Monday, June 12, 2006

AMD to chainsaw Intel

This is huge. Athlon 64 X2 3800+ to be sold at $169. That chip can beat Pentium 965 XE. Such cuts will definitely deny Intel the oppurtunity to unload its legacy inventory. I bet Hector is calling Chartered to load more wafers.

One may wonder what will happen to AMD's revenue and profits. You have to keep in mind that AMD's ASP is about $100, which means most of AMD's units are Semprons. Now with the price cut, it may increase the units on Athlon 64 and X2. In other words, the mix changes. So the net effect on ASP is less. Coupled with fast expanding capacity, I expect AMD to achieve sustained revenue growth and unit share gain.

By 4Q06, AMD's capacity will double from 1Q06 level. I expect AMD to exit 2006 with 40% market share.

The first victim of this AMD-INTEL death struggle will be Intel. The second will be DELL. With Intel's price crash, Dell will have close to zero advantage over competitors. Dell's revenue will fall because of lower unit price. To survive, DELL will have to go AMD across the board to maintain revenue growth by increasing its market share.

Intel will simply walk out 2006 dead.

55 Comments:

Anonymous Anonymous said...

Where exactly did you get the exact price drop on this particular chip? I Couldn't it be a best case scenario?

9:59 PM, June 12, 2006  
Anonymous Anonymous said...

wow, dual core x2 for under 200, would such cuts make it to the "energy efficient small form factor" line as well. if that is the case, this christmas will be very sweet.

dream pc

amd am2 3800 x2 energy efficient small form factor 35w

asus am2 x16 workstation mb 50w

nvidia 7950gx2 card 180w

4 x 533 ddr2 ecc ram 40w

1 15k scsi boot hd with pci-x ctl card 35w

4 x data raid 5 hd with pci-x ctl card 100w

total power consumption 440w, even the antec phantom 500 silent psu will have enough juice left.

sunhing

10:11 PM, June 12, 2006  
Anonymous Anonymous said...

Digitimes are now saying it is going to be less than 50%

http://www.digitimes.com/mobos/a20060612PR208.html

We'll have to see what shakes out. This will hurt AMD like it hurts Intel but that is always the case in a price war. Of course one section that is not hurting is us :) Interesting times ahead.

10:26 PM, June 12, 2006  
Anonymous Anonymous said...

With all due respect Sharikou you should have given the link yourself. Nonetheless here it is: http://www.dailytech.com/article.aspx?newsid=2800

Poster #2 that should make for a killer rig :) but personally I'd wait on 4X4 and see how it pans out.
Also I have to wonder, how did you come by the power consumption stats for all those parts?

10:28 PM, June 12, 2006  
Blogger Sharikou, Ph. D said...

This will hurt AMD like it hurts Intel but that is always the case in a price war.

AMD must prevent Intel from converting its legacy P4 chips to cash at all costs. Currently, Intel has $7 billion worth of legacy chips as inventory.

10:29 PM, June 12, 2006  
Anonymous Anonymous said...

respond to watt question:

i am currently using roughly the same components as my dream system. it is using a 550watt psu on asus mb

current system:
dell precision 380
pentium d 820
intel 955x mb dell oem
4 x 1 gb ddr2 533 ecc ram
15k u320 scsi
nvidia 7800gtx 256
2 x 200gb sata wd hd on separate pci-x card
current psu is 385watt under full load

since my current 7800 gtx use 80 w
2 x 7800 give 160w. so 2 x 7950gx2 should use 180w max

removing sata array with card save about 65w measured

removing scsi 15k with card and replaced with sata 40g hd boot save 25w measure (assume 40g sata boot hd use 10w)

this is how the numbers roughly adds up to my approximation in 2nd comment.

remaining components are cpu, heatsink, fan, mb etc. given that 955x chipset use 14w alone, the mb should consume 50w max.

cpu is probably the most power hungry component in my system right now

ddr2 533 ecc ram actually use lower voltage than ddr 400 ecc, so it should deliver about 10w max per 1gb module since they are not reg or fbdimm.


i hope i answered your question.

sunhing

10:47 PM, June 12, 2006  
Anonymous Anonymous said...

dream pc2:

amd am2 3800 x2 energy efficient small form factor 35w (passive cool)

asus am2 x16 5900sli standard mb 50w passive cool)

2 x nvidia 7950gx2 360w (active cool)

1 samsung or segate hybrid hd sata 10w as boot and data (200g plus)

total power consumption 455 w, still within the antec 550 trupower 2.0 of my antec titan case domain without burning the house down. however, i might try the silent antec 500 phantom to see how it works out.

sunhing

10:54 PM, June 12, 2006  
Anonymous Anonymous said...

