Linked by Thom Holwerda on Tue 20th Nov 2007 16:46 UTC, submitted by Luis
AMD AnandTech takes a look at AMD's new Phenom. "If you were looking for a changing of the guard today it's just not going to happen. Phenom is, clock for clock, slower than Core 2 and the chips aren't yet yielding well enough to boost clock speeds above what Intel is capable of. While AMD just introduced its first 2.2GHz and 2.3GHz quad-core CPUs today, Intel previewed its first 3.2GHz quad-core chips. We were expecting Intel to retain the high end performance crown, but also expected AMD to chip away at the lower end of the quad-core market - today's launch confirms that Intel is still the king of the quad-core market."
Order by: Score:
low yields
by kittynipples on Tue 20th Nov 2007 17:47 UTC
kittynipples
Member since:
2006-08-02

There is a reason why Intel went the way they did with fusing two dual cores instead of trying to get a true quad core to market. AMD should stop worrying about marketing buzzwords and get something out that actually competes.

Reply Score: 2

RE: low yields
by kaiwai on Tue 20th Nov 2007 19:11 UTC in reply to "low yields"
kaiwai Member since:
2005-07-06

There is a reason why Intel went the way they did with fusing two dual cores instead of trying to get a true quad core to market. AMD should stop worrying about marketing buzzwords and get something out that actually competes.


Their 'true quad core' competes alright but what I would question is whether the 5% competive advantage is worth the amount of money spent, market share lost and lag between the Intel release and AMD's.

Unfortunately AMD is a company run by engineers who are unwilling to acknowledge when a cool idea is not worth persuing because the benefits promised cannot offset the risks it entails taking a more complex path.

That is why Sun is offering Intel workstations and servers; they've slowly started to realise that when push comes to shove, in terms of reliability of product delivery, its better to place your bets on Intel. Its the same reason why Apple went with Intel as well.

Reply Score: 3

RE[2]: low yields
by chipace on Tue 20th Nov 2007 20:00 UTC in reply to "RE: low yields"
chipace Member since:
2007-11-20

Unfortunately AMD is a company run by engineers who are unwilling to acknowledge when a cool idea is not worth persuing because the benefits promised cannot offset the risks it entails taking a more complex path.


The Athlon X2 didn't perform better than Intel's P4D, but that went on to be a huge success for AMD.

Phenom, for me, is about having four cores available yet paying the power price of a single core when in idle. Cool-n-quiet 2.0 has not been delivered yet, and I look forward to those power numbers. Supposedly, Barcelona and Phenom have 4 on-die PLLs and power islands to allow for this low power mode.

Also, AMD cpu motherboards are usually cheaper than Intel ones... this could affect sales.

Reply Score: 3

RE[3]: low yields
by rayiner on Wed 21st Nov 2007 00:10 UTC in reply to "RE[2]: low yields"
rayiner Member since:
2005-07-06

What are you talking about? The X2 _killed_ the P4D. It wasn't even a ballgame until Core 2 came out.

Reply Score: 8

RE[4]: low yields
by Phloptical on Wed 21st Nov 2007 02:36 UTC in reply to "RE[3]: low yields"
Phloptical Member since:
2006-10-10

...and then Core 2 buried X2, and still does, for cheaper, I might add.

It's unfortunate, but after being a staunch AMD supporter for years and years, I had to face the facts. Intel got it right in the Core 2, even over the Opteron....numbers don't lie. I think Quad Core is frivolous, so I haven't looked into it, but it sounds like AMD is still not competing at the same price point.

Reply Score: 1

RE[5]: low yields
by rayiner on Wed 21st Nov 2007 03:18 UTC in reply to "RE[4]: low yields"
rayiner Member since:
2005-07-06

Sure, Core 2 beat X2, but what does that have to do with the comment to which I was replying?

