Tuesday, September 05, 2006

Intel's cash insufficient to layoff 80% of its workers

Intel cut its number of workers by 10%, or 10,500. I noted that Intel incurred $0.2 billion severance costs, averaging $20K per worker. Those who spent their youthful years at Intel got a $20K check. Sigh!

To layoff 80% of its workforce as a preventive measure to avoid BK, Intel needs $1.6 billion dollars pure expense. Looking at Intel's balance sheet, Intel doesn't have enough money for that.

Conclusion: Intel's BK is inevitable.

PS: Check X2 3800 stock level, it's on back order again, though relief is coming.

78 Comments:

Anonymous Anonymous said...

You are kinda mean Sharikou. There are 17,000 people here in my home town Portland, Oregon working for Intel. Just the sheer thought of 80% of these folks getting lay-off is a nightmare.

While I am an AMD-fan and anti-Intel, I cannot accept your extreme view. I really hope you are wrong and will call you on that. I will remind you of this some time next year. And I will laugh at you.

-Longan-

1:44 PM, September 05, 2006  
Blogger Sharikou, Ph. D. said...

There are 17,000 people here in my home town Portland, Oregon working for Intel.

The one who is mean is Intel. I have a lot of sympathy for those workers who spent their youth for a big company like Intel and end up jobless in a snap of finger, while Paul O and others rake in millions.

It's better for Intel go away fast and AMD grows fast, so those workers can work for a better company such as AMD.

1:48 PM, September 05, 2006  
Anonymous Anonymous said...

To layoff 80% of its workforce as a preventive measure to avoid BK, Intel needs $1.6 billion dollars pure expense. Looking at Intel's balance sheet, Intel doesn't have enough money for that.

Conclusion: Intel's BK is inevitable.


Assumption: Intel needs to lay off 80% of it's workforce to avoid bankruptcy.

Fallacious conclusion: Intel goes bankrupt because it does not have enough cash to lay off 80% of it's workforce at once.

There has been no citable evidence (vitriol and opinion don't count) presented in this forum or any other that would indicate Intel is a)in imminent danger of going bankrupt or b)needs to lay off 4/5ths of its workforce to avoid bankruptcy.

Stating it using a logical construct makes it neither logical nor true. Repeating it frequently does not change the result. It just makes you look the fool.

2:17 PM, September 05, 2006  
Anonymous Anonymous said...

We need both AMD and Intel.
If Intel is out of the picture, AMD can charge whatever they want and have no incentive to innovate, they'll become just like Microsoft.

2:21 PM, September 05, 2006  
Anonymous Anonymous said...

Conspicuous timing, I think.

They just launched 3 new product lines and I guess they don't need them anymore....so just get rid of them?

Heartless, mean and evil or just good business?

2:30 PM, September 05, 2006  
Anonymous Anonymous said...

I don't hate any companies but sorry to say this Sharikou is right. Intel just could not recover from all that has happened for the past 3 years.

2:44 PM, September 05, 2006  
Anonymous Anonymous said...

The one who is mean is Intel. I have a lot of sympathy for those workers who spent their youth for a big company like Intel and end up jobless in a snap of finger, while Paul O and others rake in millions.

Where were your crocodile tears when AMD was laying off the Americans that worked in its three Austin chip fabs?

We need both AMD and Intel.
If Intel is out of the picture, AMD can charge whatever they want and have no incentive to innovate, they'll become just like Microsoft.


Why did AMD slash prices after C/W/M were released? Good thing we have Intel around to keep AMD from scamming the consumer with artificialy high prices.

3:05 PM, September 05, 2006  
Anonymous Anonymous said...

Intel is "bluffing" everyone with the announcement of saving $3B by expensing $200M to layoff 10K people!!! (Either the savings amount is over-projected OR the number of employees let go is way under-projected).

If they are right, $3B will add .50c to thier earnings (reduce their losses ). Why is the stock down after news????


I think, really, Intel needs to save about $3B to break even, but can't say they are issuing pink slips for 30,000 employees (that will send wrong message and help AMD immensely).

Intel is down, ALL THE WAY!!!

3:08 PM, September 05, 2006  
Anonymous Anonymous said...

I wish I worked at Intel for that free 20K:)

3:09 PM, September 05, 2006  
Anonymous Anonymous said...

Intel still has more marketshare than AMD after losing the technology race for 3 years.
http://www.vnunet.com/vnunet/news/2163452/intel-holds-marginal-lead-amd
Even without Core 2, Intel has been improving these last months, from 51%, to 53%, to 58% in retail.
How will AMD carve into Intel's share when they're losing the technology race now?

3:32 PM, September 05, 2006  
Blogger Sharikou, Ph. D. said...

Intel is "bluffing" everyone with the announcement of saving $3B by expensing $200M to layoff 10K people!!!

I too find their statements suspicious. With 100K people, their annual cost is $12 billion, cutting 10% of the people and the cost reduces by 25%?

3:46 PM, September 05, 2006  
Anonymous Anonymous said...

Here is where the rest fo the saving comong from - read Intel's press release:
In addition, Intel expects to achieve a capital expenditure avoidance of $1 billion by better utilizing manufacturing equipment and space. The company expects that approximately 25 percent of the project's savings in 2007 will reduce cost of sales, and the rest will reduce operating expenses.

4:22 PM, September 05, 2006  
Anonymous Anonymous said...

$3bn savings are till 2008. Presumably ~$1bn odd every year.

5:00 PM, September 05, 2006  
Anonymous Anonymous said...

Once the cash dries-up, all the Intel Fanboy sites like THG will have to sing for their supper!
Speaking of THG, they have recently taken to faking posts from AMD enthusiasts in an effort to fake impartiality!Try posting anything pro AMD and the thread gets locked.Most of the recent AMD threads you see are posted by Intel Fanboys in a pathetic attempt to save face after some recent posts critical of THG policy.

5:12 PM, September 05, 2006  
Blogger Sharikou, Ph. D. said...

Try posting anything pro AMD and the thread gets locked.

That happened to me a few times, and I gave up educating those people.

BTW, FireFox 2 beta is great. It shows spelling errors while I am typing this comment.

5:23 PM, September 05, 2006  
Anonymous Anonymous said...

Sharikou, I respect your technical expertise immensely. I think all of us come to your site for that reason.

You have "PH.D" after your name. I have "CFA" after mine. I can assure you, Intel is not now nor is it likely to file bankruptcy in the next year...the next 2 years...the next 5 years.

Unless their market share dropped to 20% and the value of their assets (fabs, real estate, etc) fell by 30-50%, that's the only thing that would even give it pause.

Look at GM. They've been bleeding cash for most of the past 25 years, and have lost nearly half their market share in that time. Their market value is about 1/10th that of Toyota's.

Bottom Line: It takes a LONG time for an established company to go from troubled to distressed to bankruptcy. Look how long Bethlethem Steel held on.

Even if we look back on this time as the period when AMD did to Intel what Intel and MSFT did to IBM/DEC/Sperry-Rand, it will be 2015-2020 before you even HEAR about a so-called potential bankruptcy involving Intel.

Heck, even Sperry (nee Unisys) is still around !!

5:23 PM, September 05, 2006  
Anonymous Anonymous said...