4x4 makes sense now, consumers will have more reason to get two now cheaper cps, AMD will benefit because at least they will be selling, Intel will be left to cry on the side lnes boy do they need to sack some one.!

10:56 PM, June 12, 2006  
Anonymous Anonymous said...

the slimest SLI like DC x2 am2 system that i can think of is:

amd am2 3800 x2 energy efficient small form factor 35w

asus am2 standard mb with only 1 pci-e 16 slot occupied 40w (asus m2n32-sli series for example, assuming each additional pci-e 16x slot should pull 10w with device installed)

2 x ddr2 800 ram 20w

1 x nvidia 7950gx2 card 180w

1 x hybrid hd sata 10w

total power consumption 285w and should not exceed 320w max.

personally i always use ecc memory since the release of windows nt 3.51 due to stability concern.

now my dell precision has been running 76 days without reboot (using ddr2 533 ecc ram)

and my amd x2 asus custom system run 86 days without reboot (using 4 x 512mb crucial ddr400 ecc)


when the amd am2 3800 x2 energy efficient small form factor is released, i am ready to savage my dell precision 380 into a new am2 system base on workstation configuration.

;>

sunhing

11:11 PM, June 12, 2006  
Anonymous Anonymous said...

"AMD must prevent Intel from converting its legacy P4 chips to cash at all costs."

At all costs might be a little extreme since it can't mean good things for AMD.

In any case, did you notice which processors are receiving the price cuts? It appears to only be the 512kb models, which I guess means they aren't completely stupid. They're discounting the chips that are cheaper to make.

Still, you always talk about the complete breakdown of Intel's price structure, but AMD looks even worse. Intel at least still has a decent logical increased price for increased performance progression. AMD now has the X2 5000+ selling for less than the 4400+. This will wreck havok with product positioning and marketing. It also seems like the energy efficient processors aren't discounted which may impact the competitiveness against Conroe depending hope well they actually shape up power-wise.

In regards to the X2 3800+ being a killer bargain at $169, I think we still need to hold out for that. The lowest Conroe the E6300 is priced at $183 which makes it very appealing considering it's "new technology". Even the 945D is competitive since at 3.4GHz it holds up well performancewise against the X2 3800+.

http://www.techreport.com/reviews/2006q1/pentium-xe-965/index.x?pg=4

You picked the worst showing for the 965EE in your homepage, but even in gaming the 3.4GHz 950D which is really the same as the 945D performancewise, is comparable to the X2 3800+.

With the C1 stepping, it can no longer be said that Netbursts are power hungry because they've caught up with K8 in terms of power consumption.

http://www.techreport.com/reviews/2006q1/pentium-xe-965/index.x?pg=14

In the review you linked we see the C1 stepping 3.73GHz 965EE pulling in similar power consumption numbers to the 2.6GHz FX60.

All the X2 3800+ does with its price point is match the 945D and the E6300 because it does not look like it'll have any major benefits in performance or power consumption.

If you are pointing to the AM2 platform bringing benefits in power consumption that too is overrated.

http://www.techreport.com/reviews/2006q2/socket-am2/index.x?pg=16

We see slight power reductions in the AM2 X2 4200+ and the AM2 X2 4800+ over their S939 counterparts when idle with power management. However, under load the AM2 X2 4200+ actually uses more power than the S939 version with the X2 4800+s tying. The 950D and 960D numbers actually aren't accurate because they are actually from the 3.73GHz 965EE. The article specifically mentions that real Pentium Ds would have lower voltages and other programming optimizations to control their power consumption. It's also interesting to note that the 965EE uses less power than the FX62 under load. Certainly quite the change. It's probably safe to say that unless you specifically buy a non-price reduced Energy Efficient model, the power consumption characteristics of the AM2 processors have actually worsened rather than improved.

11:43 PM, June 12, 2006  
Anonymous Anonymous said...

"Definitely deny Intel the oppurtunity to unload its legacy inventory"

I amm bit unsure here ..who are the crazy people who will buy a P4/netburst now

Are we talking about the Pentium-D and Celeron Heaters dumped by Dell
??