Reply Score: 2

RE[2]: low yields
by Javier O. Augusto on Tue 20th Nov 2007 20:02 UTC in reply to "RE: low yields"
Javier O. Augusto Member since:
2005-08-10

Unfortunately AMD is a company run by engineers...

LOL!!!!

No comments...

Reply Score: 5

RE[3]: low yields
by kaiwai on Wed 21st Nov 2007 01:04 UTC in reply to "RE[2]: low yields"
kaiwai Member since:
2005-07-06

Yeah, their marketing is working wonderfully - what was their profit last year..........?

Reply Score: 1

RE[2]: low yields
by javiercero1 on Tue 20th Nov 2007 21:42 UTC in reply to "RE: low yields"
javiercero1 Member since:
2005-11-10

"Unfortunately AMD is a company run by engineers who are unwilling to acknowledge when a cool idea is not worth persuing because the benefits promised cannot offset the risks it entails taking a more complex path. "

Quotes like this makes me wish there was a Heisman-like trophy for Armchair Quarterbacks...

Edited 2007-11-20 21:43

Reply Score: 7

RE[3]: low yields
by kaiwai on Wed 21st Nov 2007 01:06 UTC in reply to "RE[2]: low yields"
kaiwai Member since:
2005-07-06

Quotes like this makes me wish there was a Heisman-like trophy for Armchair Quarterbacks...


How so? they designed something that was needlessly more complex than their competition, delivered it late and it is barely competitive with the Intel offering. Explain to me how that engineering decision of making something more complex than required, was a good idea given the lateness and lack of competitiveness with Intel's offerings.

Reply Score: 2

RE[4]: low yields
by javiercero1 on Wed 21st Nov 2007 12:58 UTC in reply to "RE[3]: low yields"
javiercero1 Member since:
2005-11-10

Most of the issues for AMD's current predicament are not due to microarchitectural decisions, but rather process/fab issues.

The reasons for the stumbling that the 65nm AMD process is experiencing (process shared by IBM and Hitachi) are far deeper and more complex than most people in this thread know or have a remote familiarity with.

So I find it amusing to see people with passing knowledge to make statements that are so off the mark as to be laughable. For some reason, anyone who can assemble a PeeCee nowadays thinks he/she is an "expert" in the field and can figure out what a company is doing right/wrong.

Being inside the belly of the beast it is amusing to see some of the comments. Naive doesn't even begin to capture the nature of a lot of these posts. Provides a nice comedic relief though...

Edited 2007-11-21 13:01

Reply Score: 7

RE[5]: low yields
by h3rman on Wed 21st Nov 2007 14:31 UTC in reply to "RE[4]: low yields"
h3rman Member since:
2006-08-09

I need more modpoints to mod you up!
I liked the armchair-quarterback-prize idea too.
Come on people, what the f*** do any of us really know what's going on at Intel and AMD??

Now back to subject.. Sure, AMD has troubles.
The funny thing is that whatever *we* think about that doesn't mean sh*t. AMD chips will sell and they will make or not make some money for AMD. But AMD will survive or it will be bought by some of the other semiconductor giants (anyone been in Korea lately?). No need to worry about whether we'll still have any choice.

By the way, Intel's way of doing business is still immoral, they still bribe stores here in Holland and Germany so that they do not sell a *single* AMD-powered device.

Come over and have a look if you wish: huge store in Rotterdam, dozens and dozens of laptops and pc's on display, not a single one of them having an AMD chip. This while dozens of models by HP, Asus, etc. etc. do ship with AMD.

This makes me, yeah I know the Core 2 are the faster chips, not feel very great when I think of buying Intel, which is (Centrino) sort of the only laptop option if you want, like I do, blob-free Linux.

I therefore hope AMD is equally wicked. ;)

Reply Score: 2

RE[6]: low yields
by kaiwai on Thu 22nd Nov 2007 04:47 UTC in reply to "RE[5]: low yields"
kaiwai Member since:
2005-07-06

By the way, Intel's way of doing business is still immoral, they still bribe stores here in Holland and Germany so that they do not sell a *single* AMD-powered device.