"That happened to me a few times, and I gave up educating those people."

I thought you were banned from THG, I didn't realize you just stop posing on your own accord.

5:31 PM, September 05, 2006  
Anonymous Anonymous said...

The news on Intel is extemely misleading.

A) 10,500 folks have not been layed off. Intel is already 5000 of the 10500 employees lower - ~1000 of which was due to layoffs. The rest were sales of businesses and standard attrition. They have not cut 10,500 as of today.

Paul O referenced the highpoint of employees at end Q2 (102,500); currently Intel is around 97500.

B) there will be additional 2500 reduction by end of 2006. this will be combinaton of layoffs and attrition (mainly layoffs)

c) There will be additionl 2500-3000 by mid 2007. This again will be mixture of attrition and layoffs.

D) not everyone of the 10,500 gets severance package - only those specifically laid off. Those moving on due to normal attrition (generally 3-5%) don't get severance packages.

5:48 PM, September 05, 2006  
Blogger Ho Ho said...

"Though, even without Intel, AMD can only charge what the market is willing to pay."

As was said, AMD cut most its prices in half just a little while ago. Not many people complained about AMD prices back then. I'd say in monopolistic world having ~5-10x price increase is not out of question.

"Microsoft does hae a great deal of competition from numerous Linux distros, Solaris, OSX, SkyOS, and etc"

I wouldn't call <10% of various OS'es vs 90% various versions of Windows a competition. I would hope that if/when PS3 gets popular with its preinstalled Linux OS things might start to look a bit different.

I personally am a Linux-only user at home and at work and can manage everything, even gaming.

5:59 PM, September 05, 2006  
Anonymous Anonymous said...

NYX - there is a difference between what the market will bear and what is a good price; just look at the FX pricing before and after Conroe.

Before Conroe, these chips were at over $1000, now they are what? Did the cost of producing them suddenly go down at the exact same time as when Conroe was released? If Conroe had not come out would you have seen his price reduction at the same point in time? When 4x4 comes out and pricing is further reduced (to be under the $1000everyone seems to claim); would this had happened if Kentsfield or Connroe were not out?

It just silly for folks to assume competition does not reduce pricing and to also assume (sharikou) if that competition went away (Intel BK's), AMD will continue to aggressively reduce pricing. Just look at past server pricing where AMD had substantial lead and price cuts were far below what was going on in mobile and desktop spae.

How much were 5000 X2's before and after Conroe again?

Sorr for typos - keyboard battery is dying! :)

6:01 PM, September 05, 2006  
Anonymous Anonymous said...

"If they jacked up the price too high, they would encourage other competition to step in which I'm sure they don't want. "

The barriers to entry are so high in the CPU space this is unlikely (except maybe IBM). The design and production cycles are so long, the time to achieve market penetration, and the IP on chip design (AMD and Intel are cross-licensed on a bunch of stuff) are such that very few companies could afford to even consider entering the space.

6:07 PM, September 05, 2006  
Anonymous Anonymous said...

LOL..

I have a CFA.. I have gone and visited a few of them jokers. Find out they would charge me 5% of my money to buy high load investments of sectors I already owned. It was laughable how ignorant people are easily impressed by people who put their titles in email, on business cards or signatures. Shut the F up you don't know anything comparing INTEL to Steel or GM is not relevant. TO compare so only shows how stupid you are. Its technology, miss one generation and you in trouble, miss two and you go bankrupt.

Somehow that CFO posting and this pretenders PhD fall into the laughable side of things.

Think about it today.. After what is 4 years of marketing a inferior product INTEL still has 70% plus of the marketshare, is more profitable then number 2-10 semiconductor company combined and is on track to make probably more then 2-10 again this year and next. Yup that is a company on the road to bankruptcy.. According to the PhD its in 4 quarters now since they aren't listneing to him and laying off 80%. That is right right Sharikou ph. d....

Pss INTEL has all the benchmarks, lead in 65nm, lead in 300mm factories, lead in evertying except 4core+... In the past they never owned the high end anyways.. so nothing has changed. Replace Sun/Power with Opetron makes no difference INTEL is king and AMD is a distant second.

Not PhD required to figure this out...

6:12 PM, September 05, 2006  
Blogger Sharikou, Ph. D. said...

After what is 4 years of marketing a inferior product INTEL still has 70% plus of the market share

You fail to see the most important factor: capacity. The barrier of entry into the x86 world is very high. Even with inferior technology, Intel was able to maintain monopoly and high profits, because AMD could only produce that much. FAB36 is a fundamental change. Intel's BK is mostly a result of AMD's expanding capacity, not technology leadership. A company can design some super chip, but if it can't get the volume output at reasonable yield, it won't fly.

AMD has struggled long enough to be able to build two modern FABs, enough to take 50% of the market.

Now, with AMD taking 50% and Intel 40ish, the result is Intel's violent collapse.

6:47 PM, September 05, 2006  
Blogger pointer said...


I noted that Intel incurred $0.2 billion severance costs, averaging $20K per worker. Those who spent their youthful years at Intel got a $20K check. Sigh!


http://www.sfgate.com/cgi-bin/article.cgi?f=/c/a/2006/09/05/BUGH2KVNQ913.DTL

The cut are mainly through attrition. Stop showing off your primary school math without able to read and use the appropriate divisor

7:07 PM, September 05, 2006  
Anonymous Anonymous said...

PhDs, CFAs... listen up...

I was going to post that Intel will not be going bankrupt. They will lay off as many as they can, borrow anything they can, issue new shares, whatever it takes to keep the executive gravy train going.

But then I got to thinking... if there was some big unfunded liability on the horizon, the execs might do it to clear the decks and keep their own gravy train going. (And screw the shareholders; they'll give themselves a fat retention package anyway)

So, it could go either way, but it's not very likely at this point.

7:10 PM, September 05, 2006  
Blogger yauipop said...

I reckon the best outcome is to have both companies playing fair and square for market share. I would really like to see Intel and AMD fight on, otherwise we would not see great CPUs like the A64 and C2.

8:34 PM, September 05, 2006  
Anonymous Anonymous said...

FACT - As long as Intel decides to stop the price war it is loosing, it will not go BK.
AMD can only supply 50% of the market, so while Intel may slip a couple of percent in market share, they will not be leaving us. What would happen is both semi firms would become profitable, and AMD would SLOWLY eat away at INTC market share until INTC gets the "house" in order.

10:36 PM, September 05, 2006  
Anonymous Anonymous said...

Wise investor:

Those that got layoff better make some smart moves. Like put their severance pay on purchasing shares of AMD. Should still have some 4 - 7$ headroom left or 25%.

11:40 PM, September 05, 2006  
Anonymous Anonymous said...

"Looking at Intel's balance sheet..."

Not that your 80%/1.6 Bil is based on any reality (I think you forget Intel still does some flash, chipsets and other products in addition to CPU's)...but anyway let's look at the balance sheet....


Cash and cash equiv (Q2'06) = 4.65Bil
Total curent assets 16.8 Bil (total assets = 46.1 Bil)
Total curent liabilities = 8.4Bil

I know you like to look at cash flow - perhaps something you haven't considered; Intel spends $1.5-3.0 Bil per QUARTER on stock repurchase and dividend. So basically suspending this for 1 QUARTER would cover the ridiculous layoff expense you calculated.

