Friday, July 07, 2006

Intel Core 2 self-fragged: stability and reliability problems with Woodcrest and Conroe

It's been 11 days since Intel launched Woodcrest, and the chip is nowhere to be found. However, before we could possibly get a Woodcrest to do some real drill, the Intel Core2 marchitecture reportedly fragged itself. It seems that in a rush to catch up with AMD, Intel has cut some corners with CORE2, and its next gen Woodcrest chips would come up with unexpected errors and run into undefined behaviours during normal operation*. We have seen reports that Woodcrest servers suffer from BSODs before. But we are talking about production Xeon 5100 systems delivered to US government for evaluation here. Woodcrest is definitely not your chip for running a nuclear reactor or a heart surgery.

You say, what about playing games? Well, Conroe reportedly posted some good frame rates, but no one has verified that Conroe was playing the games correctly. I am not even sure people verified that their Conroe got the right SuperPi digits. Intel quality is very much in question here. If Intel sent government of USA a bunch of bug laiden server chips and hoped to get millions worth of contracts, what kind of confidence can you have for Intel's consumer chips?

Since there wasn't any report of similar problems with Dempsey/Bensley, the Woodcrest stability and reliability issues appear to be intrinsic and specific to the Woodcrest/Bensley platform. Clearly, Intel America has a lot more server chip design experience than the Israeli folks, who spent all their careers patching up Pentium III.

So we have a bunch of amateurs who have been working on 32 bit laptop chips all their lives to come up with a 64 bit server CPU, we know it is not going to be easy. I personally expected Intel to announce delays for Conroe/Woodcrest. Since Intel's server chips are based on desktop designs, Intel normally launches corresponding server chips many months after the desktop part is out. For instance, the Xeon Paxville was launched eight months after Pentium D. Yet, the marketeers at Intel decided to push out a Conroe dressed up as Woodcrest Xeon 5100 in June, even ahead of Conroe launch. Such a reckless move is deemed to fail -- it's like politicians directing hopeless military efforts. Intel has lost the server market period, to challenge the advancing AMD forces with unproven troops will only invite greater defeat.

I won't be surprised if the Woodcrest has messed up some cache coherence logic in association with its complex memory preloading mechanism. Intel's competition, AMD, is a server CPU company from ground up.

I smell a Woodcrest recall soon.

I also expect Intel to send a pile of manufactured buggy Woodcrest chips and Woodcrest wafers into the dumpster - or maybe they get remarked as Conroe and sold in China?

Intel is so predictable.

* Please note that reported problem was not RAID alone. RAID was causing performance issues, but Woodcrest was crapping out all over the place. As INQ wrote: "Worse of all, several problematic situations occurred during the trial period and Intel was heavily criticised in internal memos, all ending up in a really unexpected manner."

78 Comments:

Anonymous Anonymous said...

Those problems are primarily in RAID, and there is no evidence (yet) of it effecting any other part of the system. Along with that, the first batch of chips being sold will probably go to Dell, as I have not looked at their site in quite a while (thank god), I do not remember them offering RAID set ups in their primary systems, so there is no problem for those initial Dell chips.

I am by no means trying to justify there mistake, but sales should not be a problem in the sense of "Conroe". Woodcrest on the other hand is a big problem, and I hope they resolve this quickly. They are supposed to be making a new stepping B2, cant remember where I say it though.

Glad to see your back in action.

11:27 PM, July 07, 2006  
Blogger Sharikou, Ph. D said...

Glad to see your back in action.
The problem with a blog is that there is only one person writing main articles, so when I am not connected, the story stops.

Anyway, I think Intel has stretched to limits. The CORE2 RAID problem is definitely something deep and intrinsic. But, what could it be? I think Intel knows what it is, and doesn't have a solution yet.

11:46 PM, July 07, 2006  
Anonymous Anonymous said...

Mmmhhh sound strange to me.
My opinion, my guess, is that Intel has stretched so much a old and archaic architecture that they are continuously hitting walls or facing new and unespected problems.
It's like buildind up a skyscraper over an old house, adding more and more floors: but the base is always the same, old and weak. You could patch here and there but in the end all will crumble or at least you find glitch and defects you cannot overcome.

12:16 AM, July 08, 2006  
Anonymous Anonymous said...

Does this only apply to RAID5 as this article states?
http://www.theinquirer.net/default.aspx?article=32818

If that's the case, then I don't see much to be worried about. Firstly, it doesn't affect the majority of computer users. Consumers at least. Secondly, it's good the problem was caught early on. Now some revisions can be done before the problem blossoms into something even more severe as the architecture gets deployed more commonly.

Also, what makes you think it's the CPU problem? It appears that the article I posted indicates that the issue resides in the southbridge. Before you criticize Intel's CPU, I think we should wait for more tests; ones that don't involve Intel motherboards.

But yes. You are right. This should not have happened. Intel should've been more careful. I agree with the first post; hopefully this issue will be solved quickly.

12:25 AM, July 08, 2006  
Anonymous Anonymous said...

Yep, maybe it's a southbridge problem or maybe not. Maybe it will be solved quickly but, for Intel, I think it's a big blow.
First of all, Woodcrest is a paper launch till now. We know they have a little time-window before AMD responds with its next step of Opteron; these news just stop anyone who might have bought woodcrest from buying it!!! Noone that uses its machine for critical data can afford to buy a server thata has raid problems.
Yes, if this is a CPU bug, it could be fix with a later revision but this means more time given to AMD... and that little window narrows.
IF a late june launch brings CPUs availability to some selected points only irst days of august, then a fix to this might be available not before october/novemeber: then it will be amd 65nm in a couple of months time.
Bad times for Intel.