Its immoral - so I guess you'll protest out the front of Coca Cola and many other large companies who pay funds to supermarkets to either go 100% their product or offer all the prime shelf space to their products.

Reply Score: 1

RE[7]: low yields
by Rugxulo on Thu 22nd Nov 2007 06:14 UTC in reply to "RE[6]: low yields"
Rugxulo Member since:
2007-10-09


Its immoral - so I guess you'll protest out the front of Coca Cola and many other large companies who pay funds to supermarkets to either go 100% their product or offer all the prime shelf space to their products.


Outside of Taco Bell (who I haven't been to recently), can you really name that many places that only sell Pepsi or only Coke?? I can't. It's pretty common to find both since they both are interesting in various ways (so many flavors). Same with Intel and AMD. And I guess people such as VIA (whose C7 sounds really really impressive, from what I read) are like root beer. :-)

*drinks a soda* Happy Thanksgiving! :-)

Edited 2007-11-22 06:14

Reply Score: 2

RE[7]: low yields
by h3rman on Thu 22nd Nov 2007 08:23 UTC in reply to "RE[6]: low yields"
h3rman Member since:
2006-08-09

Its immoral - so I guess you'll protest out the front of Coca Cola and many other large companies who pay funds to supermarkets to either go 100% their product or offer all the prime shelf space to their products.


I don't drink any Coca Cola, sorry. Actually I'm very rarely in the market for cola at all. ;)

But to go into what you said, are you implying that if I mention something illegitimate or immoral, I should simultaneously cite everything else illegitimate or immoral in the same genre? That is utterly absurd.
Yes, I do also oppose illegal warfare, violation of human rights, corruption, the widespread use of depleted uranium, the neocon fascists, the fascist Chinese regime, and all other criminals on the face of the planet. May we now proceed discussing AMD/Intel?

Reply Score: 2

RE[5]: low yields
by Nicholas Blachford on Wed 21st Nov 2007 22:57 UTC in reply to "RE[4]: low yields"
Nicholas Blachford Member since:
2005-07-06

>The reasons for the stumbling that the 65nm AMD process
>is experiencing (process shared by IBM and Hitachi) are far deeper
>and more complex than most people in this thread know or have
>a remote familiarity with.

You mean the process a number of companies (including AMD) are having some success with?

Barcelona is a new processor so some problems are to be expected, AMD probably expected problems so deliberately didn't produce many chips. IBM didn't appear to have any problems with POWER6 or Cell, both being produced on 65nm so I doubt it's anything to do with the process.

Reply Score: 1

RE[6]: low yields
by edwdig on Sat 24th Nov 2007 17:11 UTC in reply to "RE[5]: low yields"
edwdig Member since:
2005-08-22

Barcelona is a new processor so some problems are to be expected, AMD probably expected problems so deliberately didn't produce many chips. IBM didn't appear to have any problems with POWER6 or Cell, both being produced on 65nm so I doubt it's anything to do with the process.

POWER chips don't sell in very high volumes, so they're not a good judge of yield.

As for Cell, you've got the main core and 7 smaller cores. The chip is designed with extra smaller cores that get disabled at the factory to deal with manufacturing errors. That may be the reason AMD is/will be offering triple core processors, but with an x86 you can't really do things like that to anywhere near the same extent.

Reply Score: 1

RE: low yields
by butters on Tue 20th Nov 2007 22:07 UTC in reply to "low yields"
butters Member since:
2005-07-08

I look at the AMD/Intel landscape from a different perspective... Intel is the natural dominator in the x86 market. The only way that AMD can compete favorably is if Intel makes a huge strategic blunder. That happened with NetBurst, and it's unlikely to happen again anytime soon. Intel will always be 12-18 months ahead in process technology, they'll always have more fab capacity, and they'll always have more R&D resources to throw around.