Only 4-6 quarters (soon to be 3-5) before you eat crow; better start planning the excuses now....

12:24 AM, September 06, 2006  
Anonymous Anonymous said...

INTEL is king and AMD is a distant second.

Yeah, King in laying off and slapping two die together.

2:06 AM, September 06, 2006  
Blogger Christian Jean said...

As I'm writting this (Wednesday, 08:55), there is an Intel report on CNBC.

Here is a summary of what was discussed/debated:

1) Many anaylysts say the 10,500 job cuts are NOT enough... more are needed.

2) Its now understood by Wall Street that Intel lost its monopoly and it is no longer the company that it once was.

3) Analysts that keep proclaiming to 'buy now' (Intel stock) because the prices have never been so cheap are basically not being listened to anymore because the price has kept falling for the last couple of years and investors are afraid.

4) A CNBC comentator has stated that Intel is no longer trying to protect its decreasing profits, but have moved to strategies to prevent potential future profit losses.

5) A bit of talk on how competitive AMD has been and AMD processor inovations.

Now let me remind you that CNBC has over 300 million world wide viewers.

6:04 AM, September 06, 2006  
Anonymous Anonymous said...

Another idiot said
"But then I got to thinking... if there was some big unfunded liability on the horizon, "

You ever hear of sarbanes oxley?

Executives are personally liable now. Of course you can be like Lay and go cheat lie then die... but that was punishment enough.

INTEL BK I don't think so.. of course unlike Samsung INTEL nor AMD can sustain losing billions over quarters over a full generation cycle. No question INTEL is better positioned currently.

They need to learn from Ford/HP and sack the CEO.

But BK next year, highly unlikely.

Sharikou can you find a more relevant topic

INTEL BK and exploding batterys are getting very boring, boring boring.

a real Phd

7:05 AM, September 06, 2006  
Blogger 180 Sharikou said...

To understand the impact of the announcement and to see what the real severance per employee will likely be, visit:

http://sharikou180.blogspot.com/

180 Sharikou - the other side of Sharikou

8:34 AM, September 06, 2006  
Anonymous Anonymous said...

Let us not forget that everyday that passes Intel is getting closer and closer to facing AMD, and governments in court worldwide. These court cases are not going to bode well for Intel. We all know how Intel manipulated PC makers worldwide.

Intel needs to ax a few more thousands of people. AMD is going to get a big payday. Intel may be paying for AMD's purchase of ATI thru this lawsuit.

9:13 AM, September 06, 2006  
Anonymous Anonymous said...


You fail to see the most important factor: capacity. The barrier of entry into the x86 world is very high. Even with inferior technology, Intel was able to maintain monopoly and high profits, because AMD could only produce that much.


So now you accept that AMD's failure to gain market share was because of their capacity woos and had nothing to do with arm twisting by Intel? Wow! What a relief!!

10:53 AM, September 06, 2006  
Anonymous Anonymous said...

http://chip.seekingalpha.com/article/16404
Some analysis on Intel and AMD's financials.

12:39 PM, September 06, 2006  
Anonymous Anonymous said...

"FACT - As long as Intel decides to stop the price war it is loosing, it will not go BK."

How can that be "fact"? Even if that would be true, it has not happened yet. How can that be fact?

Intel was forced to the price war, because of its over-capacity and huge stock of aging processors. Intel will need to sell quite a few fabs and cut another ten thousand workers if it doesn't lower its processor prices to encourage sales. But then, this low-pricing trick can only work for so long, after which the market will be saturated and the arrow flies back at itself.

"AMD can only supply 50% of the market, so while Intel may slip a couple of percent in market share, they will not be leaving us."

Supplying 50% of the market means 100% growth for AMD. OTOH, losing 25% of the CPU market means losing 40% of revenue for Intel (since almost all other businesses of Intel are money-losing). It's just plain clear that which company has better outlook.

1:10 PM, September 06, 2006  
Anonymous Anonymous said...

"Why did AMD slash prices after C/W/M were released? Good thing we have Intel around to keep AMD from scamming the consumer with artificialy high prices."

No, you are wrong. The high margin (50%+) would allow AMD to upgrade its fabrication technology faster, which is how this industry work - you earn a lot of money that you can't use elsewhere except to upgrade your current production.

All these prices are artificial, anyway. The CPU, memory, flash, HD. Companies have consensus on how much they should be and when. If one company decides to lower its prices too aggressively, it will get sued.

It was Intel which decided to release Core 2 with prices lower than it can produce them (see how much street prices are higher) and cut Netburst prices like no tomorrow - that's called a price war. Basically, Intel just accelerated AMD's transition to dual-core. (AMD's single-core processor prices didn't fall as much.) Intel knew that K8's dual-core is more expensive to make than Netburst's. Intel knew that AMD doesn't have as much cash as itself in this price war. Intel's low pricing was more offensive - to crush AMD, than defensive - to save its own profit.

What Intel didn't know very well was probably that AMD could manufacture K8 dual-core very efficiently even under the price war. AMD could even go ahead to buy ATi. There could be other people/companies backing up AMD that Intel was not aware of. Definitely SUN, probably also IBM and HP.

This AMD-Intel conflict is certainly an interesting show to watch. And, if Intel BK'd, I'd not be worried. Intel could better become a few smaller companies after the restructure and bring us a even more competitive x86 (and CPU in general) market.

1:29 PM, September 06, 2006  
Anonymous Anonymous said...

I have some good news, and some bad news.

Good news is that I just bought a 3800+ x2 to refresh my system.

Bad news is that I didn't do it because of the CPU. I did it because the mobo function that I wanted lent itself better to the AMD part.

This was an upgrade with PATA and floppy support desired. It worked for me.

The really bad news is that my needs are a pimple on a gnats butt, and if I was looking at a new system (like the vast majority), I would would have gone with a C2D 6300 without blinking.

Spin this in your direction. Give me a giggle.

3:25 PM, September 06, 2006  
Blogger Sharikou, Ph. D. said...

So now you accept that AMD's failure to gain market share was because of their capacity woos and had nothing to do with arm twisting by Intel?

This was Intel's argument and was a retarded argument. AMD failed to build capacity because Intel's monopolistic behaviour denied AMD the profit to build another FAB.

5:26 PM, September 06, 2006  
Anonymous Anonymous said...

(see how much street prices are higher)

They are not inflated. At all. Prices have settled, try searching for prices.
http://www.newegg.com/Product/ProductList.asp?N=2010340343&Submit=ENE&Subcategory=343&Description=conroe&Ntk=all

5:34 PM, September 06, 2006  
Anonymous Anonymous said...

"AMD failed to build capacity because Intel's monopolistic behaviour denied AMD the profit to build another FAB."

This is a complete distortion - Did AMD not just get a 2 or 2.5Bil loan for thee ATI merger (perhaps they could have thought about an earlier loan for additional capacity? How about another stock offering to raise money? Or how about foundry capacity (like they are currently doing with Charterer) and pay as you go? AMD had plenty of oppportunity to expand capacity if they had really wanted to in the past - their management was just very conservative about it.

