TOP Intel engineers leave INTEL.

[Scali]

AMD was down a bit when I made my last rig, now they're up a bit and look to have a good solid lead for the time being.

For the time being, indeed. Ofcourse, the Athlon64 is the newer architecture, competing against the P4 which already was around for a long time. Next year Intel will be the one with the newer architecture, which will probably put them up at the least until AMD gets a new architecture out. The only thing is... Will Intel ever make such big mistakes again? Somehow I doubt it.

 

[jon67]

Jugding by sales and profit I don't think Intel sees the P4 series as a "mistake", although it may have reached the end of the line.

Regarding R&D strategies, this is actually a critical point for AMD. Intels 2004 R&D budget was 5 times higher than AMD's, i.e. 5 G$ and 1 G$ respectively. The R&D costs constitutes 14% of Intels total turnover and 20% of AMD's. The problem for AMD is that the amount of R&D work required to come up with new, competitive products is more or less constant, regardless of how many units you sell. Much due to their R&D costs, AMD's profit margins are so low right now they need to increase the production volume and sales, especially in the OEM market. And knowing that Intel each year can give away for free a number of processors corresponding to AMD total annual production and still make a healthy profit, this is not going to be easy. For the sake of competition, let's hope AMD makes it through 2005 with their heads above water (financially speaking).

 

[SSE4]

Your example is consistent with taking the humanity away from the humans pertaining to the argument (we can all agree a pretty silly exercise). Taking emotion out of the argument means analyzing the situation as objectively and honestly as possible, avoiding as much bias and conclusion-jumping as possible. The sarcasm built into the first few lines of my previous post are aimed at this sort of 'emotional' judgement.

Allow me to give you a tour of these words:
Emotion: "2 a : the affective aspect of consciousness : FEELING b : a state of feeling"

Consciousness: "b : the state or fact of being conscious of an external object, state, or fact : 2 : the state of being characterized by sensation, emotion, volition, and thought : MIND"
There was hardly any emotion in their posts, just misinformed opinions that lead to the emotion. You seem to assume that emotion is (One of it's meanings) some kind of feeling of aggravation of some sort, or letting emotions "cloud one's judgement". You're simply wrong, as having emotion in an argument can be a good thing as I had mentioned.

I'm not saying that some "emotional debate" can't be out of hand and completely wrong, but you're making the gross assumption that emotion scars all debate. Which is a horrible exaggeration of the truth, and is an estimation you can't make without being biased/invalid yourself.

And it wasn't that the entire post was made for you, despite me choosing to point out some error, therefore I never really opposed much of the opinion expressed in your post.

 

[Vapor1000]

oh god... why did anyone even post this....
it looked interesting until i saw it was from the inquirer....

 

[RawsonDR]

You seem to assume that emotion is (One of it's meanings) some kind of feeling of aggravation of some sort, or letting emotions "cloud one's judgement".

We're talking about different things. You're absolutely right in your points that the emotion one feels as a result of a debate is justified and necessary. This can be described alternatively as a passion for the subject matter. As you point out, having emotions tied into what you firmly believe is simply human. After all, you believe it for a reason.

However, including emotion in one's judgement can be disparaging. That's very obvious to everyone, just not often put into explicit words. This (different) kind of emotion is that which causes conflict within ourselves between the facts of a situation and our feelings about it. It has an impact on how we perceive those facts. In many ways its the antithesis of objectivity, which of course is one of the things we might value in a courtroom judge, or is the basis of the function/formation of an impartial jury in the legal system.

My point is that when many people read the article in question, some of the facts can get very obscured. All of a sudden it's no longer apparent or relevant that the article doesn't say they were necessarily fired, that it didn't happen all at once, that people leave companies all the time and it's not necessarily significant, that Intel is a large corporation where its impossible that people aren't ever released/replaced, et cetera ad infinitum. Instead, what may only be relevant (this is one case) is that Intel is "big brother," they financially dominate the beloved AMD, and they have fired some people as a result of a big corporate scandal. Of course this is made to an extreme, but judgements based on this sort of emotion is what leads to the misinformed opinions we referenced. Its as if to say, "I have love for AMD, hate for Intel, and an ambition to ignorantly defend my own interests." Of course similar examples could be made from opposite viewpoints.