12:10 AM, June 13, 2006  
Anonymous Anonymous said...

Hmmm, something big is missing.

Have you guys think of having a mico-heater in your room or bed in the winter ? New products will be debuted soon:

1. Pentium-D heater.
2. Celeron personal heater.

The first one is my dream hardware this X'mas.

2:13 AM, June 13, 2006  
Anonymous Ixse said...

Sharikou, re: By 4Q06, AMD's capacity will double from 1Q06 level. I expect AMD to exit 2006 with 40% market share.

That's a bit naive. Rev F is 20% larger than rev E. This transition is starting right now. Only for this transition AMD needs more capacity already.

Secondly the mix of DC/SC will need to change drastically to keep up with Intel between now and eoy. DC's are almost twice as big as SC's and will hence take a lot of capacity.

AMD is due to start 65nm shipments Q4. In this industry this typically means END of Q4. Even then Chartered and fab 30 continue shipping 90nm for all Q4. So even though the transitioning to 65nm will start in Q4 the effect on units isn't all that big yet. So it's a given that in H2 by far most of the capacity increases brought by fab 36 and fab 7 (Chartered) will go to the above mentioned events (rev F being 20% larger, and DC being roughly twice as big).

So it's impossible for AMD to exit '06 with 40% share as per your prognosis.

If you're still not convinced consider looking at this from an Intel point of view. NGA comes in above K8 and that will help them to keep margins at a level they can affort to pay for even while lowering all Netburst prices to near Celeron levels. Netburst competes with say 80% of the K8 range; with new 90nm transistors in Q3 (better strain) K8 will competes with about 50-75% of the NGA range. NGA/Netburst put K8 in a performance sandwich. Otellini stated last Q1 earnings CC that he'll keep factories at full production even if it means sacrificing ASP. Take all of this together and it's hard to escape the conclusion that:
- Intel needs full factories to maximize revenues and can do so by Celeronizing Netburst while ramping NGA as fast as they can. Intel's price cuts (both past and coming) provide evidence for this trategy.
- In order for AMD to maximize revenue they'll need to focus on the upper part of their current product ranges, which means increasing die size significantly more than unit volume (%-wise). So by far most capacity increase will go to larger dies, not unit volume. Those larger dies will sell at significantly lower prices as off July 24 (day after Conroe intro).

Basically there's not even a sliver of a chance that AMD will achieve 40% unit market share eoy (let alone rev share). This statement is btw in line with Ruiz comments that they aim for 30% market share by ~2008.

Regards,

Ixse

2:29 AM, June 13, 2006  
Anonymous Anonymous said...

"..who are the crazy people who will buy a P4/netburst now

Are we talking about the Pentium-D and Celeron Heaters dumped by Dell
?? "

They'll probably dump it on the 3rd world where people unknowingly will buy it cause it's CHEAP!

Then they'll go online and find out that they bought space heaters and can't afford the electricity... :)

7:07 AM, June 13, 2006  
Anonymous Anonymous said...

"I amm bit unsure here ..who are the crazy people who will buy a P4/netburst now..."

Well, I would. 2 months ago I had to rebuild, new cpu and motherboard. I already have good equipment, but for the price I bought an Intel 651 (65nm, 3.4Ghz).

Come July, to be able and get a 950 for $250-$280 for me its a good deal, and a quick jump to a decent dual core. Otherwise $200, new motherboard, $530 new cpu, would really want to upgrade ram from 667 to at least corsair 8500 $350+/-.

I know it will cost more in the long run, but Im gonna wait for the new ATI board, memory prices to drop a little, and again ATI's new video card R600.

8:10 AM, June 13, 2006  
Anonymous Anonymous said...

They'll probably dump it on the 3rd world where people unknowingly will buy it cause it's CHEAP!

Then they'll go online and find out that they bought space heaters and can't afford the electricity... :)


this processors will be used by people, that buy a box with a lable COMPUTER on it and who don't know who amd or intel is (probably 80% or more of private customers)

9:16 AM, June 13, 2006  
Anonymous Anonymous said...

"Then they'll go online and find out that they bought space heaters and can't afford the electricity... :) "

Don't exaggerate too much. PC is not data center, it doesn't run 24*7 with full load. People who can't afford electricity bill won't be able to afford a PC neither.

9:36 AM, June 13, 2006  
Anonymous Anonymous said...

