What is Digg?75 Comments
- fatnutz, on 10/10/2007, -0/+45Most people look at AMD like a wounded animal, we all love our Core 2's but we all know that had the A64 not been such a successful and powerful chip, Intel would still be jamming ***** egg-cooker processors up our asses at ridiculously inflated prices.
- WarMace, on 10/10/2007, -3/+37The ability to unite cores to assist single threaded applications is pretty cool. Especially since games still seem to be single thread minded.
- bunnybash, on 10/10/2007, -0/+15as an ex AMD fanboy, i feel like every purchase i make these days is with intel... makes me sigh, but there is no use staying on a sinking ship... AMD pushed intel so hard earlier this century, and essentially the A64 forced intel to innovate for the first time in years, with the C2D... maybe the C2D will force AMD to do something truly creative and powerful, and useful... hmmm here is hoping!!
- maninblac1, on 10/10/2007, -7/+17I don't know what to say, it's either brilliant or insane. The course of AMD's market will change dramatically in the coming months one way or the other. If Barcelona doesn't pan out, the desktop is going to become secondary to AMD's server and UMPC efforts. You notice AMD had lots of content concerning mobile devices and pushing x86 to those devices.
This may be in response to intel getting snugly with OLPC, but x86 isn't there now, and probably never will be. That area will go to Sparc and ARM and TI who are making simply controllers and low power asynchronous processors, along with DSP etc.
The anouncement of 3+ GPU solutions at this point is a joke, and when you go out and by your 1200-1500W PSU to handle 4 2900XT graphics cards. Quad SLI on the old G7 series will have that beat for quite some time just because it's lower power and much more affordable.
I look at today's announcements as a feint, it's not that i don't think these will come to market, they will. This is the gorilla beating it's chest to announce it's still in the room.
And let us not forget the fact that AMD is hurting for $$, taking donations from Europe, and getting IBM to help them out in research, it's smart, but it shows how desperate AMD really is. As you can see, AMD no longer has the luxury of developing at their own pace, they are now playing catchup, and have a bleeding wound in their leg. I can't wait for the next 9 months to pan out. - mpcamer, on 10/10/2007, -4/+14Phenom at 3.0 GHz with three way Crossfire: http://www.hardocp.com/news.html?news=MjcxMjcsLCxoZW50aHVzaWFzdCwsLDE=
- fuzzynyanko, on 10/10/2007, -2/+11AMD is behind, but at least they come up with some good names for their stuff.
- Darthyoshiboy, on 10/10/2007, -3/+10From the moment this whole AMD/ATI thing started to come to light, I knew it would be of no good to anyone...
What's the old adage? Anything that is supposed to do everything, usually doesn't do anything very well? - DeFex, on 10/10/2007, -0/+6No wonder they have been so slow, they have fired the engineers and hired lintels old marketing department. Way to powerpoint AMD!
- frgmstr, on 10/10/2007, -2/+8No, actually it is a Phenom part. :)
- maninblac1, on 10/10/2007, -0/+5AMD does not have any 3Ghz barcelona opterons, or they wouldn't be wasting time puting out lame 2Ghz ones. It's a phenom. Plus how do you justify 3+PCIex16 slots in a server or even a workstation motherboard? It's PHENOM.
- frgmstr, on 10/10/2007, -0/+4You will see top end Agena Core AKA Phenom parts for sale in December and we should see servers for sale with Barcelona processors in August. I would not expect to see AMD swinging for the fences though. Performance per watt is what data centers and companies are looking for now...with obvious exceptions of course.
- Daiken, on 10/10/2007, -0/+4AMD's chip architecture and designs are great as is. The issue is of course that Intel's one step ahead of AMD on transistor sizes. They've been at 65 nm for a while and now they're going to 45. Where as AMD is still getting everything over to 65 nm from 90. Course the only real solution to this is spending a couple more billion into research and fab development which AMD just can't do. Hence they're always under pressure to keep having better designs than Intel.