1:16 AM, July 08, 2006  
Anonymous Anonymous said...

"it doesn't affect the majority of computer users. Consumers at least"

yea sure I'm gonna buy a crippled cpu..not! remember the divide by zero bug.. real trivial but no-one wanted a bar of those cpu's
Deep dodo Intel

1:58 AM, July 08, 2006  
Anonymous Anonymous said...

"Also, what makes you think it's the CPU problem? It appears that the article I posted indicates that the issue resides in the southbridge. Before you criticize Intel's CPU, I think we should wait for more tests; ones that don't involve Intel motherboards."

The new INTEL CPU does not work with well established computer components, and you claimed that it is not the CPU's fault but the others, even thought the same computer components work well with AMD's CPU.

That kind of behaviour was called finger pointing, it happens when things go deadly wrong. I can understand that.

US government has, offically, thrown Woodcrest CPU out. ASSUME Intel has fixed the problem and re-submit systems with new 'Fixed" CPU today, Intel can ONLY sell Woodcrest to US government in November.

At that time, AMD will be sampling their new CPUs. Woodcrest will face the AMD 4-core CPU. That is a very different battle.

2:40 AM, July 08, 2006  
Anonymous Anonymous said...

AMD guru says Intel Xeon shared cache inferior

5:57 AM, July 08, 2006  
Anonymous Anonymous said...

War of words between AMD and Intel techs over shared cache

6:03 AM, July 08, 2006  
Anonymous Anonymous said...

The issues on Conroe seem to indicate a real hurry on Intel's part to get the part out at all costs,even if it means some incompatibility issues. As the IT manager of my home and in need of a new system, I get really scared about the issues that have been found so far. I think with Conroe I would need to buy special and proprietary CD/DVD ROMs. Also, I can't configure any memory over 800 MHz and I can't do RAID 5. Gosh, what else will be next? I will be staying on the sidelines for a while.

6:19 AM, July 08, 2006  
Blogger C2 said...

May be, may be ...
finaly the the CPU may be working or may be don't!

I can't exclude the possibility this instability to be a result from some "GENIUS" performance optimization?!

But there is a reclame w\o product, so the situation is more than normal for Intel!

7:34 AM, July 08, 2006  
Blogger Mad Mod Mike said...

"The problem with a blog is that there is only one person writing main articles, so when I am not connected, the story stops."

You can add people to your member list and other bloggers can post on your blog as well.

9:08 AM, July 08, 2006  
Blogger Ajay S. said...

Slashdot has a seperate section for

dont know when it started though

9:56 AM, July 08, 2006  
Anonymous Anonymous said...

Sharikou, go and tell AMD about this BIG news and make sure to give them the link http://theinquirer.net as it is the only site has such exclusive information (after 4 days, and no other major sites follow up on the same story). And be sure to ask them to quote it in their web site, and inform them that you know it is the Core 2 issue. They will be happy enough to give you the 4x4 platform which you might hardly afford...

10:35 AM, July 08, 2006  
Anonymous Anonymous said...

Strange problems is what you are going to get when you decide a laptop chip -- one that makes your laptop explode -- is now a server chip too!

Well we know from the family suing Apple for $15,000 for a melting MacBook that ruined their carpet, that Intel's laptop chip is not much of a floor wax.

But Intel is still determined to convince the world that Woodcrest is actually a server chip. Make that a desert topping for Intel's rapidly bleeding bottom line. If Intel can fool the world and get people to buy the chip.

The BAD bugs that we find with Woodcrest clearly indicate that the real launch date of the Woodcrest -- vs. the LaptopChipWithSomeKillerBugs launch date -- is 2007.

There will probably be a new stepping coming up early 2007 of the CPU and this is the one to buy.

Only a fool would buy a CPU that did not undergo full testing before launch.

Even the early Opteron adopters got chips that did not work as well as those one stepping later. And those started out as real server chips. And they did not hose with RAID5. Because they were designed for the server, not for a laptop.

Realistically, it may take Intel a few steppings (not just one) to get their laptop chip all the way good as a server chip... or maybe it will have to wait until Core Duo 2 SE (Server Edition) redesign... sometime in early 2007.

For most buyers, it just makes sense to wait a bit until the bad bugs get found and fixed. It won't be overnight that a laptop chip becomes a server chip. Today we have a LapServer chip... that doesn't work with things you don't find on a laptop (like RAID 5). And as you know... there is a LOT of stuff in a server not found in a laptop. So more bugs will pop up for sure.

So 2007 will have working Woodcrest systems that make sense to buy.

And 2006 will have a lot of suckers buying machines with beta-steppings of their LapServer chip that Intel dumps on the market.

One really has to wonder WTF is up with Intel management. It just makes no sense at all to rush a frakkin laptop chip into the server market.

Of course it does explain why Intel went to market with that new batch of Pentium 4 BarnHeater servers (Dempsey).

It seems to indicate that Intel knew well in advance that the LapServer was simply not ready for prime time. And that shipping beta LapServer chips was the only way to forestall the entire server market moving to Opterons this fall.