I would buy a Pentium D 940. I found a price of $158 for this bad-ass motherF***** of a chip and I can build a complete system for about $500 at this price. Imagine a duam core PD 950 for $500!!!! That is awesome. Thanks Intel.

10:40 AM, June 13, 2006  
Anonymous Anonymous said...

Where do you get 7 bil for INTC inventory? It was 3.5 bil exiting q1.

10:40 AM, June 13, 2006  
Anonymous Anonymous said...

A word about energy efficiency... Let's assume that an Intel based Pentium D 940 uses 100Watts more of power than a comparable AMD. Actually that's pretty conservative but it makes the maths easier. Anyway, if we assume 100W difference. Also where I live I am charged 8.9cents per kilowatt hour. I use my PC about 4 hours per day so every 2.5 days this results in one kWh more usage. Now lets say I can put together a Pentium D system for $100 less than the equivalent AMD system which is realistic given the current price drops of Pentium D systems. How long before I make up that difference through electic bills?

At my current usage and current electric rates, it will take me over 7 years before that $100 is chewed up by energy bills. If I put the $100 in the bank and gained 5% interest over that timeframe it is even longer... So don't be fooled if you are a typical home user, the cost of energy is negligable in this case and can be ignored. If you're running 24x7 servers, then you would care.

11:00 AM, June 13, 2006  
Anonymous Anonymous said...

Where do you get 7 bil for INTC inventory? It was 3.5 bil exiting q1.

intel hasnt sold any processors in 2nd quarter - so all have to be in stock...

11:02 AM, June 13, 2006  
Anonymous Anonymous said...

It's not too hard to figure out which people are posting from the Intel.com domain in this thread.

With these price cuts AMD is taking the wind out of Conroe's sails and preventing Intel from dumping more legacy CPUs into an already stuffed (and annoyed) channel as easily. It was an extremely smart move, as it will slash Intel's revenues and maybe even force them to dip into their dwindling cash reserves.

Combined with the anti-trust suit, the end of preferential pricing for OEMs, the defection of many of their partners* to the AMD camp and a slowdown in the market, this price war is definitely looking to be headed in AMD's favor and will hurt Intel the most of the two parties.

Remember, although AMD is a smaller company, feeding 10000 employees is a hell of a lot easier than feeding 100000, especially when you have little cash, large debts and extremely ticked off shareholders breathing down your neck. Even if they manage to sell their cash-toilet divisions quickly, the charges from severing as many as 16000 employees will hurt them in the short term severely.

I agree with Dr. Sharikou that this year is going to be very hard on Intel and winning a few 32-bit benchmarks with a suped-up Pentium-M that they won't be able to ramp up production on seriously until 2007 won't be enough to prevent some serious damage to their marketshare.

* - Intel regularly blades it's so-called partners in the back on a regular basis and competes with them directly, unlike AMD

11:07 AM, June 13, 2006  
Anonymous Anonymous said...

From that last post, it looks like Dr. Sharikou is posting as anonymous and likes to refer to himself in the third person. How else would anonymous know who is and is not posting from intel.com domain? "Sharikou Busted!"

12:16 PM, June 13, 2006  
Anonymous Anonymous said...

"Sharikou Busted!"

Thats pretty funny!

12:46 PM, June 13, 2006  
Anonymous edward said...

"At my current usage and current electric rates, it will take me over 7 years before that $100 is chewed up by energy bills."

Nonsense. You must be living in Alaska to not care about the 100W extra heat piping out from your desktop.

So in you do live in far north, and use computer around 4 hours a day, then buy P-D. If you're like me who turns on computer 12+ hours a day, then buy AMD.

12:48 PM, June 13, 2006  
Anonymous Edward said...

"Imagine a duam core PD 950 for $500!!!! That is awesome. Thanks Intel."

OTOH, with *current* pricing, you only need to spend $600 to build an x2 3800, which is faster, quieter, cooler, and smaller (ever try putting P-D on small form factor?).

This is a lot for just $100 (25%) price difference. All Intel fans should realize that their "loyalty" to Intel's cheap products is actually giving them a lower quality of life here.

12:54 PM, June 13, 2006  
Blogger Sharikou, Ph. D said...

"Sharikou Busted!"

Thats pretty funny!


It was funny indeed. I couldn't help laughing. Intel fanbois are out of ammo.

12:56 PM, June 13, 2006  
Anonymous Anonymous said...

Benchmarks don't lie. At this point, it's not whether Conroe is faster than K8 or not, it's how much faster.