- maninblac1, on 10/10/2007, -0/+4AMD announced that today. Though, it was really AMD who wanted to get GPU on die first, since intel had such a commanding lead in integrated graphics, onchip graphics would surely play nicely with AMD's platform architecture. Hence why intel is stealing it.
- Caffeinate, on 10/10/2007, -7/+11Great article as always from the [H]. AMD certainly is in need of excitement for new products - now if we can just get them on the shelves!
- maninblac1, on 10/10/2007, -0/+3I'm actualy and intel fanboy dumbass if you could read any of my other comments, and no chip manfucturer "starts" production on a chip and then abandons it, it costs too much money. There was no such concept chip, nor a platform to support it, site your sources, or STFU.
- geminitojanus, on 10/10/2007, -1/+4"You notice AMD had lots of content concerning mobile devices and pushing x86 to those devices. This may be in response to intel getting snugly with OLPC, but x86 isn't there now, and probably never will be. That area will go to Sparc and ARM and TI who are making simply controllers and low power asynchronous processors, along with DSP etc."
Everyone's concerned about x86 On-the-go because we can actually do it now. Laptops were great and all, but there's always been problems with making chips for laptops, problems that were almost completely and entirely solved going into the Core generation, and now that things are getting smaller and smaller, why the hell not package one for extreme mobility? Sell a $25 dollar piece of silicon for close to $100, and print 300+ of these chips on each wafer, they'd have to be complete morons not to consider it. Intel wants to get there first because, quite frankly, AMD has the better chip for the job and are already selling some of them (though not in very high volumes, which is sad all things considered). They even have the upper-hand in connectivity (HyperTransport), and yet they refuse to use this advantage.
AMD's biggest problem is just not realizing what they have, and trying to do things they shouldn't be. Stick to simple, quick processors. Get them out the door. Barcelona should have been the elephant in the room for AMD's management; delaying in a fight with Intel is like falling asleep in a snake pit, you just don't do it. - maninblac1, on 10/10/2007, -0/+3Because AMD has never produced ultra low power parts. The reason, SOI processing, supirior to bulk arguably, but it has it's own share of problems. Mainly keeping leakage currents under control, the irony of course was that was SOI was suppoed to do.
This is an interesting read if you're so inclined to read about the power consumption of AMD parts. They had to introduce C4 power states on desktop processors (normally reserved only for laptop processors), just to keep power down.
Source: http://www.overclockers.com/tips01182/
And: http://investorshub.advfn.com/boards/read_msg.asp?message_id=21353453
Why should AMD get into a sector it will obviously struggle with, sure there's lots of money to be made if you're the first one there. But they aren't, they're competing against technologies that use no power at idle, that only consume power when data must be processed, clockless processors, something that neither AMD or intel has lots of experience with, that's the future of ultra mobile ultra lower power.
On another interesting note, if you look really closely, there's a diamond in the rough, and it gives you an idea of what AMD things will save them. They are going to be behind Intel in process size by 12 months at least. But they're looking to jump ahead of Nvidia and get graphics processing much much smaller. I don't think we've even heard Nvidia breathe a word of 80nm let along 65 or 45nm. Look for AMD to see ATI as it's leader when it leaps ahead of Nvidia cooling power and performance with better fabrication.
The AMD ATI merger was a win for ATI and a loss for AMD. - maninblac1, on 10/10/2007, -0/+3It's archtectural genius, or so AMD would like to think. As things get more parallel things get more complicated for software. OS's are very very good at scheduling processors on single processor systems, so what if you could make 4 processors look like 1 processor. You'd get better scheduling and less overhead.
The conceptual problem is this (and this is what confuses me), if you tell an operating system their is 1 processor, and it schedules like there is 1 processor, how on earth do you get stuff into the other 4 processors!? I'd have to think that so much programming would have to be hardcoded into transistor logic to handle things for the OS. I really can't possibly see how it would work, but they're smarter than me, so i'll trust them. - maninblac1, on 10/10/2007, -0/+2Correction, not all of AMD's chips are 65nm. They have finished 65nm mobile, they still can't get it out the door.