11:48 AM, July 08, 2006  
Anonymous Anonymous said...

Only a fool would buy a CPU that did not undergo full testing before launch.

only a fool would ask for something, that isn't possible for any producer of extremly complex hardware...

12:12 PM, July 08, 2006  
Anonymous Anonymous said...

"only a fool would ask for something, that isn't possible for any producer of extremly complex hardware..."

You obviously do not know Intel's history. Intel used to be very very good about testing their servers. In fact, Intel's hard core server testing was legendary for its thoroughness.

Sadly today's Intel does not do the level of testing they used to do. Nothing even close.

That it why the government (read the original article) -- i.e. THE CUSTOMER -- had to find the bugs.

In the old days before Intel fired all their smart people, Intel would have found the bugs, NOT the customer.

If you want to be a moronic Intel customer (i.e. beta tester) then go ahead. If you have to justify it by making up some lies about how good testing is not possible, then you do that.

We, the smart buyers, will even thank you. Because it is always good to have more beta testers!!! :-)

12:21 PM, July 08, 2006  
Anonymous Anonymous said...

Come on, impossible to test.... ahahhah we are talking about a trivial Raid5 configuration, right??
I could agree if the bug is caused by a strange combination of a never-used part along with another that noone knows, but this is not the case.
In recent years, Intel's quality on desktop side was questionable, lots of errors and glitches and patches to wrong choices (Rambus RAM being the biggest)but now the virus is spreading on the server side. Too bad when marketing commands engineers.

12:47 PM, July 08, 2006  
Anonymous Anonymous said...

Blade servers mostly use onboard RAID controllers?

Then combine the above(?) with Woodcrest RAID5 problems(?) would mean Woodcrest might miss the fastest growing segment in the server space.
Standalone/Hardware RAID controllers will add a lot of expensive parts to the Woodcrest platform.

12:58 PM, July 08, 2006  
Anonymous Anonymous said...

I’m not buying into this. How could Intel be this fu**ed up. The darling of the worlds chip manufacturing producing award winning pre-fragged server chips! Surely they would test the cpu’s in all server configurations before release into the wild. Tell me it ain’t so. If true, the ramifications are huge, the biggest profits come from the server end of the business.

1:19 PM, July 08, 2006  
Anonymous Anonymous said...

mmm... Not a big deal...

If I were the one in Intel, I shall sell those server chips as laptop chips(where they are from). Then, Intel re-touts their next-version of next generation M..architecture CPU as server chip 12 months later.

Problem solved. Every one is happy ? Trust me, the official statements from Intel will be released soon.

1:45 PM, July 08, 2006  
Anonymous Anonymous said...

I am a validation engineer and worked for several high tech companies including HP and Xerox.. Let me educate some of you fanboys about validation:

As products become more complex and transistors beocme smaller and densor, validation becomes infinite space.. If you want to test every scenario and configuration possible, human race will disappear before you finish.. Companies such as AMD and Intel have to do some validation and most importantly, risk assessments. If you think that their is a high tech product that is 100% validated and bug free, you apparently have not worked in the industry. Bug escapes is one thing high techies tries to avoid, but total escape is unavoidable.. THat is why risk assessment is very important.

Here are some examples for you of bug escapes: AMD recent bug escape with around 3000 Opteron processors (with heat sensors problems), The continuing Microsoft patches, Car companies recalls, toy recalls ...

You obviously are highlighting a possible problem of an Intel chip because you are biased against the company and your blog is all about bashing this company.. (that is obvious from every article you posted and your referencing anti-intel bias sites) ..But if you think that the Opteron or the Athlon, as great as they are, do not have bugs, you must be blind.

2:34 PM, July 08, 2006  
Anonymous Anonymous said...

By the way, why do you call your site "Journal of the Pervasive 64 bit Computing" when you hardly discuss any "pervasive" computing (whatever that means)..

Isn't more appropriate to call it: "Girls do not like me", or "I have no life", "AMD fantasy", "Anti Intel fanbois", ...?

2:39 PM, July 08, 2006  
Anonymous Anonymous said...

"Woodcrest will face the AMD 4-core CPU."

K8L will not be out until Q4-2007, it will be a while.

Did anyone notice that on the DailyTech site, none of the new FX chips has reduced TDP?

Does this mean they are behind on their 65nm process?

Seems the switch should drop that down quite a bit, right?

Thanks.

3:21 PM, July 08, 2006  
Anonymous Anonymous said...

This is a great humor site. I luv it.
I wonder if anyone actually believes anything written on this site!?

3:37 PM, July 08, 2006  
Anonymous Anonymous said...

"Problem solved. Every one is happy ? Trust me, the official statements from Intel will be released soon."

looking forward to the Merrill Lynch report to investors on this one.

SPIN
"Looks like Intel's strategy may be working for now," Merrill Lynch said in a report to investors."

4:57 PM, July 08, 2006  
Anonymous Anonymous said...

You’re probably wondering why we have all gathered at this blog today; this post is in memory of Intel and those bothers who played SuperPi all day on their Intel big iron.

4:58 PM, July 08, 2006  
Anonymous Anonymous said...

"You’re probably wondering why we have all gathered at this blog today; this post is in memory of Intel and those bothers who played SuperPi all day on their Intel big iron."