2:34 PM, June 13, 2006  
Anonymous Anonymous said...

Edward, have you looked at Shuttle.com? The have a WHOLE lineup of SFF PCs that accept Socket 775 chips from Intel. They run pretty cool and quiet too. I think you need to get your head out of your ass and realize that AMD exists because Intel lets them exist. Don't want no pesky FTC antitrust lawsuits again.

Yeah "Sharikou Busted!" that was funny. I'm glad Sharikou has a sense of humour.

2:50 PM, June 13, 2006  
Anonymous Anonymous said...

"AMD must prevent Intel from converting its legacy P4 chips to cash at all costs. Currently, Intel has $7 billion worth of legacy chips as inventory."

why, exactly?

2:59 PM, June 13, 2006  
Anonymous Anonymous said...

dear sharikou,

when i read your posts, it seems you want to replace one market leader, that can dictate the cpu price with another. are you really sure, thats what you want? you want to pay tree time the price for a cpu than now?

i prefer two companies, that compete with another and keep the prices low! but thats just my opinion...

3:17 PM, June 13, 2006  
Anonymous Anonymous said...

"All Intel fans should realize that their "loyalty" to Intel's cheap products is actually giving them a lower quality of life here."

I see what your saying, and I know its in regards to Netburst (or at least hope it is), but it is purely BULLSH^T...

How is AMD going to change my quality of life?

Will you say the same thing "if" Conroe ends up being better than K8?
(I know its a 3 year old design, but thats whats out there.)

Some people have just goten used to what they use, whats wrong with loyalty? We buy cars based on loyalty. Shoes, clothes, pretty much everything.

Lower quality of life huh... well I am living quite well, Intel or not, I am happy with my computer, as are most people who play games and maybe do some video editing with their machines.

3:23 PM, June 13, 2006  
Anonymous Anonymous said...

Sharikou, Intel fanbois are not out of ammo and in fact are packin' heat. At least Intel are packin' The ammo? How about 3 fabs running on 65nm process by end of year. Owned by Intel not by someone we have to pay to fab... Wasn't it Jerry Sanders who said "Only real men own fabs."? I guess AMD are not "real men" or at least half men. Another piece of ammo... Conroe, Woodcrest, Merom. Conroe is at least as good as FX-2s and will be significantly cheaper. Intel mobil has always been better and Merom will keep it that way. Core 2 architecture will ramp production faster than you think. Trust me. Last piece of ammo... who will last longer in a price war? Intel. Has to do with cash and assets. Intel can shed losing businesses with short term losses but longer term increased profitability. IE, same thing that happened when AMD ditched Spansion can happen with Intel just on a larger scale. If it was good for AMD it's also good for Intel. Sharikou, why do you hate Intel so much? Did Intel hurt you in some way? Did they hurt your feelings? Did you get fired from Intel at some point? Because your pathological hatred of a company is just well ... pathological.

3:26 PM, June 13, 2006  
Blogger Mad Mod Mike said...

"How about 3 fabs running on 65nm process by end of year"

How about 3 fabs that have 33% yields and are equal to 1 FAB from AMD? Intel fanboys think they're packin' .357 when it's really .22 cal.

"Benchmarks don't lie. At this point, it's not whether Conroe is faster than K8 or not, it's how much faster."

I'm sorry, but you are a f*cking dumb@$$ if you think benchmarks don't lie.

4:23 PM, June 13, 2006  
Anonymous Anonymous said...

"I'm sorry, but you are a f*cking dumb@$$ if you think benchmarks don't lie. "

Remember when Nvidia got busted for optimizing their drivers for some particular benchmarks? I want to say it was around the 5000 series cards when it happened. Correct me if I am wrong.
I also believe ATI resorted to some funky tactic optimizations thru drivers. Just to look good on benchmarks.

7:21 PM, June 13, 2006  
Anonymous Edward said...

"How is AMD going to change my quality of life?"

Without AMD's competition you won't buy today's Pentium-4, let alone tomorrow's Conroe. Instead, you'll be paying $1000+ for just CPU+MB for an Itanium.

It's amazing to see how ignorant people can be!


"Will you say the same thing "if" Conroe ends up being better than K8?
(I know its a 3 year old design, but thats whats out there.)
"

Conroe IS better than K8 in some aspects, but not all. However, as you said, K8 is what's out there, while Conroe is not. :)

I'm sorry that you have to put up with current Intel machine's heat and noise while dreaming about tomorrow's Conroe. As for me, I am enjoying quiet and fast and cool AMD today.

