Tuesday, May 15, 2007

Why is Intel QX6700 quad core a $125 CPU

Look at this picture. Michael Dell was excited, I heard, he is going to switch to near 100% AMD by 1Q08, just one quarter ahead of Intel's projected BK.

I think AMD's seemingly painful 1Q07 was Hector's plan. By taking a huge hit in 1Q07, AMD builds massive amount of inventory to flood Q2 -- those $59 X2 3600s just keep coming wave after wave. Then in Q3, K10 is out to push all of Intel CPUs down below $100 (except QX6700, which will be sold at $125). Intel's planned price cut in July is too little-- they only cut 50%, they need a 75% cut.

Intel's Penryn may see some clock speed increases, but those will be way short of narrowing the huge IPC gap. Also keep in mind, AMD is migrating to better SOI transistors in 65nm, which will result in 40% clockspeed boost.

Rahul thinks AMD made Intel stronger than ever. How ridiculously wrong. He can't see that Intel is at its last breath. Every sign tells us that Intel is doomed. The reason is simple. AMD has become a major player in server, desktop and mobile and has sufficient capacity. All AMD needs is one killer chip and Intel will be mortally wounded. This is because Intel is huge animal which needs a lot of food to survive, once AMD cuts Intel's revenue by 50%, Intel will die within two quarters, insufficient time for making any adjustments. The situation for Intel is unlike others. SUN always has this Sparc niche to enjoy its meal, so it can last a long time dying slow. Intel and AMD CPUs are software and hardware (except MB) compatible, and AMD has become a smarter choice.

Now, you ask, why should Intel die?

The reason is simple: Intel has been lagging behind AMD for five generations for four years, and there is no sign of catching up. Intel is behind on (a) virtualized 64 bit computing (b) IMC (c) direct IPC (ccHt) (d) direct Core-Core communication (e) virtualized HyperTransport I/O.

The Core 2 architecture somehow reduced the gap in 64 bit and Core-Core communication. But it was a kludge, the solution can't be extended to quad core. Intel's 64 bit architecture is incapable of IOMMU and now AMD made it to IOMMU 1.2.

The K10 will widen AMD's five generation architecture lead.

47 Comments:

Blogger PENIX said...

Those results are pitiful. I wouldn't even pay $125 for garbage like that.

8:28 AM, May 15, 2007  
Blogger Unknown said...

This comment has been removed by the author.

8:40 AM, May 15, 2007  
Blogger Unknown said...

wakakaka... Phd, I think you should say "why is intel qx6700 QC a $12.50"

hehehe..

8:40 AM, May 15, 2007  
Blogger R said...

I stated in an earlier post their can be only one of two possible conclusions to the launch of the K10, either it lives up to its claims or it doesn’t. Soon one side of the argument is going to look very silly. Words mean things and a lot of fanboies are in deep. However enthusiastic brand loyalty is as American as apple pie. Your passion is why I keep coming back to this blog.

8:45 AM, May 15, 2007  
Blogger Hornet331 said...

http://www.sweclockers.com/forum/showthread.php?s=&threadid=680996

ops....

nice sharikou that you post fake screens as fact, that screens where intent to punk fudzilla and even he pulled his "story" after a few hours...

he even admint that he knew they where fake:

http://www.xtremesystems.org/forums/showpost.php?p=2191373&postcount=192

9:14 AM, May 15, 2007  
Blogger Unknown said...

AMD is more desperate than ever before to fabricate data. They know K10 will be a flop like R600. Clovertown is so advanced it has prefragged everything on AMD's roadmap.

9:20 AM, May 15, 2007  
Blogger Unknown said...

Clovertown is so advanced it has prefragged everything on AMD's roadmap


What i know, the colovertown cant even compete the old 90nm AMD 8222SE.. Wakakakakkaa... Congratulations, You are the stupidest donkey among the stupid Intel donkeys in the world.

9:44 AM, May 15, 2007  
Blogger Evil_Merlin said...

All the poor little AMD fanbois are in a tizzy!

How does it feel to have been punked?

How does it feel to be yet more of a laughing stock?

10:45 AM, May 15, 2007  
Blogger Tsukune said...

What kind of drugs are you on and what fantasy world do you live in?

10:50 AM, May 15, 2007  
Blogger Unknown said...

You guys are missing The Big Picture:

(1) Sell-side firms have reversed on AMD. Several have gone from SELL to BUY...OVERNIGHT !!

