. Obviously they don't want to hurt sales of Thuban too much.

Sounds crazy I know but Thuban should beat an i7 920 and Athlon II X6 won't be far behind an i7 in performance, costing 450m transistors and probably not much more than $100...while plugging into current AM2+ mobos.


But load distribution isnt the same as clock for clock comparison.

Global foundries is saying a 50% increase in performance at 32nm and hkmg at 28nm meaning AMD is relying more on nodes than metals that increase electron ballistics. While they have done a stellar job at tuning an aged and highly reworked mArch it remains to be seen if 32nm will be released at optimal perf levels or as with phenom and 65 45nm chips in general that performance will evolve as chips sell and revisions are made.

AMD has a slower tick tock but at least it finally comes around.
m
0
l
November 12, 2009 4:18:05 PM

Whats optimal? G0?
Steppings are needed. Bugs get fixed etc.
Using SOI, maybe they already have their solution? Too early to say
m
0
l
November 12, 2009 4:49:27 PM

it remains to be seen if SOI can scale to 28nm jaydee. Alot of people expected it to die at 45nm yet the promise is that 32 will also be soi. and point 2 on that is afaik hf02 is incompatible with soi and they arent talking HKMG until 28nm so clock for clock AMD's electrons move slower since they retain older metals that dont allow higher electron ballistic thresholds.

Edit: You know as well as I that Intel was trashing AMD from the point of core2 until phenom 2 was released and phenom 2 finally matured with 955. That was a long painful period of 2-1/2 years. I think the devil in the details post Hector leaving is that it was likely he who held AMD back, but Dirk still has some clockwork to do of his own.

On a side not Douglas Grose is the perfect candidate for ceo of GF as he was a key innovator at IBM in packaging, all in all things are looking better but with the devestating past 2.5 years AMD has alot to atone for.
m
0
l
November 12, 2009 5:19:20 PM

Links for the SOI scaling? I remember something awhiles back
m
0
l
November 12, 2009 6:04:12 PM

http://www.research.ibm.com/journal/rd46-23.html for 10 dollars :D 

http://money.aol.com/article/leti-caltech-workshop-to-p...

Looks like SOI as with bulk has simply got hurdles not restrictios, I would imagine that the hurdles as with any other silicon chip process are metallurgic, or rather how the metals are doped and alloyed.

http://www.semiconductor.net/article/199938-Bulk_or_SOI...

David Lammers, News Editor -- Semiconductor International, 7/31/2007
Advanced Micro Devices (AMD, Sunnyvale, Calif.) is still mulling whether to use silicon on insulator (SOI) or bulk silicon technology for its future high-end and mobile products.

“This is in the exploratory phase, and AMD has not made any statements about when it would make a decision to produce next-generation processors in bulk or SOI,” a spokesman said, adding that AMD technologists are asking questions that “are not answered yet, so we are leaving the question open as they look at the issues.”
from the feedback section.

One of the major differences between the SOI and Bulk technology for the 45nm and beyond is to control the electostics or the short channel effects. For the bulk technology used by Intel the quantum confinement of carriers is controled by a combination of Hallow source/drain implant, and retrograded channel/substrate doping. On the other hand, for the SOI technology the quantum confinement of carries in inversion layer is carried out by physically reducing the SOI thickness, Tsoi by narrowing the space between the gate oxide and the buried oxide. To mitigate the short channel effects, 45nm SOI may require 50nm~40nm Tsoi, 30nm~20nm Tsoi for 32nm, and 10nm or less Tsoi for 22nm technology. Such a thin Tsoi causes significant carrier mobility degradation and increase in threshold voltage, Vt. Furthermore, for the scaled devices, the strain induced mobility enhancement techniques become less effective. This is particularly more so for the thin SOI technology simply because in such a thin ~10nm junction and isolation depths, and channel inversion layer thickness it is extremely difficult to implement GeSi S/D junctions and a large lattice mismatch induced by the relaxed Ge-Si substrate in the channel. Even for the 45nm SOI technology, the manufacturability of the strain induced mobility enhancement techniques used for 95nm and 65nm may not be feasible. In this respect, the SOI technology for the 45nm and beyond has a significant disadvantage over the bulk technology. IBM and AMD are at the crossroad today to determine extenderability as well as manufacturability of the SOI technology for 45nm and beyond. The conversion from the SOI to the bulk 45nm technology node has enormous technological and manufacturability challenges. This is because IBM and AMD do not have the required learning experiences such as process, design, reliability and device yield gained from the 95nm and 65nm bulk technology development and mannufacturing. Furthermore, two major new materials were introduced in the bulk 45nm technology: the thermal oxide, SiO2 that was used for 40 years is replaced by HfO2, and the polysilicon gate that was used for over 30 years is replaced by the metal gate. Today Intel is the only company that is manufacturing the bulk 45nm. If that is true, Intel has enormous advantages over its competitors, particularly if IBM and AMD have to adopt the 45nm bulk technology. This is because Intel must have resolved most of the device, process, reliability, and manufacturability issues as a result of introduction of the new materials and processes.
When the new materials and processes like HfO2, metal gate, and their new processes are introduced, new or unknown faiure mechanisms will be also introduced. Therefore, it is crucial to design test structures so as to bring out the unknown failure mechanisms for early detection, and develop effective E-test and reliability test screens. Such experiences gained through the 95nm and 65nm bulk technology development cycles will give an edge to Intel in successful development of the 45nm technology and beyond.




Sang U. Kim, Ph.D - 8/1/2007 3:29:00 PM CDT
» More
m
0
l
!