 

[BillR]

Regardless of whether or not AMD people might "enjoy" this more, it belongs in the Intel thread. However, in consideration to the thread itself...
So? Intel will do just fine. I almost think they could have been quietly let go, or they knew that Intel was working on some micro economics, taking a long hard look at some of the development. It happens, but I think they're trying to blow this out of proportion.

And "top" Intel engineers wouldn't have to beg to get into AMD, believe me.

You could indeed put or move this to the Intel forum, there are four whole threads in there right now, they need something to read

 

[ViN86]

im an AMD !!!!!! (ever since my athlon xp) so im gonna stick with em. as was stated before, i do care about intel as they are AMD's motivation.

 

[Epicenter]

Imho it's just dumb luck that AMD can fit dualcore on the same chipset. Because the memory controller and everything related is in the CPU, that factor of compatibility is taken out of the equation.

Uh, the onboard memory controller and very wide bus to the CPU (939 pin socket) are design aspects, not accidents or 'dumb luck'. Do you think the memory controller or socket fell from the sky? No. AMD planned for multicore CPUs from the very start with this design.

The A64 will have 5 sockets by the end of next year. S940, S754, S939, M2, S1. Intel has had what, 3 sockets for the Pentium 4? And LGA775 looks to be around for quite a while, like Socket A.

Also, I'm on a socket 754 setup, and AMD will make no more desktop CPU's for it, aside from value semprons. Way to never leave a socket

If you went with Socket 754 it's your own fault for buying technology you should've known damn well would be obsolete in no time. It was your mistake and you have to deal with it. AMD has no part of the blame in you buying hardware you can't upgrade.

Socket 939 works with almost every Athlon 64 and Athlon 64 FX CPU in existance, and mobile A64s, and will work with dual-core A64s with just a BIOS update. LGA775 does not provide enough bandwidth for effective bandwidth to 2 cores on one die (that is why the Pentium-D's have a slower FSB, 800 MHz and not 1066) and the fact that both cores must communicate through the MUCH higher latency chipset link, compared to AMD's direct core communication, means 775 can't support an effective (see: competitive next to AMD's Athlon 64 X2) dual-core architecture.

Who cares if Pentium-D's can be had for Socket 775? The use of these dual core chips necessitates a CHIPSET change and thereby, a MOTHERBOARD upgrade. So if you are upgrading your board anyway, does it make you feel good that you're moving to another comprimised bottlenecked chipset and socket? The smart choice on Intel's part would be to introduce a dual-core chipset and board solution with closer to ~1000 pins, or more, all at once. But they did not TAKE the smart choice, they took the fast marketing-friendly choice to try and one-up AMD with inferior rushed products and try to beat them with hype.

So, the P4 (Pentium-D's qualify as P4s. They're just 2 attached P4 cores) will have been on 423, 478, 775, and 775 with a new chipset, and will soon have to move to a newer more-pinned socket. That adds up to 4 unique sockets, and what may as well be a fifth socket. AMD's desktop CPUs have had 754, 939, and now M2/S1. That's 4 as well. If you had looked into it a bit more you'd see 940 is meant for SERVERS with OPTERONS, not desktop use, and Intel has its OWN socket for Xeons for use in servers as well. "Whoops."

 

[SuddenDemise]

this is from the inqiuer you people beleive this i dont

I concur, Never belive the shit you hear in the inquier, Next thing they'll print will be; "Intel CEO abducted by Aliens, CEO claims they gave him the plans for the next Pentium...After severly anal probing him".....

 

[Scali]

Uh, the onboard memory controller and very wide bus to the CPU (939 pin socket) are design aspects, not accidents or 'dumb luck'. Do you think the memory controller or socket fell from the sky? No. AMD planned for multicore CPUs from the very start with this design.

I don't see how you come to that conclusion, given the presented facts. The facts indicate at most that multicpu was planned from the start, which is obviously true.
Also, since Intel still uses the 775 socket for dualcore, how come you want to use the socket design as an indication that AMD planned for multicore, while Intel didn't?