I hear if you really ask Intel nicely they will give you a special version of SuperPi that has a big "I &hearts Intel" graphic... they call this special version "SuperBoi".

6:52 PM, July 08, 2006  
Anonymous Anonymous said...

I found an interesting post on another web site highlighting the weakness of the core architecture.
http://www.realworldtech.com/forums/index.cfm?action=detail&id=70202&threadid=70202&roomid=11

6:55 PM, July 08, 2006  
Anonymous george said...

But This archictecture is more complex than amd's so bugs are easyer to get http://chicagrafo.blogspot.com/2006/05/steam-engines-versus-diesel-and.html

7:36 PM, July 08, 2006  
Blogger Sharikou, Ph. D said...

I shall sell those server chips as laptop chips(where they are from).

Intel is having problem getting Conroe down to 35 watts

8:05 PM, July 08, 2006  
Anonymous Anonymous said...

Intel is having problem getting Conroe down to 35 watts

I do not know that you work for Intel, Sharikou. I have no such information even we are 'colleague'... Or most probably, 'someone' high up or down below whispered to you while you were dreaming ...

9:27 PM, July 08, 2006  
Anonymous Anonymous said...

"I do not know that you work for Intel, Sharikou. I have no such information even we are 'colleague'... Or most probably, 'someone' high up or down below whispered to you while you were dreaming ..."

It is well known, even from the Intel paid pumper sites, that Conroe needs big heatsinks. The chip runs very hot (due to bad design).

I would not be surprised if not only does Conroe run hot but runs on more than 35W. I would guess Conroe at full tilt is dragging down 75W, about 100% over spec.

10:13 PM, July 08, 2006  
Anonymous Anonymous said...

By the way, why do you call your site "Journal of the Pervasive 64 bit Computing" when you hardly discuss any "pervasive" computing (whatever that means)..

Isn't more appropriate to call it: "Girls do not like me", or "I have no life", "AMD fantasy", "Anti Intel fanbois", ...?

Why does "The Journal of Pervasive 64-bit Computing" love AMD so much? Answer: because AMD chips can actually do real 64-bit computing. Intel chips don't.

And why would an Intel fanboi like you resort to name-calling? Answer: desperate times need desperate measures.

11:16 PM, July 08, 2006  
Anonymous Anonymous said...

It is well known, even from the Intel paid pumper sites, that Conroe needs big heatsinks. The chip runs very hot (due to bad design).

I would not be surprised if not only does Conroe run hot but runs on more than 35W. I would guess Conroe at full tilt is dragging down 75W, about 100% over spec.


ha, i'm not sure which site claim that conroe run very hot and it has 'bad design' beside this blog.

secondly, the normal Conroe TDP is 65W, not 35W. Sharikou was refering to Meron version (implicitly, for once, i defend for him that he knows this fact). Anf for you that do not know anything, please do some research before commenting. Btw, you know how to google?

12:06 AM, July 09, 2006  
Anonymous Anonymous said...

"You obviously are highlighting a possible problem of an Intel chip because you are biased against the company and your blog is all about bashing this company.. (that is obvious from every article you posted and your referencing anti-intel bias sites) ..But if you think that the Opteron or the Athlon, as great as they are, do not have bugs, you must be blind."

Following logic, you are suggesting that it is common and acceptable to have buggy Intel chip.

mmm... I gusse you are well educated in Intel's camp.

In the name common practice, I proclaim that thee computer vendors should buy Intel chips regardless of the quality, NO MORE complaints will be allowed.

btw, 65W Laptop CPU!? guess I am going to have a 65-nm laptop heater this X'mas.

2:54 AM, July 09, 2006  
Blogger C2 said...

Hey just look at this!

http://www.warp2search.net/modules.php?name=News&file=article&sid=29021

Actualy for a CPU isn't enough to be fast, it also must be stable :)

"Intel hints that Conroe is going to release at B-2 Stepping as Intel Core 2 Duo processor. As for the previous version, a problem was found to make the system full loaded. It’s only solved in the new stepping. We don’t encourage anyone to buy the engineering sample from the web. The retail version is going to release in the end of this month, and it’s much stable. "


"much stable" ?

actualy how much?

3:47 AM, July 09, 2006  
Anonymous Anonymous said...

So, we know that Conroe/Woodcrest given out for testing were bugged and not destined to sell.
Now I doubt those chips were prototype ones tweaked for SuperPi and benchamrk calculations. Now I understand WHY you couldn't have access to those Intel-prepared machines. They would have crashed.

6:49 AM, July 09, 2006  
Anonymous Anonymous said...

Hey, I just noticed another post at this forum: http://www.realworldtech.com/forums/index.cfm?action=detail&id=70211&threadid=70202&roomid=11

Seems someone that works here posted a bunch of junk there, and then claimed to be an anonymous poster that innocently found some anti-Intel dirt, and failed to mention the original source, which also included some AMD dirt. I can't imagine how such a silly mistake could occur! Think anyone here will fix that? Oh yeah, I just did. Let's see if this comment is allowed to be posted.

8:15 AM, July 09, 2006  
Anonymous Jeach! said...

"You obviously do not know Intel's history. Intel used to be very very good about testing their servers. In fact, Intel's hard core server testing was legendary for its thoroughness."

Funny... you obviously don't know Intel's history because Intel has never produced a server (at least a real one)!