AMD's "profits" over the last year is what? ~200Mil? and you're saying THIS is what enabled them to finance a 2-3Bil Fab?!?!? Not to mention this fab is significantly subsidized by the German government.

I guess you could blame everything on Intel, but it was a lack of "cajones" by senior AMD management to bet on additional capcity earlier then they ended up finally doing so. They could have finance this through a number of means...

6:09 PM, September 06, 2006  
Anonymous Anonymous said...

"It was Intel which decided to release Core 2 with prices lower than it can produce them (see how much street prices are higher)"

So you are saying Intel is selling COre2 chips at a loss! (Not even Sharikou would make this kind of statement). Not to mention this might be considered dumping and illegal if Intel is considered a monopoly.

Street prices are higher because demand is high and/or initial supply is low and perhaps (I'm going out on a limb here) - Intel's prices are based on 1000 units (like AMD) so it is ridiculous to think buying a single chip from a distributor will cost the same as Intel quoted price (Does the distributor just pass them on based on goodwill? Or does it cost $0 for the distributor to sell them?) Intel does not control the price newegg, etc sells chips for. They will mark up intially to get the bleeding edge consumers for whom price is not a major consideration (just like any other technology when it is new - HDTV, HD-DVD, Blu-ray, plasma TV's...)

6:16 PM, September 06, 2006  
Blogger Sharikou, Ph. D. said...

AMD had plenty of oppportunity to expand capacity if they had really wanted to in the past - their management was just very conservative about it.

This is twisted logic. So, Intel claimed that AMD should have borrowed a high interest loan instead of making money and using profit to support capacity expansion.

Intel owes AMD $15 billion, just wait till the lawsuit. I have talked about how to compute the damages before.

6:46 PM, September 06, 2006  
Anonymous Anonymous said...

Sharikou said: "Intel owes AMD $15 billion, just wait till the lawsuit. I have talked about how to compute the damages before."

Too bad for AMD that Intel will BK soon, eh?

8:02 PM, September 06, 2006  
Anonymous Anonymous said...

So now you accept that AMD's failure to gain market share was because of their capacity woos and had nothing to do with arm twisting by Intel? Wow! What a relief!!

You could not even find motherboards when the Athlon was first launched. Why? Intel arm twisting motherboard manufacturers. If those Taiwanese companies like VIA and Asus had not put on a brave face and went ahead with AMD motherboards (remember, Intel tried to cut out AMD by using Slot-1 and therefore AMD could not ever get their cpu in your computer in the socket-7 fashion anymore) there would be no AMD. These companies tasted the AMD Athlon and took their risk by getting on the AMD bandwagon.

AMD's failure to gain market share were due to Intel arm twisting and their propaganda. Now that those activities have been curtailed, I am going to enjoy the huge number of choices I have for AMD AM2 cpus today.

8:27 PM, September 06, 2006  
Blogger Sharikou, Ph. D. said...

Too bad for AMD that Intel will BK soon, eh?

BK doesn't mean there is no asset left.

8:58 PM, September 06, 2006  
Blogger pointer said...

AMD's failure to gain market share were due to Intel arm twisting and their propaganda.

are you kidding me? AMD products was bad, real bad at that time. I'm not sure if the CPU used in the Tomshardware test was K5 or K6, which smoked when the fan is taken off the CPU and not able to function again even after put back, and that's the only x86 CPU in the history that would smoke (this is different from Sharikou's conspiracy theory and I beleive Edward would know this, and i believe you should still able to get this information out of the web). AMD start to gain market when it has good product (K8, i'm not sure about K7).

Btw, I still give credit to the K5 design team though (despite it was running hot and slow) as it is the first in-house design that give AMD's ability to come out with later design. it is a feat by itself of such effort.

9:33 PM, September 06, 2006  
Anonymous Anonymous said...

Sharikou said: "Intel owes AMD $15 billion, just wait till the lawsuit. I have talked about how to compute the damages before."

Too bad for AMD that Intel will BK soon, eh?


Hey... just a thought... Do you think after Intel goes BK (and is unable to pay the $15B you claim they owe) due to AMD's success, that whatever is left of Intel can be reassembled into a company that can sue AMD for running them out of business?

Sorry for the silly question. I think I mistakenly drank your bong water.

12:57 AM, September 07, 2006  
Anonymous Anonymous said...

are you kidding me? AMD products was bad, real bad at that time. I'm not sure if the CPU used in the Tomshardware test was K5 or K6, which smoked when the fan is taken off the CPU and not able to function again even after put back, and that's the only x86 CPU in the history that would smoke (this is different from Sharikou's conspiracy theory and I beleive Edward would know this, and i believe you should still able to get this information out of the web). AMD start to gain market when it has good product (K8, i'm not sure about K7).

Ha! It is biased people like you that give AMD a bad name. You are wrong on both counts. It was neither K5 nor K6 in the Tomshardware smoking cpu test that registered over 200 degrees Centigrade. It was the Athlon or K7. K5 and K6 are socket 7 cpus and therefore not subject to Intel arm twisting in those days since Intel could do nothing to prevent someone from plonking a AMD K5/K6 into the motherboard. That was why Intel came up with the stupid Slot 1 and AMD decided to go with Slot A.

1)Socket 7 days: some of AMD K6 killed Intel's PIII rubbish thanks to a higher FSB (history repeating itself LOL). AMD's products were cheaper and better performing.

2)Incompatible motherboard days: Intel almost choked AMD to death with its crooked anti competitive arm twisting, AMD Athlon (K7) almost was not born although it had come as far as the womb's mouth. Bad product? The AMD Athlon beat the crap out of Intel's P4 (in fact, Intel's P4 actually was SLOWER than a Pentium MMX in some benchmarks so go figure whose product was real bad) and yes, it ran hot but I never lost a AMD cpu and I still have a Duron 700 running. The melting cpu was resolved with the next revision which is known as Athlon XP. All the Athlon's were hot processors but so are/were Intel's P4 crap. All you Intel fanboys can remember is that dumb Tomshardware video of a melted AMD Athlon but none of you remember/know of another video where the reviewer had to replace quite a few motherboards, and in fact gave up later, while trying to benchmark an Intel Pentium Extreme and I cannot remember where I saw to prove it too X-(. The replacements were due to the motherboards' capacitors being knocked out by the Intel hot iron and so those benchmarks never ran more than 2 hours before the box becoming unstable.

When I put together an Opteron server three years ago, the shops around here in Hong Kong could not believe that I was building a 10 disk drive, dual Opteron 242 box (8 IDE + 2 SCSI) with a redundant power supply that was only rate 300W. That box is still running and has only crashed once since I delivered the thing.

Are you kidding me? You know nothing of the history of Intel vs AMD. I have been through the 486DX 33Mhz till today and I know whose blinking products were/are better. After my Pentium 90Mhz cpu, I am never ever going to get an Intel box again. Not after all the crap they have pulled. You can join the ranks of the phony computer shop salesmen/dealers here in Hong Kong who only see Intel.

1:55 AM, September 07, 2006  
Anonymous Anonymous said...

if intel BKed, and it would be split into smaller companies, amd could pick the cherries out of them using the 15B...
i bet they would love to get their hands on the intel sw team, to optimize some compilers for them... :)
maybe even get some of the fabs, and make ati happy by making their gpus there..