Well... I'd never go back to the "quality life" you have with Intel's hot & noisy machine (I used to work in an office that has 2 of them). At least not in my home.

7:55 PM, June 13, 2006  
Anonymous Anonymous said...

sfftech.com has reviews of many small form factor computers. A Shuttle SN21G5 with a (non-energy efficient) Athlon X2 3800+ under load has a noise level of 56 dBA at 12". A Shuttle SB83G5 with a 3.6 GHz P4 under load has a noise level of 60 dbA at 12". Worse yet, the P4's CPU temperature was 7 degrees hotter. Note that this is for a P4, not a Pentium D.

8:07 PM, June 13, 2006  
Anonymous Edward said...

"The have a WHOLE lineup of SFF PCs that accept Socket 775 chips from Intel."

Have you built an Intel SFF and an AMD SFF yourself? If you have and know how much more difficult it is to dissipate heat with Intel's you know what I'm talking about; otherwise sorry it's you whose head's inside his ass.

Just because shuttle.com list those boxes doesn't mean they're good choices. 80W more heat under load is still 80W more heat, no matter what form's you're in.

8:09 PM, June 13, 2006  
Anonymous Anonymous said...

"I'm sorry, but you are a f*cking dumb@$$ if you think benchmarks don't lie." You forgot to mention that benchmarks still shows Conroe is FASTER and STILL BEATS FX-62 (even if some of the bnechmark margin is small). LOL..

At this rate, Sharikou looks like is going out of business as as July 23rd.. maybe make that by 4Q06. AMD already sold their alchemy unit, thus its not foreseeable they may cut Sharikou's "phd" funding... LOL :P

9:52 PM, June 13, 2006  
Anonymous Anonymous said...

"How about 3 fabs that have 33% yields and are equal to 1 FAB from AMD?"

Link? Speculation? Fantasy?

11:23 PM, June 13, 2006  
Anonymous Anonymous said...

Speculation at best; notice the link to "The Tech Report" benchmarks showing the 3800+ X2 beating the 965XE on ONE benchmark. Okay, so if you're saying that benchmarks lie, then you're denying your own evidence. NUMEROUS sites have published that Conroe is definitely packing heat; it will not get "fragged" by the FX-62 or any of its overclocked derivatives any time soon.

12:32 AM, June 14, 2006  
Anonymous Anonymous said...

Here is your original quote...

"All Intel fans should realize that their "loyalty" to Intel's cheap products is actually giving them a lower quality of life here."

Here is my question...

"How is AMD going to change my quality of life?"

Here is your rebutal...

"Without AMD's competition you won't buy today's Pentium-4, let alone tomorrow's Conroe. Instead, you'll be paying $1000+ for just CPU+MB for an Itanium.

It's amazing to see how ignorant people can be!"

Yes it is, so why dont you try and answer my question... How is AMD going to change my quality of life? Am I going to get laid more... "Hey baby I have an AMD", get real man.

My point is that a computer doesn't change your quality of life, Intel or AMD, may make for a better computer experience but thats it.

"Well... I'd never go back to the "quality life" you have with Intel's hot & noisy machine (I used to work in an office that has 2 of them). At least not in my home."

Have you heard of quality fans, good cases? You have got to be talking about socket 478, those ran really hot, moving to the 65nm 775 the temps are quite good, I idle 32 on a overclocked 651. The loudest part of my case is the ATIx1800xt. Damn that thing is loud!

Here is a screenshot...

http://www.enumae.com/computer-temps.html

Hey man, not trying to argue with you, but think about the meaning of "quality of life", maybe I am missing something here but I can not connect it to a computer.

8:05 AM, June 14, 2006  
Blogger Mad Mod Mike said...

"NUMEROUS sites have published that Conroe is definitely packing heat; it will not get "fragged" by the FX-62 or any of its overclocked derivatives any time soon."

Numerous = All of them at Taipei

Some people can't read, either that or they just dun wanna!

PS: You can't fix stupid

9:03 AM, June 14, 2006  
Blogger Entrophos said...

To all the morons saying that Dr. Sharikou posted the comment dated "11:07 AM, June 13, 2006":

I posted that before I signed up for a Blogger account. If you were halfway intelligent you'd notice that our writing styles are nothing alike.