Oh, do you mean those 32-bit Xeon's? I can tweak my old Nintendo and call it a server too, but it doesn't mean its a server!!

If there had not been any 64-bit servers in the last 10 years then I could say that Intel was a true server player, but until they can actually do native 64-bit computing... it's just a 'toy' server company!

1:09 PM, July 09, 2006  
Anonymous Jeach! said...

Sharikou, I don't quite understand how your blog has mutated from a simple isolated RAID problem to the assumption that the Core architecture is flawed all around. Time will tell on the seriousness of this issue (although Intels silence is not a good indicator).

You use to post facts! Or when you posted theories, you at least attempted to back yourself up with great logic and/or data.

If you know more on this, we'd all be interested on hearing it!

1:23 PM, July 09, 2006  
Blogger Sharikou, Ph. D said...

Sharikou, I don't quite understand how your blog has mutated from a simple isolated RAID problem to the assumption that the Core architecture is flawed all around.

Read the second report by INQ, it didn't attribute the faults to RAID alone, what it said was Woodcrest was crapping out all over the place. In the first report, INQ reported performance problems with RAID5, nothing else. In the second, there were far more problems. Also, I reasoned that since the same problems didn't occur with Dempsey, the problem must have something to do with Woodcrest. These faults were consistent with earlier observations of BSODs in Intel's lab.

2:33 PM, July 09, 2006  
Anonymous Anonymous said...

Sharikou, do you see Apple choose
AMD in the long run?
How will they act now, that is, if there is a real problem with Intel? (How will the rest of the world acually?)

4:26 PM, July 09, 2006  
Blogger Sharikou, Ph. D said...

do you see Apple choose
AMD in the long run?


Apple is a small fish, it's irrelevant now.

4:28 PM, July 09, 2006  
Anonymous Anonymous said...

AMD “This Bud’s for you”.
Intel engineers have had four years to best the Opteron. It seems they have squandered a lot of time. I personally was hoping for some great new science built on 65nm. Now were debating about an overly complicated cpu that has yesterdays architecture and needs monster size cache just to compete. Post launch we have BSOD and RAID problems. What’s wrong with this picture?

5:10 PM, July 09, 2006  
Anonymous Anonymous said...

You use to post facts! Or when you posted theories, you at least attempted to back yourself up with great logic and/or data

Thanks for the joke of the day :)

7:37 PM, July 09, 2006  
Anonymous Anonymous said...

Time will tell on the seriousness of this issue (although Intels silence is not a good indicator).

do you realize that INQ is the only site reported that? I am just wondering why no other site follow up on the same story? Just go and do a google news search and you know what i mean.

While i do not know for sure the news is fact or fiction, just to be sure to differentiate 2 chips, the woodcrest and the conroe. while they are having the same architecture, they do no behave excatly the same as one for the server workload and one for the desktop client workload. and one is a launched product as of today, and another one is soon to be launched product.

7:55 PM, July 09, 2006  
Anonymous Anonymous said...

however, those wallstreet analysts are still recommending buy on intel and intel fanboys everywhere are wetting their pants in anticipation of conroe and woodcrest.

8:01 PM, July 09, 2006  
Blogger Sharikou, Ph. D said...

do you realize that INQ is the only site reported that?

INQ is a British site, UK's case laws allow them to publish the information they got without fear of lawsuits from Intel. The INQ story provided enough detail on the problem. If you can get hold of a Woodcrest, you can verify it yourself. But Woodcrest is nowhere to be found.

Now, you understand why Intel only allowed controlled tests on Woodcrest. Intel knew the problems, and they were trying very hard to hide them and hope to find some suckers who are willing to pay for substandard products.

9:47 PM, July 09, 2006  
Anonymous Anonymous said...

When I google Intel, it seems most articles declare Intel’s price cuts are working and most believe Intel will regain market share and most put a buy on Intel shares. AMD has announced a 9% decrease in sales to be expected Q2. Dr Sharikou, do you need to reassess any of your predictions or will time change the outcome? Intel has a huge stock pile of outdated cpu’s that they are most likely glad to dump at any price; that’s a lot of ammo.

Be Well

10:51 PM, July 09, 2006  
Anonymous Anonymous said...

Funny... you obviously don't know Intel's history because Intel has never produced a server (at least a real one)!

Oh, do you mean those 32-bit Xeon's? I can tweak my old Nintendo and call it a server too, but it doesn't mean its a server!!

If there had not been any 64-bit servers in the last 10 years then I could say that Intel was a true server player, but until they can actually do native 64-bit computing... it's just a 'toy' server company!


Intel has produced many white box servers, server boards, server cases, server components, etc. Intel has done many major server designs for its various server customers and partners.

The testing that used to put into all these (and more) "server" systems was legendary for its thoroughness.

I used to work with Intel (not for them). I've seen the testing labs and talked to the test engineers.

As most of the servers that exist in the world today are 32-bit Intel servers, I'll let intelligent readers decide whether or not your Nintendo is a server. Sheesh.

11:02 PM, July 09, 2006  
Blogger Sharikou, Ph. D said...

AMD has announced a 9% decrease in sales to be expected Q2. Dr Sharikou, do you need to reassess any of your predictions or will time change the outcome?