LGA775 does not provide enough bandwidth for effective bandwidth to 2 cores on one die (that is why the Pentium-D's have a slower FSB, 800 MHz and not 1066)

Erm, where is the logic in this?
LGA775 supports both 800 and 1066 MHz FSB. The reason why Pentium-D's don't use the 1066 MHz FSB is not because the socket cannot handle it (obviously it can, when using a single-core P4EE), but probably to keep the costs of the dualcore CPUs/chipsets down (same reason why only the EE has a 1066 MHz bus, and none of the other single-cores, so no difference between single or double).

and the fact that both cores must communicate through the MUCH higher latency chipset link, compared to AMD's direct core communication, means 775 can't support an effective (see: competitive next to AMD's Athlon 64 X2) dual-core architecture.

It is still perfectly possible to have the CPUs communicate internally, yet use the LGA775 socket. In fact, it would be easier, since it wouldn't require specific chipset support.

The smart choice on Intel's part would be to introduce a dual-core chipset and board solution with closer to ~1000 pins, or more, all at once.

Really, with all those serial technologies replacing the parallel ones, you should know that the number of pins doesn't have a direct impact on the bandwidth.
Also, you can't compare directly to AMD, since AMD's pins also have to feed a dual-channel memory controller, which in Intel's case is still in the chipset.
Now, compare a single-channel socket to Intel's socket, and you'll see that the number of pins match up quite nicely.

AMD's desktop CPUs have had 754, 939, and now M2/S1. That's 4 as well.

You forgot Socket A, which is still used by Semprons.
Also, you are looking at a much larger timespan for Intel's sockets. 423 was long phased out by the time the 754/939 arrived on the market.

 

[(cf)Eclipse]

It is still perfectly possible to have the CPUs communicate internally, yet use the LGA775 socket. In fact, it would be easier, since it wouldn't require specific chipset support.
.
.
.
You forgot Socket A, which is still used by Semprons.

why aren't they then? intel is already moving to a new chipset to support dual core, might as well do it the right way

also, i'm fairly certain that production on socketA processors ended a few months back

 

[TheOneBarton]

U guys know the AMD is backed up by Intel, just like ECW is backed up by WWE, right?

 

[Scali]

why aren't they then? intel is already moving to a new chipset to support dual core, might as well do it the right way

Because it's a lot cheaper, I suppose. The dualcore is pretty much copy-paste of 2 P4 cores, and the chipset probably got some copy-paste of some of the logic from the dual-CPU Xeon chipsets.
Doing it the right way would require them to build something like AMD has... which Intel will have in its next generation, I suppose. But they really couldn't wait that long before introducing dual-core obviously.

 

[(cf)Eclipse]

yeah, that does make sense. a redesign that large might have taken a few extra months..

 

[Scali]

yeah, that does make sense. a redesign that large might have taken a few extra months..

Yup, and since they wanted to get rid of the dated P4 architecture, that would not have been a good investment... Basically they're doing what you're saying... a redesign, trying to get it right... but it's bigger than just dualcore. It will be an entirely new processor... I can't wait to see it

 

[Epicenter]

I do not need to reiterate AGAIN why LGA775 is not well suited to dual cores. This is a temporary stopgap "copy paste" as was said, technology by Intel to try to one-up AMD until they have a real engineered dual-core solution, and nothing more. We'll see something interesting from Intel on the DC front in 2006 or 2007.

 

[(cf)Eclipse]

but it's bigger than just dualcore. It will be an entirely new processor... I can't wait to see it

same. it's about time intel pushed netburst away for something else.

 

[Scali]

I do not need to reiterate AGAIN why LGA775 is not well suited to dual cores.

Actually you do, because I countered all your previous reasons.

This is a temporary stopgap "copy paste" as was said, technology by Intel to try to one-up AMD until they have a real engineered dual-core solution, and nothing more.

The CPU is, yes, but I don't see how that would relate directly to the socket or chipset.
After all, AMD still only has one memory controller in its dualcore, and only one FSB to the rest of the chipset. So there really isn't a big difference between Intel and AMD there.

 

[Epicenter]

You're so stubborn it's unbelievable. With all the facts right in front of you, you still won't admit that your precious chipmaker made a mistake, and that their dual-core solution is a mess compared to AMD's. That's just moronic from any standpoint. AMD's Athlon X2's are faster, more streamlined in design, run cooler, draw FAR less power, are more overclockable .. and the boards they are on won't require to be REPLACED to add one.