AMD has consistently made wise design decisions to mitigate their inherent disadvantages. The Athlon K7/8 rested on two essential pillars: the on-die memory controller (IMC) and the serial, point-to-point system bus (HyperTransport). Intel didn't have them, and they still don't, because they don't need them. Because of their process lead, they can shove double the cache onto their chips, making them less vulnerable to bus performance.

This will change next fall with Nehalem, which will all of a sudden catch them up to AMD in terms of design sophistication. It will feature an IMC, a system bus called QuickPath (formerly CSI), and a unified L3 victim cache--just like Barcelona/Phenom. AMD needs to stay at least a year ahead of Intel on design because they're always a year behind on execution.

That's going to get a lot more difficult for AMD next year. I don't know how they're going to continue to remain competitive against Nehalem, especially after the tick-tock cadence takes this architecture down to 32nm in 2009. Intel has completely recovered from its NetBurst misstep, and they're catching up to AMD's design lead.

Then there's the battle between Fusion and Larrabee, where, IMHO, Intel has the better strategy. By treating graphics as a first-class citizen of the ISA, they're developing a true MPU with pipeline integration rather than just die/bus integration. Combined with HyperThreading (to which AMD has no answer), Intel could dispatch ALU and SP instructions in parallel through the same core.

AMD looks all kinds of screwed in their head-to-head competition with Intel. They desperately need to find a way to keep making x86 processors, but ones that don't compete directly with Intel's. Because what you see right now, with Barcelona up again Penryn, is the best it's going to be for AMD from here on out. Intel's going to pulverize them next year, and their new friends in Abu Dhabi are going to want them out of Intel's bread-and-butter markets.

Reply Score: 5

RE[2]: low yields
by psychicist on Tue 20th Nov 2007 23:12 UTC in reply to "RE: low yields"
psychicist Member since:
2007-01-27

What exactly is the importance for AMD of following Intel into an x86 death match in which it can impossibly ever compete, let alone win? I'd choose for making Intel EMT64 and AMD64 each their own architectures and create optimised OS's for either, at least in the longer term.

I've been working on an OS port to little-endian MIPS because it's free from the shackles of x86 in that it doesn't have to run Windows, is under no pressure to have a clock frequency high enough to make the bloated Vista seem fast and doesn't have any legacy software base to support (it's all free software anyway).

The advantages are streamlined RISC cores with low power consumption, low cost to manufacture since it doesn't have so many transistors and therefore doesn't have to be manufactured in the latest and greatest process, i.e. it can easily be done in 90 nm.

I'm looking forward to ARM (Cortex A9 and OMAP34xx), MIPS (Tilera64 and Loongson 2F/3) and PPC(Pwrficient and AMCC's Titan processors based on Intrinsity's Fast14 technology), because they don't even try to compete with Intel and AMD, but instead try to achieve what we were headed for 10-15 years ago.

That is, low power RISC processors with efficient UNIX implementations, except this time around it's called GNU/Linux. Wintel has set back computing for more than 10 years, but its domination is coming to and end.

The end result is the same software running equally well on all architectures, so you can choose the best one for your needs. If you need raw speed, choose x86 or POWER6. If you need low power (as most consumers do, particularly now that Vista is out) then choose ARM, MIPS, POWER (which is a superset of PPC nowadays) or another efficient one.

People are asking me to completely wipe Windows from their systems or even buy completely new computers with GNU/Linux. Not because they haven't tried but because they're tired after years of dissatisfying computing experiences.

They'd even choose AIX or Solaris on their desktops over Windows, if they had to. So what's the importance of x86 for most people nowadays compared to 10-15 years ago, really?