It seems to me that AMD successfully transitioned to AM2 during 2Q in a rather short period of time. AMD's 1Q06 revenue was $1.33b, 2Q06 is $1.215, that is a 8.6% drop from 1Q, but year/year, it will be a 50% growth. Did Intel gain market share? Let's wait and see Intel's results. Some analysts projected a $7.8b revenue for Intel's 2Q06, a 12.3% billion drop from 1Q06's $8.9b, a 15.5% drop from 2Q05's $9.23b. AMD is gaining substantial amount of market share.

From my observation, Intel's pricing has completely collapsed already. Those P4 D 805s are selling at $112. You need to understand that Intel's ASP was $150 in Q1. So these processors are selling at 25% below its previous ASP. I project that Intel's 2Q06 revenue will be even lower than $7.8b.

Going forward, Intel's 3Q06 revenue will be even lower. Based on this Woodcrest reliability report, Intel's architecture is fundamentally flawed.

12:42 AM, July 10, 2006  
Anonymous Anonymous said...

"Why does "The Journal of Pervasive 64-bit Computing" love AMD so much? Answer: because AMD chips can actually do real 64-bit computing. Intel chips don't."

So Why the self proclaimed doc made it his mission to bash intel rather than just promote how great AMD procs work in 64 mode?????

1:09 AM, July 10, 2006  
Anonymous Anonymous said...

"Following logic, you are suggesting that it is common and acceptable to have buggy Intel chip."

Apparently you did not get it Doc.. Let me speak slower: All product in the world come with bugs. N O P R O D U C T I S B U G F r e e!

That includes AMD's, Intel's, HP's, Xerox, Sharikou's, ... This is just real life sonny boy..

1:16 AM, July 10, 2006  
Blogger mork said...

*Clearly, Intel America has a lot more server chip design experience than the Israeli folks, who spent all their careers patching up Pentium III.*

Ang i guess that you, ohh so very great and mighty sharikou with your so called ph.d and big know-it-all about science can do a better job? Give me a break dude. The Isreali-team has done a remarkable job with both Pentium M and Core tech.

This is second blogg entry from you where u bach out at the Isreali, im rally beggining to think that you are a antisemiat.

And on a diffrent level. Ive just read the lastest reviews from mathbe (french test site). Again Core beats the living hell out of s939 and AM2 cpus. What are you gonna tell us now, that all sites are bought by Intel and that this is a big concpiracy against AMD? Give us a break with this shiet now and accept the fact that Intel has finally after a few years in the dark finally made great cpu tech.



PS: While are responding, maybe u should visist swec.com ;

http://www.conman.se/pi3479.JPG

http://www.conman.se/aquac3420c635c760.JPG

http://www.conman.se/05c3420c635c760.JPG

For those that dont do 3Dmark and SuperPI thats AMD being owned :)


Grettings from sweden btw
/MOrK

4:00 AM, July 10, 2006  
Anonymous Anonymous said...

That’s what happens when you release a new processor on old (netburst) platforms.
If core 2 isn’t compatible with current Motherboards (platforms), why didn’t Intel released new chipsets and sockets for their new Processor?

And about other Intel problems, is the Core Duo USB battery problems solved?
Is the problem from the OS ?
If it is does anyone tested with Turion and other older Intel systems to see of the problem also exists there?

4:16 AM, July 10, 2006  
Anonymous Anonymous said...

By now I am sure most of you have already read Reverse hyperthreading does not exist.

7:05 AM, July 10, 2006  
Anonymous Anonymous said...

DigiTimes also joins the group of paid pumpers!!!

Can you believe what they have reported now?

http://www.digitimes.com/mobos/a20060710A1001.html

NO AMD K8L processors until 2008!!!

We all know K8L will be out by December this year...

Don't we?

7:42 AM, July 10, 2006  
Anonymous Anonymous said...

Wow. Reverse Hyperthreading does not exist!!!

http://theinquirer.net/default.aspx?article=32885

The author is a paid pumper, I guess.

7:43 AM, July 10, 2006  
Anonymous Anonymous said...

AMD seems to have some problems too!

A Gruesome Roadmap...
http://www.overclockers.com/tips00991/

I hope this roadmap is a fake else AMD is in deep trouble.

7:47 AM, July 10, 2006  
Anonymous Anonymous said...

AMD is going down..!

http://www.theinquirer.net/default.aspx?article=32885

8:28 AM, July 10, 2006  
Anonymous Anonymous said...

"A Gruesome Roadmap...
http://www.overclockers.com/tips00991/"


newsflash. Ed is an intel paid pumper, period. i used to like reading his commentaries because they were constructive, now it's just garbage. his nose is so way up intel's ass that you can smell the stench all the way from new york where he lives. yea yea, let conroe enjoy its glory days(sorry it won't last long). it will be a good time to upgrade with amd next year for me. i couldn't help but laugh when i read those xtremesys forum threads about the availability of conroe. i must admit, eventhough these fanboys are diehard, they do have a sense in checking out the gouged prices. unfortunately, there are some that simply can be rescued from wasting money.

10:06 AM, July 10, 2006  
Anonymous Anonymous said...

Re: Apple and Intel/AMD

Apple took the money from Intel to make the switch, so they need to be nice for a while, presumably.

From here on, they would be very wise to play Intel and AMD against each other.

10:12 AM, July 10, 2006  
Blogger Sharikou, Ph. D said...

newsflash. Ed is an intel paid pumper, period.