- maninblac1, on 10/10/2007, -1/+3I'm going to have to say that is one of the most idioitic things i've ever heard, and kind of insulting. As a computer engineer, evolution is important absolutely, but if you never think outside of what you know, then nothing ever gets better.
To say that all the best things have been designed so all we need to do is tweak them is probably the most naive thing i've ever heard.
So let's get one thing straight, Netburst, was a good idea, and it did work, 5 years it worked, but when P = C*V^2*F+(stuff), and you need more V for faster F, P goes up, and when P is directly related to TDP, we have problems. It was an evolution that lead to extinction. Every so often, a revolution is necessary.
This is why you see both companies have opted for "tick-tock", revolution, die shrink evolution, revolution, evolution etc....why do it this way. Because both companies understand an archtecture can only get them so far. Both Intel and AMD know now that multi core will stop, there will be a maximum amount they can put in, and when that happens, there will be another HUGE revolution, likely opting for bigger chip sizes, to offer more cores, but then more heat will be generated, and we're back to where we started. That's where "evolution" gets you.
So pardon me, while i think of a revolutionary idea to avoid this fate, cause that's what engineers do. - cbreaker, on 10/10/2007, -0/+2If you've checked out the big box stores lately, there's a lot of AMD there. Half the notebooks, and somewhere less then half of the desktops all have AMD processors it seems. HP seems to push them the most.
The Athlon and Opterons are still excellent CPU's. For the workstation, it might make sense to get a Core 2 Duo because you'll get a few extra % power, but for server operations I still recommend the Opteron servers; especially if you're going to be doing heavy SMP and RAM usage. The HyperTransport really makes those servers reach full potential. No Northbridge bottleneck. (We just recently purchased two Dell 4-way dual-core Opteron servers with 32GB for VMware. These machines scream.) - RawOysters, on 10/10/2007, -0/+2It's all fluff till I see benchmarks.
- diggonit, on 10/10/2007, -1/+3It's all about where the rubber meets the road these days. This PR doesn't mean nearly as much as it used to. Just look at how fast Intel made up the market share that was lost to every ones darling AMD. They better put out some quality products and soon because I doubt Intel is just sitting around all fat dumb and happy this time.
- maninblac1, on 10/10/2007, -0/+1"fat dumb and happy"
I think that's the funniest way i've heard of refering to intel in the P4 days. - maninblac1, on 10/10/2007, -1/+2There is only one tiny difference between then and now. Intel had 5 years to squat and think it over buying time by ramping up frequency. AMD has had essentially 18 months.
- maninblac1, on 10/10/2007, -1/+2Agreed, CPU companies don't announce features they can't deliver on, a delay is one thing. Canceling an entire feature, is millions of dollars wasted, and could destroy a whole architecture if integrated to deeply.
You don't see well, we announce this to be in these chips 6 months ago....but we don't think it's going to pan out, so we're going back to the drawing board. It doesn't happen.
But i'm a little skeptical myself as to how well it will work. They didn't go into enough detail about how it works IMHO. - Reziarfg, on 10/10/2007, -1/+2Stick with "don't count your chickens before they hatch". This is AMD. It's not vaporware and they're more than capable of delivering.
- geminitojanus, on 10/10/2007, -2/+3Yeah, because that feature was both mentioned and exists in reality.
I really hope I don't need a sarcasm tag; if it sounds too good to be true, it is. - dstz, on 10/10/2007, -0/+1Great news. For what I experienced their processors seem so much better with single threaded, real time, heavy duty floating points operations, as in sound synthesis softwares (I have no idea why).
- VanD, on 10/10/2007, -3/+4It's really too bad because AMD has done so well for themselves in the past, but really just lost their step.