..one can dream..

for now, i keep looking forward to the first amd cpus with bundled accelerators on die, hope to see them before the end of 2008. or at least desktop level htx/socket accelerators.

if half of the rumours are true, amd is executing intel right now, brad pitt style (from troya, the reocurring jump thing). amd64 is running towards, k8l is the leap, the accelerator stuff is the finishing blow. enjoy the show...

btw..remember when enron fell, the movies about it were kinda lame, cuz it missed some spice.
imagine a movie about the amd-intel struggle. it has a good guy turn bad (intel going on rampage killing competitors), a small guy take the challange and succeed (amd standing up, with the help of the killed competitors), with some romance in the transistor world (nvidia chipsets and the ati marriage...), and the golden spy hunt for the conspiracy. add some twists and witty dialogues, and here comes the oscar for sharikou! (who i, after reading the phone transcripts, hope would write the script)

oh..and dont take this too seriously. i post it for entertainment value mostly (mine in the first place..).

keep up the good work! (and start working on the script, if you have some spare time! ;) )

2:36 AM, September 07, 2006  
Anonymous Anonymous said...

"Do you think after Intel goes BK (and is unable to pay the $15B you claim they owe) due to AMD's success, that whatever is left of Intel can be reassembled into a company that can sue AMD for running them out of business?"

They wouldn't even have to do that. By causing their only competitor to go out of business, AMD will have broken a whole host of federal laws, and will be in the deepest sh*t with the DoJ.

4:05 AM, September 07, 2006  
Anonymous Anonymous said...

" pointer said...

AMD's failure to gain market share were due to Intel arm twisting and their propaganda.

are you kidding me? AMD products was bad, real bad at that time. I'm not sure if the CPU used in the Tomshardware test was K5 or K6, which smoked when the fan is taken off the CPU and not able to function again even after put back, and that's the only x86 CPU in the history that would smoke (this is different from Sharikou's conspiracy theory and I beleive Edward would know this, and i believe you should still able to get this information out of the web). AMD start to gain market when it has good product (K8, i'm not sure about K7).

Btw, I still give credit to the K5 design team though (despite it was running hot and slow) as it is the first in-house design that give AMD's ability to come out with later design. it is a feat by itself of such effort.

9:33 PM, September 06, 2006 "

Yes, its always so smart to remove the heatsink when the instructions tells you not to :D
very smart indeed!!

also you forgot that slot A kicked Pentium 2 and then III's all aboard? and they had to use their marketting to call that Mhz was the real importance ( then when AMD beat intel to the race of 1 Ghz they pulled back the statement? ) :P

indeed very smart MR. Point

6:54 AM, September 07, 2006  
Anonymous Anonymous said...

btw.. the inquirer posted some info claiming that shareholders were sure that 10,500 fired were not enought, so
Id expect way more layoffs soon.

6:56 AM, September 07, 2006  
Blogger Sharikou, Ph. D. said...

I was building a 10 disk drive, dual Opteron 242 box (8 IDE + 2 SCSI) with a redundant power supply that was only rate 300W.

I built a 3U server 2P Opteron server with 14 disk drives (12 hotswap, 2 internal) and a 460 watt power supply. Never had any problem running for 1 and half years up to today.

8:07 AM, September 07, 2006  
Blogger S said...

For years and years AMD has been in the Red. It never went bankrupt. In the past 10 years, AMD did not have 1/10 th as many profitable quarters as Intel.

So bankruptcy of Intel is much further on the horizon than AMDs.

9:03 AM, September 07, 2006  
Blogger pointer said...

Yes, its always so smart to remove the heatsink when the instructions tells you not to :D
very smart indeed!!


So, a faulty fan or a disconnected fan or misconfigure DIY will kill your CPU! You can't think of this? not so smart in indeed!! :D Anyway, it was history.

also you forgot that slot A kicked Pentium 2 and then III's all aboard? and they had to use their marketting to call that Mhz was the real importance ( then when AMD beat intel to the race of 1 Ghz they pulled back the statement? ) :P

wow, since when Intel P2 and especially PIII is bad (comparatively to AMD's)?? The K5 and K6 are inherently hot per the standard at that time. it was a known issue.

It was the Athlon or K7. K5 and K6 are socket 7 cpus and therefore not subject to Intel arm twisting in those days since Intel could do nothing to prevent someone from plonking a AMD K5/K6 into the motherboard.

Thanks. Now i know even part of the K7 was hot too. :)

So, after AMD come out with good chip K8, it picks up the market with time. (you need time to get acceptance and recognition)

And whining why intel used different slot? Why not? Did intel has to care about AMD on which slot it used? ATI stop future Intel chipset to focus on AMD is perfectly fine too. Stop whining and grow up.

9:28 AM, September 07, 2006  
Anonymous Anonymous said...

Sharikou said: "Intel owes AMD $15 billion, just wait till the lawsuit. I have talked about how to compute the damages before."

Too bad for AMD that Intel will BK soon, eh?


I am sure that Sharikou fails to see the irony in that. When you spin so many lies it gets hard to keep them from conflicting with each other.

9:30 AM, September 07, 2006  
Anonymous Anonymous said...

"I'm not sure if the CPU used in the Tomshardware test was K5 or K6, which smoked when the fan is taken off the CPU and not able to function again even after put back, and that's the only x86 CPU in the history that would smoke (this is different from Sharikou's conspiracy theory and I beleive Edward would know this, ... "

Hey... I don't know whether K5 or K6 smokes; that didn't happen to me (K6-2). However, you are wrong about these being the only x86 CPUs that smoke. FWIW, P-III and P-4 both would, and also K7. You have to be out of your mind to stop the fan of any chip consuming more than 30W without some form of internal protection. There is no rocket science here - just look up the processor's spec and see how much power it consumes!

BTW, K5 and K6 are fine products. They are 6-10 months late but they are fine for their intended competitors. Of course at their release Intel also released higher performing processors, but 1) that's why competition is good for customers 2) those higher performing Intel processors are considerably more expensive.

11:00 AM, September 07, 2006  
Anonymous Anonymous said...

"I guess the Chinese are trying to be more self sufficient and not have to depend on imperialist western technologies (lol)."

The Chinese are still light years away from making their own computers, even if they could make CPUs today.

How easy is it to run a MIPS based personal computer? Is it even possible? The Power-based personal computers with IBM and Apple backing and a group of highly dedicated/loyal users only had to die out recently.

Even if the Chinese make their own computers, it'll be less performing as a supercomputer, more expensive and less useful as a personal one. The smartest thing the Chinese should do is to support an open, global, and competitive CPU (x86, MIPS, Power, etc.) market. Yet, I guess they're not smart enough to see that.

11:11 AM, September 07, 2006  
Anonymous Anonymous said...

"AMD's "profits" over the last year is what? ~200Mil? and you're saying THIS is what enabled them to finance a 2-3Bil Fab?!?!? Not to mention this fab is significantly subsidized by the German government."

Aren't the profits what's left after AMD repay some debt (which was used to build the new fab) and expansion spending?

Had Intel been able to continue its monopolistic behaviors, AMD wouldn't have had profit for the past few years!