Additional info:

Linux 2.6.21.7 #1 SMP PREEMPT Wed Oct 17 04:27:14 CEST 2007 x86_64 AMD Hammer Family processor - Model Unknown AuthenticAMD GNU/Linux

Linux 2.6.21.5 #1 PREEMPT Mon Jul 9 23:00:10 CEST 2007 i686 Pentium III (Coppermine) GenuineIntel GNU/Linux

Linux 2.6.18.1 #2 Sat May 12 18:57:10 CEST 2007 mips GNU/Linux

I have a SPARC machine too and will have POWER hardware soon. They're all running the same software. Are AMD and Intel as important as they think they are or have they fooled themselves into thinking so and do they live by the grace of Microsoft and Windows?

Edited 2007-11-20 23:28 UTC

Reply Score: 3

RE[3]: low yields
by renox on Wed 21st Nov 2007 09:04 UTC in reply to "RE[2]: low yields"
renox Member since:
2005-07-06

>So what's the importance of x86 for most people nowadays compared to 10-15 years ago, really?

Depends on the market: while the market for Smartphone and other appliance has grown with RISCs CPUs, the server/workstation market went from RISCs and x86 to mostly x86.

Intel has announced that they will target low-power x86: so it's possible that x86 will gain market share from RISCs in embedded markets too: even Linux OSs works better on x86 than on other CPUs..

Reply Score: 2

RE[3]: low yields
by Downix on Wed 21st Nov 2007 14:17 UTC in reply to "RE[2]: low yields"
Downix Member since:
2007-08-21

@psychicist

I am saying much the same thing, but I'm focusing heavily on SPARC lately due to the ease of licensing, and the availability of open sourced chips. What is needed is the will to say "No more market domination" and push for diversity.

Reply Score: 1

Don't discount AMD just yet
by porcel on Tue 20th Nov 2007 22:26 UTC
porcel
Member since:
2006-01-28

AMD has been discounted time and again. Yet they offer the best bang for the buck in just about every market.

Most SME servers and desktops are already powerful enough. As a customer, I am interested in power consumption, stability and price.

Reply Score: 7

Multiple problems
by Nicholas Blachford on Tue 20th Nov 2007 22:40 UTC
Nicholas Blachford
Member since:
2005-07-06

AMD has a number of problems with Phenom, the biggest of course being the bugs which are keeping the clock speed right down. Unfortunately these months to fix.

The second problem is Intel have got 45nm with high K much sooner than tAMD. Being a process node ahead isn't that much of a performance leap these days but high K makes a *huge* difference.

Even if they fix the bugs Intel can just up their cock speed and stay ahead. AMD wont be able to respond until they have the same tech, and that could be some time away.

The other problems are more subtle - Intel, despite being dual core are still designing processors for fast single threaded performance. AMD's is clearly more designed for multithreaded performance. Single threaded software is much more common on desktops and thus can't take advantage of AMD's design.

Once AMD have the bugs ironed out and have 45nm tech and especially high K they'll be able to compete, If (and that's a big If) by then there's more properly multithreaded software around it'll show. - assuming AMD still exists by then.

The current tests appear to be on embarrassingly parallel tasks - tasks which won't stretch inter-core communications they completely ignore Barcelona's major advantage.

I suspect Barcelona will do rather better in multi-socket server benchmarks, their interconnect will be properly tested there. They were already in the lead there and Barcelona, even with it's problems should increase the lead.

Reply Score: 2

why the death knell?
by Rugxulo on Wed 21st Nov 2007 00:16 UTC
Rugxulo
Member since:
2007-10-09

Why such a rush to declare AMD "losing ground" and "needs to compete" and "might disappear". FUD FUD FUD! ;-)

No, seriously, AMD is quite good. Who cares if it's like a small percent slower than Intel? It's good enough. You act like Intel is the only decent processor, and that's not true. I don't see it as worth the extra money just to feed Intel's ego. We need competition (from more than just Intel and AMD, and I don't think I have to remind any of you of that).