Hopefully they can come through big with their new products. - majortom1981, on 10/10/2007, -7/+8AMD has to do something. Intel is also working on fusing gpu with the cpu and also putting the memory controller on its cpu just like on AMD's. AMD needs something HUGE to just keep up with intel.
Go ahead and digg me down if you want but thats what I think. - Tenoq, on 10/10/2007, -0/+1What, you mean like a Windows PC with MS Office?
- geminitojanus, on 10/10/2007, -0/+1You see what the problem with that statement? Complacency. Both companies rested on their laurels while the opponent toiled away in the lab. 18 months is plenty of time to spin an architecture as long as you're basing it on an older one, and this kind of development should be ongoing and not "stop-and-go"; this should not be a revolutionary business at this point, it should be almost entirely evolutionary from Intel and AMD (and revolutions should come from other parties wanting to break into the market, or at least by sub-parts of the larger companies; tasking your better engineers on revolutionary ideas are the kinds of things that kill companies).
Intel's Netburst was one of the smartest things Intel could have done for itself (and for AMD). AMD squandered that time by getting into a clock race that they couldn't win; even if Intel couldn't build the chip to produce less heat, it sure could build a lot more of them. Instead, AMD's engineers toiled away at making "true n-core" solutions instead of looking at what the market wanted and was eating up. And their marketing teams are going to have a nightmarish job of passing of "true quadcore" as an advantage to "quad core for $300, $500". - LaughingMan11, on 10/10/2007, -0/+1Has anyone actually confirmed with AMD that they are going to do this? Reading the article, it sounds more like [H] interpreting a slide from the AMD presentation and making a leap to saying that they somehow magically got more than one core to work on a single thread.
From the article : "The statements that most needs to be focused on are, “Continued Scaling for single-thread performance,” and “Partitioned for future scalability and modularity." That would tell me that Bulldozer could be morphed according to the importance of the application whether it has specific multi-thread needs or single thread needs and still take advantage of all its resources."
That sounds like the HardOCP guy talking hopefully, NOT AMD confirming that the feature exists in the form that the HardOCP guy imagines. - maninblac1, on 10/10/2007, -1/+2There's a saying in the industry about fabrication, from a guy i knew who worked for TI. "Intel moves first, always."
Who had the first mass produced 130, 90, 65nm anything, intel...
AMD could throw all the money in the world at it, and probably still not get ahead. - thatbox, on 10/10/2007, -0/+1Did you just call using four graphics cards "lower power" and "affordable"?
- LaughingMan11, on 10/10/2007, -0/+1No one has confirmed with AMD that such a feature exists! The quote from HardOCP sounds like the author wishing for this imaginary feature, not a confirmation from someone at AMD.
- frazw, on 10/10/2007, -0/+1Intel and AMD have been swapping roles for years. One minute Intel are top dog in performance, then AMD take it back then Intel again etc. Remember when AMD 64 was released? Intel had previously been top dog and got complacent so were panicking back then, (although I'm sure some intel fanboys will argue with me there). Their panic lead them to release a series of substandard chips just to keep in the game whilst they were developing their Dual core chips. AMD as the performance leader got complacent and lost the lead.
The only reason I think intel weathered their low points so well is because of their aggressive advertising, e.g. the intel noise and manufacturer deals.
AMD never seemed to advertise in the same way, I've never seen an AMD ad on TV here in the uk.
AMD from my standpoint have always been the better option because of the price, I always buy the best I can afford at the best value for money and that has always been AMD I suspect that is why they are able to weather the bad times. I expect that even if this Bulldozer doesn't turn out swing things back in their favour, they have something else in development.
Intel presumably also have their next big thing in development too and I wouldn't be surprised if they either make an announcement shortly or wait until just before the launch of these new AMD chips before announcing.
Regardless I think the bulldozer WILL swing things back to AMD, then Intel will release and the market will swing back to them.