(2) GPU high-end market is not key for AMD/ATI. Much less important than server/desktop high-end. Nvidia can have that market, AMD wants the channel, OEM's, and next-gen 'Fusion' type sales.

(3) AMD should win $5-$20/per share ($3 - $10 billion) from Intel lawsuit. Figures could be 3X if anti-trust comes in.

18 months from now, AMD may double from much higher levels within a week or so on the lawsuit alone!

11:44 AM, May 15, 2007  
Anonymous Anonymous said...

Why just FUD when you can FUDZilla!

Former Inquirer writer Fuad Abazovic released "benchmarks" of the upcoming AMD quad-core Barcelona. Abazovic believes his benchmarks are "an AMD exclusive" of a new processor he has in his possession.

Unfortunately, not is this tale completely fabricated, but the purpotrators who architected the hoax came forward early this morning on Swedish enthusiast forum, SweClockers.com.

The poster "2good4you" details the process he used to convince Abazovic that his benchmarks were authentic -- which essentially entailed a single email with the image attached. Abazovic then republished the benchmarks as his own, and not for the first time alludes "2good4you."

Rest assured, real Barcelona benchmarks will arrive shortly. AMD typically releases two spins of silicon before retail samples, EVT followed by DVT. As of May 2007, several EVT samples are floating around, though it's difficult to predict the actual retail performance of K10 via these engineering samples. AMD still had serious bugs in its EVT samples before AM2 launched, and the only real change presented on that architecture was the DDR2 memory controller -- for Barcelona we're talking about a dramatically new design.

With Computex coming up in less than a month, it's fairly reasonable to expect a flood of authentic benchmarks to make waves just after or during the show. And yet even then, don't be so quick to believe the Fear, Uncertainty, and Doubt until we start seeing DVT and Retail samples.

Update 05/14/2007: As of noon today, Abazovic removed the article without an official statement. Abazovic declined to speak to DailyTech about the matter.


This is the only way AMD gets around, fabricating data and paying off people to "release" these benchmarks.

AMD BK 2Q'08

11:56 AM, May 15, 2007  
Blogger abinstein said...

"This is the only way AMD gets around, fabricating data and paying off people to "release" these benchmarks."

Seems more like an silent retraction caused by NDA violation...

but we're and probably will never know for sure, of course.

12:29 PM, May 15, 2007  
Blogger Hornet331 said...

sorry poke but i dont see how amd is involved in this incident, its just shows on what level fuad operates, every news (even if its fake) that will generate site traffic for the quick bug, is good enough for him.

I dont think amd payed anything, the guys of sweclockers just made fun of fuad and he was dump enough to swallow it, as well as some hardcore fanbois that live in there "INTEL BK 2Q'08" world. ;)

12:57 PM, May 15, 2007  
Blogger Unknown said...

pezal
wakakaka... Phd, I think you should say "why is intel qx6700 QC a $12.50"
when its at $12.50 for a quad core, consider how many people will flock out to buy them?

just as Mr. Rahul Sood says, AMD needs a new way to do business. Have the best chip but selling at $999 a piece isn't going to beat a lesser chip, but at $266 a piece.

12:57 PM, May 15, 2007  
Blogger Unknown said...

Denial denial. What part of fudzilla don't you get?

12:59 PM, May 15, 2007  
Blogger Unknown said...

I would hold off judgment too soon!
Realistically if AMD's 2900 is going to be a killer at 65nm, and if those pics of Phenom can be turned into quantity supply and the motherboard "700" chip set tips up real soon in volume, " I'm already installing volume 690 pc's" Then maybe Intel will feel some heat.. the perception
needs silcon to back it up?

1:04 PM, May 15, 2007  
Blogger Unknown said...

Whoops!

US Govt. opens patent infringement investigation into AMD:

yahoo.reuters.com/news/articlehybrid.aspx?storyID=urn:newsml:reuters.com:20070515:MTFH85655_2007-05-15_20-52-58_N15450766&type=comktNews&rpc=44

2:21 PM, May 15, 2007  
Anonymous Anonymous said...

"Seems more like an silent retraction caused by NDA violation..."

More AMD fanboy denial. Typical.

5:10 PM, May 15, 2007  
Blogger Ho Ho said...

bluevettekid
"GPU high-end market is not key for AMD/ATI. Much less important than server/desktop high-end. Nvidia can have that market"

Did you know that NV earned more revenue from its from professional Quadro cards than entire ATI/AMD graphics lien put together? Considering these GPUs are sold at >>$1000 the profits earned from them are redicilous.


abinstein
"Seems more like an silent retraction caused by NDA violation..."