Pumping is useless. At the end of the day, what matters is what product you can sell. Intel made all their current products total crap. That's the reality. There is the fantasy of Woodcrest/Conroe, I predicted that it would be fragged in real performance tests. But Intel is playing hide-n-seek and denying us the oppurtunity to frag them.

11:35 AM, July 10, 2006  
Anonymous Anonymous said...

this is a bit offtopic, but i'm currently testing windows vista beta build 5456 ultimate edition. i'm starting to like vista alot and it runs great on my x2 4400. i also dual boot with winxp 64bit edition and i just can't go back to using winxp pro anymore. got a little story to tell. a friend recently built a whole new system. this guy has been an amd user for a long time, but decided to build an intel rig. he bought a 940D, a gigabyte board(forgot chipset #), and a 7900gt. he got it up and running fine, but was having frequent bsod. he figured that his 400W psu couldn't handle the load so he went out and got an ocz 600W psu. long story short, he has been down for a day since he since he switched the psu. he even called me up, but i couldn't do much for him. i also couldn't help but laugh. the story continues and i will report back the progress when i hear from him.

12:28 PM, July 10, 2006  
Anonymous Graham said...

"But Intel is playing hide-n-seek and denying us the oppurtunity to frag them."

Sharikou, as I have already mentioned to you, go out and get a server from Dell.com with a Xeon 5100 series processor. That can be your test box. Why do you insist INtel is playing hide and seek? All you need to do is plunk down some cash and put your money where your mouth is.

1:11 PM, July 10, 2006  
Blogger mork said...

*Pumping is useless. At the end of the day, what matters is what product you can sell. Intel made all their current products total crap. That's the reality. There is the fantasy of Woodcrest/Conroe, I predicted that it would be fragged in real performance tests. But Intel is playing hide-n-seek and denying us the oppurtunity to frag them.*

News flash for you. A so called "crappy product" like Core Duo T2600 makes a 400$ more expensive FX-60 look weak, not to mention what a OC`d T2700 does to a 1100$ overpriced FX-62. Its time you learn to set the facts straight.

If Intel core technology was so crappy, dont you think atleast one testsite would say this? Or are all of them bought by Intel? We have word for that in sweden, and i do believe its the same in english, and that word is "paranoid".

12:11 AM, July 11, 2006  
Anonymous Anonymous said...

"If Intel core technology was so crappy, dont you think atleast one testsite would say this? Or are all of them bought by Intel? We have word for that in sweden, and i do believe its the same in english, and that word is "paranoid".

Sites that are unpaid are showing Woodcrest and Opteron doing better than the other on different tests. Neither is the clear-cut best at everything vs. the other.

Yes, Intel has gone back to the Pentium Pro design and found it really was a good design. They added some improvements and have an architecture that is similar to AMD's and also similar in overall performance.

While Intel has done better on performance than they ever have, they also have a chip that has ominous bugs, including RAID crash and data loss bugs. The chip is not ready for prime time even if it does run programs that fit in its cache very quickly.

Now let us wait and see AMD's newest chips and see how they compare. We are comparing Intel's latest to AMD designs that are over 3 years old (shipping) and over 5 years old (on paper).

2:01 AM, July 11, 2006  
Anonymous Anonymous said...

"We have word for that in sweden, and i do believe its the same in english, and that word is "paranoid"."

wow a dumb swede thats "never" been seen before ;) greetings from ur neighbours the Danes, we just love making fun of those crazy swedes :P.. it seems intel are a bit desperate in many of theire acts, but it would be about time intel had something that could look nice.. for once ^o).. but no matter what, scaling and server is AMD's corner, it always has been since Opteron architecture. It also does seems suspecius, i can se that Core Duo is pretty good, but how does Core2 become extremely more powerfull?, maybe this is the answer, ;), i dont really care anymore, i buy AMD and always will, i've never gotten an unstable CPU or a slow computer when using AMD, but my laptop is a piece of crap(3GHz P4)

3:00 AM, July 11, 2006  
Anonymous Anonymous said...

"Now let us wait and see AMD's newest chips and see how they compare."

The new chips will come only in 2008 as K8L... the present K8s will be around till then...

The K8s were always compared to the Netburst which were 3 years old when they were released.

4:34 AM, July 11, 2006  
Blogger mork said...

*Sites that are unpaid are showing Woodcrest and Opteron doing better than the other on different tests. Neither is the clear-cut best at everything vs. the other.

Yes, Intel has gone back to the Pentium Pro design and found it really was a good design. They added some improvements and have an architecture that is similar to AMD's and also similar in overall performance.

While Intel has done better on performance than they ever have, they also have a chip that has ominous bugs, including RAID crash and data loss bugs. The chip is not ready for prime time even if it does run programs that fit in its cache very quickly.

Now let us wait and see AMD's newest chips and see how they compare. We are comparing Intel's latest to AMD designs that are over 3 years old (shipping) and over 5 years old (on paper).*

Ive actually not been able to find one site that gives AMD the preformance crown, when comparing against core 2 duo (comparing with am2). Well, AMD normally wins Synthetic test like Mathematica due to to its bandwith. Real life test like compression and encoding goes to Intel. Prove me wrong on that one!

Similar in preformance? I think not. Even a cheap E6600 gives a 1100$ FX-62 a match in most benchmark test, not to mention test in encoding divx and mp3.