11:16 AM, September 07, 2006  
Anonymous Anonymous said...

graham said: "Sharikou said: "Intel owes AMD $15 billion, just wait till the lawsuit. I have talked about how to compute the damages before."

Too bad for AMD that Intel will BK soon, eh?

I am sure that Sharikou fails to see the irony in that.
"

If a chapter 11 is filed, not only AMD but the whole industry will suffer. The problem is not debt - which I believe the debtor is still obligated to repay, just later - but that chapter 11 would gives the failing company some operational advantages. In terms of Intel, it actually started the price war itself, and a chapter 11 would be the reasonable escape hatch if it could not keep on its overcapacity under the price war.

But the greatest loss will be by the shareholders, whose stocks are all nullified.

11:29 AM, September 07, 2006  
Anonymous Anonymous said...

wow, since when Intel P2 and especially PIII is bad (comparatively to AMD's)?? The K5 and K6 are inherently hot per the standard at that time. it was a known issue.

Who ever posted that thing about K7 wiping out PII and PIII is a total dipstick. The K6-2 350Mhz started wiping out a PIII with a similar clock rate. And no, K6-2 had no heat issues. For that matter, anything beyond a Pentium needed a cpu fan but nothing really fancy until the P4 and K7 were introduced. PIII, K6-2 and lower were not major furnaces.

And whining why intel used different slot? Why not? Did intel has to care about AMD on which slot it used? ATI stop future Intel chipset to focus on AMD is perfectly fine too. Stop whining and grow up.

You are a total blind fool pointer. Intel changed its cpu motherboard interface and also did not renew its patent licenses to AMD to cut AMD out. Of course Intel does not have to care...but in fact they DO CARE. Why on earth did you think Intel changed the slot/socket? So that YOU, the consumer, are forced to pick your entire box based on the cpu you want to buy. Like today's SLI/Crossfire. All this incompatibility means that we get less choice. If you could still get motherboards that could take an Intel cpu or an AMD cpu, I can tell you that Intel will get the short end of the stick because that was the reason why Intel had to come up with something other than socket 7.

We are the ones who are denied easy choices and therefore, I definitely am going to complain about Intel's lousy practices. Intel made it acceptable for Nvidia's SLI because this sort of thing has become common place. Locking out the competition on the basis of an arbitary socket/slot and not any particular technology. Ha! At least this paved the way for AMD to develop Hyper Transport. Which they are quite happily opening up to others. Love the Sparc and AMD on the same motherboard possibility. Intel can take its ball and play by itself in the corner. In fact, Intel better just leave the room and stay away.

12:04 PM, September 07, 2006  
Anonymous Anonymous said...

Locking out the competition on the basis of an arbitary socket/slot and not any particular technology.

So tthe socket/slot only is a physical connector with no technology involved whatsoever? Pin density, power supply, heat dissipation, whether or not it forces a PCB designer to move from 4 to 6 layer boards at substantial cost- none of these are technology drivers? Wow!

That's not to say there aren't marketing points factoring into the form factor changes, but to claim that is the only driver is a simpleton's world view.

Let's look at it another way- if AMD spent >$20M (warning: made up number, likely low) designing a new socket infrastructure (testers, packaging, PCB traces, validation and stress testing-the works), would you think they would be happy if Intel took advantage of their capital investment and switched to the same design? As in "Woohoo! Thanks AMD- we can just go order up these parts from all the suppliers and you paid all the startup costs. I owe you one dude!" Probably, AMD and AMD "supporters" not so happy. Now go switch AMD with Intel and vice-versa. Get my point?

3:13 PM, September 07, 2006  
Blogger pointer said...

Why on earth did you think Intel changed the slot/socket? So that YOU, the consumer, are forced to pick your entire box based on the cpu you want to buy.

Stop whining and be real to the business world. What on earth that a company A selling its product and yet 100% depends on its competitor to come out with a compatible platform for it to use? It is a BIG mistake and deserve to be a case study in MBA.

Hey... I don't know whether K5 or K6 smokes; that didn't happen to me (K6-2). However, you are wrong about these being the only x86 CPUs that smoke. FWIW, P-III and P-4 both would, and also K7. You have to be out of your mind to stop the fan of any chip consuming more than 30W without some form of internal protection. There is no rocket science here - just look up the processor's spec and see how much power it consumes!

BTW, K5 and K6 are fine products. They are 6-10 months late but they are fine for their intended competitors. Of course at their release Intel also released higher performing processors, but 1) that's why competition is good for customers 2) those higher performing Intel processors are considerably more expensive.


I was just lazy to look up the clip again and as someone pointed out, the chip was K7. Anyway, The K5 was more than 1 year late in schedule if not mistaken and plagued with manufactring issue. It was a FAILED product to the market (slow comparatively), a SUCCESSFUL product internally to AMD (for the first time, it has a compatible x86 in-house). The K6 make no much improvement either.

btw, the P4 may run hot, but it will not smoke, so do the rest of the Intel chips and latest AMD chip (partial K7 and all K8) due to the thermal throttling protection.

7:33 PM, September 07, 2006  
Anonymous Anonymous said...

Let's look at it another way- if AMD spent >$20M (warning: made up number, likely low) designing a new socket infrastructure (testers, packaging, PCB traces, validation and stress testing-the works), would you think they would be happy if Intel took advantage of their capital investment and switched to the same design? As in "Woohoo! Thanks AMD- we can just go order up these parts from all the suppliers and you paid all the startup costs. I owe you one dude!" Probably, AMD and AMD "supporters" not so happy. Now go switch AMD with Intel and vice-versa. Get my point?

Heh. Intel refused to license its stuff to AMD. I bet they would rather die than license stuff from AMD. Look at what Intel did once AMD bought up ATI. Intel forced AMD to come up with HyperTransport and whatever sockets they now use. Now Intel is trying to catch up rather than license HyperTransport. What makes you think AMD would not let Intel make compatible processors if Intel was willing to pay for a license? Not that Intel would ever do that.

8:03 PM, September 07, 2006  
Anonymous Anonymous said...

"So, Intel claimed that AMD should have borrowed a high interest loan instead of making money and using profit to support capacity expansion."

SHARIKOU First off INTEL ddn't claim anythng, this was my opinion. Second , AMD could have outsoured to foundry (like they are now doing to a small extent with Chartered) without taking out a huge loan, or they could have raised money thru other means (loan, stock, VC loan/warrants).

My point is if AMD had taken more risk they could have added capacity earlier instead of waiting to become marginally proftable first. Would AMD not have made back even the cost of high interest loan over the last yaers had tey been able to build more chips - this clearly is what has been limiting them over at least the last year, if not more

8:57 PM, September 07, 2006  
Anonymous Anonymous said...

"Aren't the profits what's left after AMD repay some debt (which was used to build the new fab) and expansion spending?"

Normally yes, but if I recall correctly (and someone please correct me if i'm wrong); AMD has yet to show any substantial equipment depreciation (whch is probably the largest expense and I think is estimated at something like $850Mil for 2006)

9:03 PM, September 07, 2006  
Anonymous Anonymous said...

"pointer said...
So, a faulty fan or a disconnected fan or misconfigure DIY will kill your CPU! You can't think of this? not so smart in indeed!! :D Anyway, it was history."