I guessed that some Intel whackjobs found this blog and decided to post their BS because I don't see how anyone could be so stupid unless it was on purpose. Are some people at Intel getting a little afraid of the fact that we can see through the benchmarketing and are spreading the word?

Seems I was right, judging by some of the responses.

10:44 AM, June 14, 2006  
Blogger Mad Mod Mike said...

"Seems I was right, judging by some of the responses."

It's the classic case of "FUD" for the Intel fanboys. Once people prove them wrong, they start insulting or saying obsurd remarks relating to nothing to say "Intel pwns", it's quite disturbing.

Having a huge loyalty to a company is great, it means you'll always have a product. But being loyal to a company who is inferior to another on all aspects (morality, products, etc.) and not accepting the fact that there is better, is just downright stupid.

11:14 AM, June 14, 2006  
Blogger Sharikou, Ph. D said...

To all the morons saying that Dr. Sharikou posted the comment dated "11:07 AM, June 13, 2006":

I posted that before I signed up for a Blogger account.


They knew I didn't do it. Everyone knew. But they said it, expecting me to claim innocence. But then I couldn't offer any proof, one way or the other. And they would successfully derail the disccusion to be about me, instead of the question about whether Intel can survive the price war.

11:34 AM, June 14, 2006  
Anonymous Edward said...

"why dont you try and answer my question... How is AMD going to change my quality of life?"

I've already answered - competition, and AMD's offering a good one. It's been playing very constructively in the way it approaches its customers and OEMs. It offers better technologies when its bigger foe focused on hypes and brands and marketing. For 3 years, till now, if you buy an Opteron-based machine, it's more powerful, less power-hungry and noisy. Don't these improvement everyone's quality of life?

PLEASE.. DO YOU EVER READ OTHER'S WORDS (not just quote them)?


"Am I going to get laid more... "Hey baby I have an AMD", get real man."

Yeah... if THAT's what you mean quality of life, then no, AMD doesn't do that.


"Have you heard of quality fans, good cases? You have got to be talking about socket 478, those ran really hot, moving to the 65nm 775 the temps are quite good, I idle 32 on a overclocked 651."

1)Your 651 is a 86W typical CPU, while the A64 3500+ is a 67W max CPU. There's no match here.

2)For P-4 to reach the same low temp it's speed must be higher (whether under load or idle). That severely decrease the life of fans, and your fans will surely get noisy much sooner than those with AMD inside.

12:30 PM, June 14, 2006  
Anonymous Edward said...

"A Shuttle SN21G5 with a (non-energy efficient) Athlon X2 3800+ under load has a noise level of 56 dBA at 12"."

Exactly!

I have an X2 3800+ in SFF, built myself, under my desk at home. When idling, the whole thing consume < 100W power. At night, I can hear my neighbor snoring better than that box. Think you can match that noise level with Pentium-D or even Pentium-4?

BTW, not all noise or heat or point of failure comes from CPU/fan. When you have a more power-hungry CPU and motherboard (w/ memory control @ high freq), the fans and temps of power supply become more important. I have had two PS failures due to packed space around the box, one K7 and one P4. The reduced overall power consumption of Opteron-based system makes every component's life easier inside the box.

Now, Conroe might change that. In fact people've been using Dothan or Yonah in SFF for this very purpose. Congrats to Intel, that its 10-billion-revenue company still has something hot&cool for consumers. And thanks to AMD to push Intel the right way!

12:57 PM, June 14, 2006  
Anonymous Anonymous said...

Numerous = All of them at Taipei

Some people can't read, either that or they just dun wanna!

PS: You can't fix stupid

Seems like you were pretty happy with the taipei tweak town benchmark with conroe showing that it was beating fx-62 by 15% instead of 20%. I guess that benchmark is alrite huh.

PS: You cant fix BIAS either....

2:50 PM, June 14, 2006  
Anonymous Anonymous said...

"PLEASE.. DO YOU EVER READ OTHER'S WORDS (not just quote them)?"

I did read your statements, several times, trying to be sure not to miss anything, but somehow maybe I have misinturpreted (sp?) what your saying.

Quality of life I quess is different for everyone. Competition between AMD and Intel... well it may be, but not for me.

Not trying to argue, but understand what your saying.

3:19 PM, June 14, 2006  
Anonymous Anonymous said...

Google search definiton : quality of life...

“feeling of wellbeing, fulfillment, or satisfaction resulting from factors in the external environments.”