Necessity is the mother of invention and this has been proved in the past. - Giga, on 10/10/2007, -0/+1They didn't really lose their step, their hardware has been improving the whole time. Intel just decided to get off its arse and replace the netburst system, and had a whole lot of cash to do it properly.
- maninblac1, on 10/10/2007, -2/+2IBM won't let AMD die, if AMD goes under, intel's next target is PPC, and the back server room. IBM will buy AMD if AMD is about to go under.
- maninblac1, on 10/10/2007, -1/+1What are you talking about, their mobile market dreams are dead because their processors hemorrhage current. Software hasn't dictated the success of hardware in 10 years.
- iapx, on 10/10/2007, -0/+0Bla bla bla...
AMD show graphs and presentation while they didn't do anything new these 3 last years!
While Intel delivering desktop graphic with GMA950 and GMA3000/X3000, desktop power with Core2 Duo and Core2 Quad (core2 Quad is NOW available for under $1000 deskop PC), Xeon (based on Core microarchitecture as Core2 CPUs), AMD still wasting time, decreasing price of outdated CPU that couldn't cope in terme of performances, and showing dreamland vaporware presentation!
Is it the same AMD that gives us 386 at 33Mhz and 40Mhz while Intel ended-up at 25Mhz?
Is it the same AMD that creates Athlon?
Is it the same AMD that extends x86 architecture to 64bits, leading Intel to deliver "AMD64 compatible x86 processors" ?
Where is AMD gone? - Wasyu, on 10/10/2007, -0/+0Actually An AMD chip combined with a radeon just might produce said high performance low watt 3d graphics which right now is nothing more then a dream and with intel's previous performance in producing 3d chips don't count on them ever doing it.
We'll see some Taiwanese or Chinese company produce a good low power mobile graphics chip before intel . - christianw, on 10/10/2007, -1/+1what about this article makes "the [H]" great? posted a picture of a slide, told you what the slide said (which you could get by reading the slide in the first place), and placed their banner ads on it. "the [H]" as you call it is the most biased, unprofessional, ignorant, and talentless group of "tech journalists" on the net today. i place them at the bottom just below the inquirer for accuracy. if you want REAL tech news, check ars technica or anandtech.
kyle bennet and steve sold out just like tom from toms hardware. although i wonder what the size of the paychecks he and steve are getting to bash apple on their front page daily are. Maybe they arent getting paid for that though, maybe they just realize their content is so completely worthless innacurate and inconclusive (and most of the time, just a rehash of the digg tech index anyway), that they have to do that in order to get hits. kyle should be ashamed of himself, but he isnt. that guy is the karl rove of the tech industry. - Giga, on 10/10/2007, -0/+0Except that Hardware T&L doesn't really work yet on the X3000 as the drivers don't yet support that feature. Beta drivers are available, but the feature shouldn't be considered functional until it hits production drivers.
- inactive, on 10/10/2007, -1/+1Geez, rather than bragging about the future of AMD, they should start first how to cope up with C2D and their future as a company because no one give a damn about their current processor like the X2.
- Iam9376, on 10/10/2007, -1/+1Jack of all Trades.
Master of None. - xaxxon, on 10/10/2007, -2/+2This makes me wish it was a bbspot article. amd is going the way of cirrus logic (i.e. nowhere). Remember when cirrus used to make 486 clones?
- Wasyu, on 10/10/2007, -0/+0If we had only intel we'd get far worse then over priced egg cooker cpus we'd get crappy useless DRM chips just look at that pile of steaming dog crap called VIIv that we hear very little of now days.
- Wasyu, on 10/10/2007, -0/+0AMD has ATI they have good gpu technology Intel has shown they cannot make a good performing GPU to save their lives it took until the GMA x3000 to even get hardware transform and lighting a very basic GPU feature that has been around since the first geforce cards.
-
Show 51 - 73 of 73 discussions
Check out the new & improved 
© Digg Inc. 2009
— Content created and posted by Digg users is