For ex-inquirer worker that would be the first time. You can bet your life on it that this was not removed because of NDA.

5:15 PM, May 15, 2007  
Blogger Unknown said...

Hey Moron, Why don't you go see a good shrink immediately

5:44 PM, May 15, 2007  
Blogger Unknown said...



What i know, the colovertown cant even compete the old 90nm AMD 8222SE.. Wakakakakkaa... Congratulations, You are the stupidest donkey among the stupid Intel donkeys in the world.


False. Clovertown is 79% faster than Crapteron.

http://tweakers.net/reviews/661/4

5:45 PM, May 15, 2007  
Blogger DaSickNinja said...

Penix
"Those results are pitiful. I wouldn't even pay $125 for garbage like that."

Kids, don't do drugs. Or else you end up like this moron. *Points upwards to the farce masquerading as intelligent life*

7:01 PM, May 15, 2007  
Blogger Randy Allen said...

AMD is also currently being sued by SGI, Opti Inc and Microunity. Given AMD's dire financial state and the fact that they are bleeding a lot of cash this is just more bad news for AMD.

Did you know that NV earned more revenue from its from professional Quadro cards than entire ATI/AMD graphics lien put together? Considering these GPUs are sold at >>$1000 the profits earned from them are redicilous.

Wow. That's a lot of cash. I had no idea the high-end workstation market for Nvidia was so successful and profitable.

It's clear that anyone wants high end graphics will go with Nvidia. Whether for high end rendering or high end gaming Nvidia is the smarter choice. AMD's graphics products is all junk. They don't even have Linux drivers for the 2900. Pathetic.

AMD BK Q2'08.

8:07 PM, May 15, 2007  
Blogger Unknown said...

False. Clovertown is 79% faster than Crapteron.



Wow... the latest 65nm 4 cores Intel Tortoise 3.4GzH totally get frogged by AMD 8220SE 2.8Gzh.. In fact, the 2 cores Intel Colovertown is much better than the Tortoise one. However, both of the Intel Colov & Tortoise still no match to the fastest old 90nm AMD Opteron 8220SE..

9:38 PM, May 15, 2007  
Blogger abinstein said...

"You can bet your life on it that this was not removed because of NDA."

This is absurd. Why would I bet anyone's but your life on such a presumptious claim like this?

Prepare to kill yourself if this turns out to be caused by NDA violation (and hope you're not a catholic) :-p

10:39 PM, May 15, 2007  
Blogger abinstein said...

"More AMD fanboy denial. Typical."

The whole response here actually seems more like Intel fanboy denial.

K10 is going to perform great... but I know none of you are interested because you've got Clovertown, which is really a Core 2 "3.3" but in your minds it prefrags everything without Intel inside.

It's okay... as long as you're not one of my employees. :-D

10:45 PM, May 15, 2007  
Blogger Ho Ho said...

abinstein
"Why would I bet anyone's but your life on such a presumptious claim like this?"

Sorry, that was a typo, I meant my own life.

Btw, are you willing to bet on something when resulst from that specific benchmark are considerably different than on that screenshot? Say 20% or bigger difference at same clock speed.


"K10 is going to perform great"

Considering we have no benchmark data how can you say that? Just a couple of months ago 2900XT was supposed to obliterate GTX also.

To say the least my oppinion about AMD has dropped considerably during last half a year or so. They were doing fine when they were in the lead but now as they have dropped behind it has all fell apart.

12:24 AM, May 16, 2007  
Blogger Ho Ho said...

Another fun thing about R600. In theory, it has more than twice the shader power of GTS and roughly 65% more memory bandwidth but it still can barely compete with it.

2:16 AM, May 16, 2007  
Blogger Tanrack said...

Why X6000 is a $5 chip. It is 5 generations ahead acording to Sharikou and it is still slower than the ancient cache bloated c2d. Maybe with K10 widening the gap AMD can be more on par with performance and give us some healthy industry competition.

6:29 AM, May 16, 2007  
Blogger Christian Jean said...

hoho said...
Another fun thing about R600. In theory, it has more than twice the shader power of GTS and roughly 65% more memory bandwidth but it still can barely compete with it.


Do your homework hoho and you'll realize why!

When you introduce something new, you can't have full support for it.

This card is a lot more advanced than a traditional video card.

64-bit took a long time, just like writing the drivers/software for the R600 will.

AMD will rely heavily on the open-source community for this task!