You are right that it was history.

FWIW, Intel's earlier CPUs DIDN'T have thermal protection either. Guess Tom chose a convenient time to do this test in the short gap after Intel introduced it, and just before AMD did ;)

12:15 AM, September 08, 2006  
Anonymous Anonymous said...

"Granted these companies are not Chinese, yet it's the same region. I don't see why China couldn't become a technological powerhouse considering South Korea, Tiawan, and Japan managed to."

The original comment used the Chinese "Godson-2" as an example of Chinese home-grown computing technology, which I believe is just not going to succeed. It is a yet another new ISA and all, and as I said, it will be low-performing for supercomputers and too expensive for PCs.

The examples you gave about other Asian countries are irrelevent, either. VIA bought the remain of previous Cyrix (after many hands and partitions), and to date AFAIK they still do all processor designs in the U.S. It is solely x86-based and those CPUs aren't particularly competitive compared to what Intel and AMD have to offer.

The other companies are all motherboard makers. A modern motherboard consisting of tens of ICs would worth about the same or less than the processor plugged onto it. To say a company making motherboards has the ability to make computers is like saying a company making car frames can make good car engines. It's just not true.

There is however a slight hope for the "emerging" Asian market to make x86 or MISP compatible processors for their particular needs. There is no hope that, like what China was/is doing, they can make an alternative processor (ISA) line successful. To do that, they would need a complete set of paripherals, compiling tools, architectural design and roadmap, and more.

Japan however has its own computing technologies which, although not comparable to those of the U.S., are good and solid. But even Japan would not succeed commercially or performance-wise with a custom-built proprietary system, and the difference between Japan and China/Taiwan/Korea is huge (at least presently).

1:47 AM, September 08, 2006  
Anonymous Anonymous said...

To Pointer -

AMD's K5 is technologically more advanced than its targeting competitor, the Pentium. It was late for more than a year, and it didn't have good FPU performance. Nevertheless, its IPC is better, and it didn't succeed commercially partly due to general public's belief in megahertz.

AMD's K6/K6-2/K6-3 are very competitive. They do not gain the absolute performance crown due to weaker FPU implementation (though 3DNow somewhat changed it later), but for the same price they perform better than P-II's. The problem was that AMD did not have enough capacity to make enough of them, nor did it have the marketing power! Basically Intel had monopolistic resources that kept K6-family being a real success.

Also, both P3 and P4 could fry. AthlonXP before the Barton core can also fry. I've seen/heard examples of all three. Had TomsHardware tried to they would've been to smoke not only the K7 but also the P3/P4 around the same time (early Thunderbird). Using TomsHardware's weak result as your proof that K5/K6/K7 suck is really lame, IMO.

2:21 AM, September 08, 2006  
Anonymous Anonymous said...

Stop whining and be real to the business world. What on earth that a company A selling its product and yet 100% depends on its competitor to come out with a compatible platform for it to use? It is a BIG mistake and deserve to be a case study in MBA.

Real business world = profits above everything else. That includes laws, 'ethics' and whatever else that is in the way or can be sacrificed. Revelation 11:18 shows who will rid humankind of this plague.

What ever happened to honesty and not seeking the downfall of others? Very little is for the benefit of others nowadays. 2 Tim 3:1-5. When will all mankind be united in peace and doing things for the benefit of all and not trampling on the dignity of others? Then we will see proper use of technology and not systems built for the sake of designing weapons or 'safeguarding' piles of WMD.

Yes, OT, and human governments and false religion will go too according to Daniel 2:44 and Revelation 18. When the Creator eradicates the three main parts of today's system, politics, false religion and big business, then we will see real progress when we start making good use of his physical laws under his direction.

2:28 AM, September 08, 2006  
Blogger pointer said...

AMD's K5 is technologically more advanced than its targeting competitor, the Pentium. It was late for more than a year, and it didn't have good FPU performance. Nevertheless, its IPC is better, and it didn't succeed commercially partly due to general public's belief in megahertz.

with all respect, i read the above statement as AMD K5 was inferior as compare to the intel product at that time. The K5 might as you said, have designed in higher IPC, but i do not know how effective it is. And again, frequency was KING at that time. This is not marketing myth but technicaly sound. A 10% increase in frequncy at that time, you would see clear difference in term of the system performance.

AMD's K6/K6-2/K6-3 are very competitive. They do not gain the absolute performance crown due to weaker FPU implementation (though 3DNow somewhat changed it later), but for the same price they perform better than P-II's. The problem was that AMD did not have enough capacity to make enough of them, nor did it have the marketing power! Basically Intel had monopolistic resources that kept K6-family being a real success.

few thing to say here, again, i read above as intel CPU had advantage of AMD CPU at that time due to FPU. And it DIDN'T change with 3DNow. How many software/game support 3DNow at that time compared to MMX? You can say intel has more resource in enabling the ecosystem to enable its succsee, but why use the word 'monopolistic'? Just because intel has a lot of money, did it prohibit Intel from using it to promote MMx, etc? Btw, to be fair, I do not know how 3Dnow as compare to MMX. I do know later AMD adopt MMx too (and subsequently SSE).

And about 'same price they perform better than P-II's'. Yes, high percentage this statement will be true as AMD was fighting using price at that time, instead of technology.

And how AMD tried to overcome the FPU disadvantage at later stage? it used 3 FPU instead of improving the FPU. (i'm really lazy on finding which AMD CPU did that). I guess this is one of the reason their CPU ran hot. Having 3 instance of FPU is not a bad idea if they have good FPU and manufacturing power (bigger die size). Coupled with not so good manufacuring at that time, how do you expect AMD can supply for the demand?

Also, both P3 and P4 could fry. AthlonXP before the Barton core can also fry. I've seen/heard examples of all three. Had TomsHardware tried to they would've been to smoke not only the K7 but also the P3/P4 around the same time (early Thunderbird). Using TomsHardware's weak result as your proof that K5/K6/K7 suck is really lame, IMO.

P3 and P4 are impossible to smoke without having their thermally throttling logic broken. intel implemented this technology since P6. AMD users might have to thank tomshardware of forcing AMD come out with a safer CPU. Btw, i'm not using Tomshardware as saying K5/6/7 is bad, but ust one instance of example why AMD CPU failed to gain significant market share at that time. Another instance will be what you said, FPU performance. And again ust to re-iterate, frequency DOES matter MOST at that time, technically.

8:27 AM, September 08, 2006  
Anonymous Anonymous said...

What makes you think AMD would not let Intel make compatible processors if Intel was willing to pay for a license? Not that Intel would ever do that.

Think like a businessman. Assume that Intel would want to license the technology. What would motivate AMD to allow them to do so? Certainly not the goodness of their hearts. The primary motivator is cash- AMD has a big need for cash in order to service/buy down their substantial debt load. They have been reducing it as of late by diluting equity (converting debt to shares), but shareholders rather see them achieve value from their technology. That said, AMD would need to determine the ROI of such a transaction with Intel- if they thought for a second that Intel +HyperTransport would substantially hurt their ability to generate a return on their K8/K8L/K9/K10^10 due to reduced competitiveness, do you think they would do it? For how much?

8:47 AM, September 08, 2006  
Anonymous Anonymous said...