Agreed, lets wait and see on the launch date of Conroe to se who`s the winner, NDA gets lifted on 28th i think. ill bet u an old isa network card with some coke on it that Conroe will prevail.

On a diffrent subject. Nice to have a resonable guy to argue with here, most ppl are fanatics.

/Mork

Ps: Just to point out, im not a Intel fanboy, i actually use a Opteron 144 as my current gaming comp. I just buy have has the best gang for the buck atm

5:25 AM, July 11, 2006  
Blogger Steel Smack said...

"The new chips will come only in 2008 as K8L... the present K8s will be around till then..."

INQ is reporting K8L is still due out in 1Q07.

http://www.theinquirer.net/default.aspx?article=32948

10:28 AM, July 11, 2006  
Blogger mork said...

*wow a dumb swede thats "never" been seen before ;) greetings from ur neighbours the Danes, we just love making fun of those crazy swedes :P.. it seems intel are a bit desperate in many of theire acts, but it would be about time intel had something that could look nice.. for once ^o).. but no matter what, scaling and server is AMD's corner, it always has been since Opteron architecture. It also does seems suspecius, i can se that Core Duo is pretty good, but how does Core2 become extremely more powerfull?, maybe this is the answer, ;), i dont really care anymore, i buy AMD and always will, i've never gotten an unstable CPU or a slow computer when using AMD, but my laptop is a piece of crap(3GHz P4)*

You call me dumb? Yet you say your self that Core technology is powerfull. Only stupied thing in the computer world is fanboys, why in the living hell would someone pay more for less? Its the product that offers best bang for the buck that is of intrest and nothing else. I dont give rats ass if it called Intel, Via or AMD.

Btw: My father happens to be danish, so im not a full blood swede, shitface :) Besides that obvious fact, I also live in the most southern part of sweden (read skåne) and dont real call myself a swede, we have more in common with you danish ppl then the once living in sealgulf land up north.

4:33 AM, July 12, 2006  
Blogger mork said...

"You say, what about playing games? Well, Conroe reportedly posted some good frame rates, but no one has verified that Conroe was playing the games correctly. I am not even sure people verified that their Conroe got the right SuperPi digits."

I can have a hell of alot ppl confirm their results in both games, superpi and 3DMark, not to mention divx and mp3 encoding. Maybe should take a trip to xtremesystems or sweclockers among others.

"ohh no say it aint so"

http://conman.se/crystalc3153c630c720.JPG

http://conman.se/05c3153c630c720.JPG

Want better scores then maybe u should join xtremesystems forum, dont think they would appriate you laying so much.

"I personally expected Intel to announce delays for Conroe/Woodcrest."

Strange. Many europeen retailers are allready (2 weeks in advance) shipping conroes.. ohh that must hurt..

"to challenge the advancing AMD forces with unproven troops will only invite greater defeat."

You used intel-nazi analogy not long ago, and describing this as horrible behavior. And yet for your own part you use Nazi-like propaganda here. How does this add up? *fanboy caugh fanboy*

8:01 AM, July 12, 2006  
Anonymous Anonymous said...

"Now let us wait and see AMD's newest chips and see how they compare. We are comparing Intel's latest to AMD designs that are over 3 years old (shipping) and over 5 years old (on paper).*"

Ok... by your logic the new memory controller for the ddr2 is not making it a new chip, just revamping it correct?

So if I go by your logic, we could also say that core architecture is just a revamped P3, right?

Well looking back it would seem that the P3 which came out in 1999... 2006-1999=7 years. Does that mean a (by your logic) seven year old design is kicking the crap out of a (by your logic) a three year old design.

Sure would be embarrassing if that were the case, wouldn't it?

But its not, AM2 is by all accounts a new design.

To compete with Conroe they have some work to do, and there are so many rummors floating around about K8L, you cant be sure it will be out before the end of 2007.

You have to compare whats out there vs what your best is. That is all Intel is doing, pitting Conroe up against AMD FX62, and you AMD fanbois just dont like the fact that they are winning, so stop making excuses and grow up... AMD will be back, and it only benefits the consumers.

12:45 PM, July 12, 2006  
Anonymous Anonymous said...

/Mork... Sorry about the last post I put up, I cant stand reading this comment...

"Now let us wait and see AMD's newest chips and see how they compare. We are comparing Intel's latest to AMD designs that are over 3 years old (shipping) and over 5 years old (on paper).*"

Was that your comment?

If not, it only applies to who wrote it.

Again, sorry.

12:56 PM, July 12, 2006  
Blogger Dr Blog said...

This comment has been removed by a blog administrator.

1:26 PM, July 15, 2006  
Blogger Dr Blog said...

I read above that if only RAID 5 was not working it wouldn’t be such a problem.
RAID 5 is RAID 0 with parity and becomes a RAID 0 when a drive is lost. I believe the only RAID array that will work on the Conroe is a mirror.

I've been using AMD since the K6-2\500 and have never seen a "BSOD" that I didn't cause. (Wrong drivers) Stability problems on a CPU not even on the market are a huge problem.
(I’m running Vista 64 beta with 4 drives in RAID 0 config, using modded x64 raid drivers someone on the net hacked, and the only problem so far was when I tried to burn 4 DVD's at once. I had to close explorer with task manager, but no "BSOD". This may very well be "Farewell to Intel"

Roger

1:27 PM, July 15, 2006  

Post a Comment

Links to this post:

Create a Link

<< Home