The R600 is the first step of the first phase (of 3 phases) to achieve Fusion!

6:46 AM, May 16, 2007  
Blogger Evil_Merlin said...

Um, no, no one ever said the K10 wasn't going to perform well. Hell, I don't see that anyplace in the past few months.

Intels current generation of chipe perform better than AMD's current generation of chips and it's next generation is going to perform great... but I know none of you are interested because YOU ARE A FANBOI.

7:12 AM, May 16, 2007  
Blogger Ho Ho said...

Errr, I've done more homework than pretty much anyone else in this blog. I've kept a close eye on GPU architectures starting from FX/R9700.

G80 is considerably more efficient thanks to its fully independent scalar shader units. R600 has two types of them, one complex for every four simple ones. G80 has one complex and one simple for each and every of its 128 SPs. That means when you have lots of complex instructions in your shaders G80 will mop the floor with R600.

Also the entire architecture of R600 is not that nice. FX was VLIW also, so is Itanium. We all know how well those work and how much are they depending on good compiler technologies.

As for open source community, is that the new AMD way to cut costs? Sure, they can finally have some basic 3d drivers correctly working under Linux but hoping they would create well optimized ones for windows is a pipe dream. Intel opensourced their drivers a long time ago and during the years very little has been made in terms of performance. R600 will long be history by the time half-good drivers hit the streets.

7:13 AM, May 16, 2007  
Blogger pointer said...

Jeach! said...

When you introduce something new, you can't have full support for it.

This card is a lot more advanced than a traditional video card.

64-bit took a long time, just like writing the drivers/software for the R600 will.


I thought they have claimed to have best Windows driver and yet ... and you also implied NVIDIA driver support is better than AMD's here? ...


AMD will rely heavily on the open-source community for this task!


I have never heard that they open sourced the Linux driver. Are you sure on this? In comparison, NVidia has much better open source support than AMD's (anyway, this is comparing 1 to 0, because AMD never have one open source graphic driver)

4:14 PM, May 16, 2007  
Blogger Ho Ho said...

pointer
"I thought they have claimed to have best Windows driver and yet ."

They have said lots of things. Havin better Windows drivers is just one of those things. They also promised hard launch of R600 at all pricepoints just a couple of weeks ago.


"I have never heard that they open sourced the Linux driver."

You should read more news :)


Of cource that is only for Linux, it doesn't help Windows at all.


"anyway, this is comparing 1 to 0, because AMD never have one open source graphic driver"

Their driver for older than X800 series driver is in fact opensource and has zero official support.

10:34 PM, May 16, 2007  
Blogger pointer said...

Ho Ho provided links to AMD's linux driver plan

well, it is a plan, with a word WILL ... means not yet right? :) and look at their recent track of record when they said they are going to release their GPU ...

11:16 AM, May 17, 2007  
Blogger Christian Jean said...

hoho said...
Errr, I've done more homework than pretty much anyone else in this blog.


Ok there Einstein! I wouldn't go that far in your claim.

As for open source community, is that the new AMD way to cut costs?

And if it was... what the hell is wrong with that? They are a hardware company after all, why not 'outsource' the drivers!

Sure, they can finally have some basic 3d drivers correctly working under Linux but hoping they would create well optimized ones for windows is a pipe dream.

If your even implying that open source coders couldn't optimize drivers better than 'paid' programmers... your dreaming!

Intel opensourced their drivers a long time ago and during the years very little has been made in terms of performance.

The performance problems had nothing to do with software or drivers :)

11:59 AM, May 17, 2007  
Blogger Ho Ho said...

pointer
"well, it is a plan, with a word WILL ... means not yet right? :)"

Yes it that what it means. I also should have mentioned that ATI promised to make their Linux driver working ... over four years ago.

jeach!
"Ok there Einstein! I wouldn't go that far in your claim."

Considering how much false information I've read here I think I'm not too far off with my claim.


"And if it was... what the hell is wrong with that?

Only what I said: it won't help Windows.


They are a hardware company after all, why not 'outsource' the drivers!"

You know that without the software hardware is meaningless. If they would start relying only on opensource drivers then they would be pretty much doomed.


"If your even implying that open source coders couldn't optimize drivers better than 'paid' programmers... your dreaming!"

Where did you read that one out? I was only trying to say that there won't be an opensource driver for Windows. As you can pretty much say that the people who make commercial programs are the same who make open source ones there is no difference in their code quality. Though you could say that with lot more eyeballs viewing the code opensource has an edge.