Its a stretch, but external enviroments could be considered computers.

3:27 PM, June 14, 2006  
Anonymous Edward said...

"Quality of life I quess is different for everyone."

Okay... maybe I should rephrase it so that it's less subjected to my subjective view:

If you use computers 24/7 like me and most professionals and enthusiasts, you will welcome Opteron's power efficiency and quietness (and for that matter, Dothan's and Turion's), which brings better quality to your everyday life.

If you only use a computer few hours a day to play PC games with earphones on, then you could just order any prebuilt machine and don't care what's inside (80% are Intel P4, though).

9:08 PM, June 14, 2006  
Anonymous Anonymous said...

Wow, this blog turned into a crazy flame-fest. A few points:

Conroe will, most likely, outperform current X2s, though I'd expect a speed bump when Conroe launches (an X2 5400+, IMO). Of course the way some of the benchmarks out there were achieved is a bit dodgy but a right out lie about performance would destroy Intel's credibility forever. At best (for AMD) I'd expect the FX to be out-performed by around 10-15%, which is not bad at all, in my opinion (remember that the XE version of Conroe WILL draw more than 65W).

4x4 is hardly a desperate stop-gap solution because AMD built support for it into socket AM2, and we've known its specs for quite a while. AMD probably kept it secret because it wanted to be a surprise to be revealed at a time of its chosing (Taiwanese companies leak like sieves).

Personally I'm pretty hyped up about it because it actually brings workstation-class performance to the desktop (I'm not that big of a gamer) at a much more affordable price, not to mention that we'll probably get great overclocking features on 4x4 motherboards (I do love my overclocking).

I'd guess that, like Mad said, X2s will work fine with 4x4. Intel's performance lead will take enthusiasts away from AMD but the ability to thinker with 2-socket systems will keep others (like myself) in AMD's camp. Torrenza sounds absolutely tasty, too, but will take some time to come to life.

The Pentium XE 965 matches the power consumption of the FX62, which is a remarkably hot K8. the X2 5000+, for example, consumes almost 30W less at full load.

Intel's 65nm yields are probably not that great right now and they'll probably take another hit as Core 2 in thrown into production. AMD is able to ramp up production much faster than Intel simply because it has the need to maximize its fab efficiency and has invested heavily to do so. Intel's massive fab capacity advantage more than makes up for this, though. Then again, Tulsa, Itanium, flash and the various other pet projects at Intel also take from this.

The X2's price drop will probably do very good things for AMD if its capacity doesn't limit it. The cost of silicon is hardly the only cost that goes into making CPUs and most of the other costs (packaging, transportation, advertising, etc) are the same regardless of the size of the die being thrown into the CPU.

I don't believe AMD has to do anything to block Intel's crazy Netburst sales. It's a nice thing to do because it allows them to gain market share and increase profits while using AMD massive momentum. Neither Intel nor AMD will be selling any chips below cost, by the way.

7:25 AM, June 15, 2006  
Anonymous Anonymous said...

Conroe looks like it'll be a great chip.

The problem for Intel is that if the rumours of channel stuffing, margin drops and sales figures are true, they f#$cked themselves over starting six months ago, and things will get much, much worse before they get better.

They've rung the bell, they've cranked the helm hard-aport, but the cash-crunch iceberg is still dead ahead for The Unsinkable Intel.

Now, they're not going to vanish. No doubt they can raise whaty they need by issuing shares or debt, but existing shareholders are going to take it in the nuts.

12:17 PM, June 15, 2006  
Anonymous Anonymous said...

AMD will only have one sub 200$ dual core part. Intel will have it's cheapest dual core at 90$. Do you really think a general consumer will fork out 169$ when they can get dual core for anywhere starting at 90$. AMD will sell some 3800's but the reason they are not dropping prices across the board is because they will lose a lot of money if they drop prices on dual core parts across the board. And I bet the reason they are cancelling the 1MB cache parts is because of capacity. Intel is going to use 65nm and Presler two chips in one package manufacturing advantages to squeeze AMD on pricing. They will convert the entire market to dual core faster than anyone can say Jack and AMD will not be able to follow them cos they don't have the capacity. Guess who can afford to bleed less...I don't think the company with 10+ bln dollars in the bank. AMD will survive but they are in a street fight with knives and Intel is bringing guns.

9:52 AM, June 18, 2006  

Post a Comment

Links to this post:

Create a Link

<< Home