Reply Score: 5

RE: why the death knell?
by psychicist on Wed 21st Nov 2007 06:03 UTC in reply to "why the death knell?"
psychicist Member since:
2007-01-27

I think most people here haven't seen anything other than Intel vs AMD, it looks like their lives revolve around it. Whereas in the real world people choose the right solution for their servers and desktops based on several criteria such as price, performance and power consumption.

AMD may be in a less favourable position now than it was a few years ago, but that doesn't mean their processors are worthless. They just have lower performance, which will probably be offset with lower prices like it is now with their K8 processors.

They are fast enough for most people with efficient operating systems. You only need to keep this contest going on if your world consists only of x86 and Windows. The rest of the people are enjoying their computers for some real work.

P.S. thanks for voting my other post down. Even though it was slightly off-topic, it was actually directed at butters because he's someone I can actually have a discussion with along with kaiwai, who's finally enjoying his MacIntel (and doesn't have much choice for a processor) and a few others. The rest of the posters I don't care much for actually.

Reply Score: 2

RE: why the death knell?
by Soulbender on Wed 21st Nov 2007 06:41 UTC in reply to "why the death knell?"
Soulbender Member since:
2005-08-18

Death knells sells, balanced facts doesn't.

Reply Score: 1

Price?
by cjcox on Wed 21st Nov 2007 04:53 UTC
cjcox
Member since:
2006-12-21

At least for now, price/performance, the AMD has Intel beat (by a significant margin). I wouldn't count AMD out just yet.

Reply Score: 4

RE: Price?
by irbis on Wed 21st Nov 2007 15:52 UTC in reply to "Price?"
irbis Member since:
2005-07-08

Yep, from customer point of view it is the price/performance ratio that counts (plus reliability, power consumption and such things), not so much who happens to have the fastest product on the market at the moment. Newer and faster processors are developed all the time, and the fastest products are usually all too expensive for most people.

Besides, when AMD seemed to be leading CPU development and Intel seemed to be a bit slow, it was in fact very active time in Intel development labs. The results show now. The same could perhaps be said about AMD/Ati now, and I'm sure they have lots of internal development going on nowadays.

Anyway, Intel has been doing very well in the CPU power consumption front lately, and that may actually mean a lot more for their business success than who happens to have a slight lead in the most recent CPU speed comparisons.

Reply Score: 2

the real AMD advantage
by unclefester on Wed 21st Nov 2007 10:17 UTC
unclefester
Member since:
2007-01-13

When you buy AMD you know you will be able to upgrade.

I built a new system a few weeks back. With AMD you just buy an AM2 board and know any AM2 cpu will work. Intel is about constantly changing chipsets, BIOS updates and general friggin' around.

AMD is a fraction slower. Who cares when my AMD 4000+ is norally running at 2 to 5 % cpu load.

AMD@ motherboards are often much cheaper too.

Reply Score: 4

i just want a cheap quad core
by jokinin on Thu 22nd Nov 2007 10:15 UTC
jokinin
Member since:
2005-11-07

I've been running this Athlon64 3400+ for almost 3 years now, and next summer i'll search for a new computer to replace it. What will i look for? As i am not a fool, i won't spend like 500 euro on the microprocessor, but more like 200 or even less if possible.
If AMD can give me a quad core for that price that has competitive performance with the intel equivalent, and lower power consumption i'll go for it, if not i will have to go for intel.

Reply Score: 1

RE: low yields.
by flywheel on Sun 25th Nov 2007 04:26 UTC
flywheel
Member since:
2005-12-28

Naturally IBM have problems with their Power6 and Cell processors. Both the Cell and Power6 is complex and a radical change when compaired to its predecessors.
The difference is that AMD is expected to perform a full production from day one, and if not everything is wrong with AMD, they are having major problems, their new processor suck.

Naturally AMD are having problems, just like IBM, VIA and Intel.

the funny part is, Intel never have to deal with that sort of spotligt (They also never have to deal with any form of negative past).

Reply Score: 1