"i read the above statement as AMD K5 was inferior as compare to the intel product at that time."

I really don't care how you read things. The fact is fact. K5's microarchitecture is comparable to Pentium-Pro (or P-II minus MMX). It surpass its class (Pentium). Or are you going to say P-II is bad because it was inferior to Alpha at that time?

"The K5 might as you said, have designed in higher IPC, but i do not know how effective it is. And again, frequency was KING at that time."

Oh, frequency was KING at that time? Whose king? Stupid buyers', I bet. At that time, cache size actually helps performance the most, followed by IPC, FSB speed, then clockrate. The IPC improvement is how P5 outperforms x486 and P-pro outperforms P5 at the same clockrate. Again, K5's microarchitecture is comparable to that of P-pro. K6's is better than that of P-II.

You kinda mix up marketing or commercial success with technical ones.

"How many software/game support 3DNow at that time compared to MMX?"

You are gonna love the truth, that is 3DNow didn't require a software compiled specifically for it to give performance advantage. At 3DNow's release, most games with few exception perform better on K6-3 than P-II.

Intel later changed the SIMD microarchitecture significantly to become SSE, with its own set of registers and all. Then it becomes so unique that other technologies could not take advantage of the SSE-compiled software. Figure yourself why Intel do that.

(p.s. It's not that Intel couldn't do a better SSE design. For example, AMD64 took a completely open approach to extend the ISA. I don't believe Intel was so incapable of something like that.)

"P3 and P4 are impossible to smoke without having their thermally throttling logic broken."

Well, then I've seen both a broken P3 and a broken P4. They were smoked, anyway, because the fan died. What you said "impossible" just happened.

"Btw, i'm not using Tomshardware as saying K5/6/7 is bad, but ust one instance of example why AMD CPU failed to gain significant market share at that time."

So you ARE using toms as saying K5/6/7 are bad for the market share, which they aren't. AMD lacked the marketing strength; AMD lacked the capacity; AMD lacked the resource for a price war. K6 was competitive enough, and K7 trashed P3 performance most of the time. Toms report and K5/6/7 design had nothing to do with their low market share.

"Another instance will be what you said, FPU performance."

You use little FPU except in heavy gaming. How much percentage PC users are heavy gamers? 5%? 10%? Weaker FPU doesn't forbid market share, unless amidst some misleading marketing from the competitors.

"And again ust to re-iterate, frequency DOES matter MOST at that time, technically."

Oh yeah, frequency matters, so does many other things. But it is by far not the most factor of performance. You just know nothing of computer architecture since the Pentium time if you think otherwise.

1:14 PM, September 08, 2006  
Blogger pointer said...

Oh, frequency was KING at that time? Whose king? Stupid buyers', I bet. At that time, cache size actually helps performance the most, followed by IPC, FSB speed, then clockrate. The IPC improvement is how P5 outperforms x486 and P-pro outperforms P5 at the same clockrate. Again, K5's microarchitecture is comparable to that of P-pro. K6's is better than that of P-II. ...

Oh yeah, frequency matters, so does many other things. But it is by far not the most factor of performance. You just know nothing of computer architecture since the Pentium time if you think otherwise.


please do not get confused on the chronologycal release of the CPU. AMD didn't able to release higher frequency than intel at that time. No matter how you claim AMD's microarchitecture advantage was, overall K5 and K6 is a inferior chip compare to the Intel CPU at the same time frame.

And again, to add in silicon feature sich as multiple FPU unit is a silly thing to do if you do not have the capacity. Similarly, you can do dual core x86 at 1990 too, but why neither AMD not intel has done that? Everything has a trade off.

you can claim at the same frequency that the AMD CPU at that time might beat intel's CPU by a bit because of the so call technologicall advatage in the micro architecture. But, can AMD manufacture those chip at the same frequency at a reasonanle volumn as compared to intel's at the same time frame? No! The CPU is not purely depends on tehe microarchiteure alone. It depends on the manufacturibility (inclisve of testability), capacity, and the technology environment such as the software.
I understand that there are other factor that increase the performance of a chip, but at that time, frequency is the MAJOR factor (with the superscalar, cache, etc already presented at that time - i do not expect a 386 running at the same frequency can beat the 486).

it really depends on which factor is the major bottleneck at a given time. Take the memory latency as current example. AMD resolve this by having IMC, and intel resolve this by having bigger and smarter cache. No matter how you want claim the FSB bottleneck, intel with its manufacturing capacity, the bigger and smarter cache not only resolve it, but beat AMD on that, statistically (do not look at single access, but the whole duration of the system usage).

Then, take the memory bandwith (which is critical to certain application, mostly server related), AMD resolve it with NUMA, intel temporarily pushing higher DDR frequency, while pursuing it's own NUMA with CSI. Here, AMD has advantage for system 4p and above.

At the coming future, GPU integration and multicore are key, because now everyone has the silicon budget. Can you claim if AMD has technology advantage if it come out with this 10 years ago, using very big die? No! I call it stupid.

9:18 PM, September 08, 2006  
Anonymous Anonymous said...

"please do not get confused on the chronologycal release of the CPU. AMD didn't able to release higher frequency than intel at that time. No matter how you claim AMD's microarchitecture advantage was, overall K5 and K6 is a inferior chip compare to the Intel CPU at the same time frame."

Let me just skip all your later comments which are quite irrelevant to our discussion. But here, there's nothing chronical that made K5/K6 architecturally superior than Pentium/P-II. K5's microarchitecture IS superior than Pentium, period. K5's price was comparable to that of Pentium, and although it was more than a year late, it was intended to compete and replace the Pentium.

You can't say K5 is bad just because it came out when P-pro was available. The same argument would've made Pentiums/2/3 all bad products because at those time frames Alpha or even Sparc was superior than them. You have to compare things of the same level, that is, production complexity and cost.

The K6 family complete very favorably against P-II and P-III, except perhaps FPU, with the exception of 3DNow. There is really no much ground to spin around when your focus is technical merits. Commercially, or market-wise, AFAIK, AMD didn't produce that many processors, not because of low yield or bad design, but because of limited capacity (# of fabs). These two reasons above mean world difference!

No matter how you spin it, what is fact is fact:
1) K5 was late.
2) K5 was microarchitecturally superior than Pentium, its supposed competitor.
3) K6 was superior than P-II, except FPU before 3DNow.

11:12 AM, September 09, 2006  
Anonymous Anonymous said...

"I have a CFA.. I have gone and visited a few of them jokers. Find out they would charge me 5% of my money to buy high load investments of sectors I already owned. It was laughable how ignorant people are easily impressed by people who put their titles in email, on business cards or signatures. Shut the F up you don't know anything comparing INTEL to Steel or GM is not relevant. TO compare so only shows how stupid you are."

I doubt you have a CFA, because if you did, you'd know what it stood for. Instead, you obviousley confused it with CFP, CLU, ChFC, or some other designation.

CFA's aren't in the business of selling investments with 5% loads or 1% loads.

The analogies to GM and Bessie Steel are entirely appropriate from a cash flow perspective. If you had a CFA, you'd know that.

Nice try trying to impress us with your knowledge. Next time, know what you're talking about.

12:05 AM, September 10, 2006  

Post a Comment

<< Home