Another point I made was that by the time opensource creates half-decent drivers R600 will be history.

Btw, have you ever compared opensource ATI drivers and closed source ones? If not then just google a bit and find out yourself. Those drivers have been opensource for a long time already and ATI itself has stopped working on their binary drivers of the same time so you can't really say that they have continued developing them in parallel.

Btw, ATI had a bug in their Linux drivers that halted your computer when you tried to use AA. IIRC that bug was known to exist well over a year* and throughout several official driver releases before it got fixed.

*) It was there when I started using a Radeon around two years ago and it was still there when I last looked their readme around a year ago.

"The performance problems had nothing to do with software or drivers :)"

What I meant was that performance compared to the first more or less working drivers haven't changed a lot. You can optimize drivers but you can't do it forever. I expect that in best case scenario after fixing the obvious bugs AMD can get additional 5-10% performance out of R600 thanks to optimizing their drivers.

3:21 PM, May 17, 2007  
Blogger Unknown said...

The QX6700 might just be a $125 processor by the time AMD gets Barcelona out the door. ;-) Motherboard suppliers are now reporting that AMD is delaying Barcelona from June to August or possibly September. The K10 has so many bugs in the EVT release that they will have to respin it at least twice now. Early reports have Agena moving from December to February now due to the Barcelona delay. K10 production is only expected to account for less than 8% of AMD's volume by the time Agena is launched. Tell me once again how Intel will be BK by Q208 if AMD cannot even ship their "Killer" core in enough volume to make a dent in their overall processor sales for at least the next year?

4:57 PM, May 17, 2007  
Blogger Joshua Smith said...

Wow, your totally deceived.

5:48 PM, May 17, 2007  
Blogger Ho Ho said...

Talking about CPU prices, after 22'th July AMD fastest CPU will be fighting against $163 Intel CPU. How does that affect its margins and profitability?

2:00 AM, May 18, 2007  
Blogger Unknown said...

Intel's price cuts will totally finish AMD off once and for all. 6000+ will be forced to be sold at sub $150 prices. AMD's BK is inevitable.

AMD BK Q2'08.

5:14 AM, May 19, 2007  
Blogger Randy Allen said...

Also keep in mind, AMD is migrating to better SOI transistors in 65nm, which will result in 40% clockspeed boost.

You keep saying this, but do you have any proof of this? Right now, AMD's fastest CPU is at 3GHz on 90nm. It's fastest 65nm CPU is 2.6Ghz 5000+.

AMD is clearly have troubles with 65nm. Why else would they have such a low clockspeed as the highest performing 65nm part?

10:31 PM, May 19, 2007  
Blogger Ho Ho said...

I guess he is only saying that by the time AMD is moving to 45nm its 65nm CPUs will be up to 40% higher clocked than its current 90nm ones. Of cource those 45nm ones will be at midrange speeds of 65nm, at best.

1:11 AM, May 20, 2007  
Blogger Scientia from AMDZone said...

jimmie

"Early reports have Agena moving from December to February now due to the Barcelona delay. "

Where did you get this silly idea from? AMD's last statement from Rivet was that Desktop X4's would be ready for Christmas.

8:06 PM, May 25, 2007  
Blogger Scientia from AMDZone said...

randy allen

"Right now, AMD's fastest CPU is at 3GHz on 90nm. It's fastest 65nm CPU is 2.6Ghz 5000+."

There have been reports of 65nm X4's clocked to 2.97Ghz.

"AMD is clearly have troubles with 65nm. Why else would they have such a low clockspeed as the highest performing 65nm part?"

Clearly you aren't familiar with AMD history. AMD always does this. At introduction, the new process is normally a little slower than the old. However, the new process keeps getting faster and passes the old. This is in contrast to Intel which has not increased C2D dual core speed in a year.

8:10 PM, May 25, 2007  
Blogger Scientia from AMDZone said...

ho ho

"Talking about CPU prices, after 22'th July AMD fastest CPU will be fighting against $163 Intel CPU. How does that affect its margins and profitability?"

If it were only cpu's then you would have a good point. However, Intel has not yet announced anything competitive with mini-DTX. This should allow AMD to compete even with higher chip prices. Essentially, mini-DTX should make it impossible for AMD to lose anymore share in the lower half desktop market after mid year.

8:15 PM, May 25, 2007  
Blogger Scientia from AMDZone said...

sharikou

Do you know of any reason why IOMMU could not be included in Nehalem?

8:17 PM, May 25, 2007  

Post a Comment

<< Home