 

[Scali]

You're so stubborn it's unbelievable. With all the facts right in front of you, you still won't admit that your precious chipmaker made a mistake, and that their dual-core solution is a mess compared to AMD's. That's just moronic from any standpoint. AMD's Athlon X2's are faster, more streamlined in design, run cooler, draw FAR less power, are more overclockable .. and the boards they are on won't require to be REPLACED to add one.

Erm, I think you failed to read my posts properly.
I never said the CPUs themselves weren't bad (on the contrary, I agreed with you). I just disagreed with the things you said about the LGA775 socket (like how it doesn't have enough bandwidth because it runs at 800 MHz with the current dualcores, even though it runs at 1066 MHz with certain CPUs. Or how the number of pins has anything to do with it).

So calling me stubborn or moronic is not exactly the right answer to that, is it?

Oh, and for the record, my 'precious chipmaker' at the moment is AMD, since I am using an Athlon.

 

[(cf)Eclipse]

Oh, and for the record, my 'precious chipmaker' at the moment is AMD, since I am using an Athlon.

i thought it was ibm with the PPC

 

[SSE4]

My point is that when many people read the article in question, some of the facts can get very obscured. All of a sudden it's no longer apparent or relevant that the article doesn't say they were necessarily fired, that it didn't happen all at once, that people leave companies all the time and it's not necessarily significant, that Intel is a large corporation where its impossible that people aren't ever released/replaced, et cetera ad infinitum. Instead, what may only be relevant (this is one case) is that Intel is "big brother," they financially dominate the beloved AMD, and they have fired some people as a result of a big corporate scandal. Of course this is made to an extreme, but judgements based on this sort of emotion is what leads to the misinformed opinions we referenced. Its as if to say, "I have love for AMD, hate for Intel, and an ambition to ignorantly defend my own interests." Of course similar examples could be made from opposite viewpoints.

I wholeheartedly agree. I personally don't feel that putting this in the AMD section was the best choice for the thread itself, because we were presenting it to an audience that may very well just ignore all common sense. I agreed with all your points, I just felt that maybe it was in part the lack of personal consideration as opposed to emotionally misguided information. Regardless, I see where I may have branched off from your original intent and I apologize.

 

[BillR]

It’s also interesting to note that insofar as the average consumer is concerned all Intel has to do is produce any new chip, good bad or indifferent and hype it. Dell et al will take it and the public will purchase it just because it has “Intel Inside”.

AMD on the other hand has had to prove itself over and over and is just now starting to make serious inroads into the server market where with luck they can actually make some money.

AMD has yet to “stick” the consumer with a bad or poorly thought out product, such as RDRAM (Ram Buss) and funky sockets and sometimes very questionable chipsets. In fact for the past several years AMD has followed a very common sense line of marketing (despite the confusion about CPU model numbers) all based on longer term viability for their customer base, both professional and consumer.

A better product doesn’t always win the battle but “poor” little AMD is sure heating things up a bit at the moment.

 

[Erasmus354]

Oh, and for the record, my 'precious chipmaker' at the moment is AMD, since I am using an Athlon.

You could have fooled me, you are one of if not the most ardent anti-amd pro-intel people on these boards. At least that is how you come off in nearly every single one of your posts, be that your intentions or not.

 

[Scali]

i thought it was ibm with the PPC

Well, the only IBM I ever owned was a 486DX2-80 (or whatever it was called, the one that was also used by Cyrix). Never actually used it, since it wouldn't work properly on my motherboard.
I took 'chipmaker' literally, as in the company that built my chip.

 

[Scali]

You could have fooled me, you are one of if not the most ardent anti-amd pro-intel people on these boards. At least that is how you come off in nearly every single one of your posts, be that your intentions or not.

Well first of all, that seems to just be your interpretation of my posts, I don't think everyone would agree with that, and they certainly weren't meant to be anti-AMD or pro-Intel. I just try to stick to technical facts.
Secondly, is it just my interpretation or do you get annoyed by what you think are anti-AMD or pro-Intel posts? You sure have reacted incredibly rudely in my direction for absolutely no reason, other than perhaps this one.

 

[(cf)Eclipse]

I took 'chipmaker' literally, as in the company that built my chip.

hehe, proudly made is dresden, germany!

though, no offense intended, Erasmus is not the first one to think that you are very biased, even if you say you aren't. that's all past history though

 

[RawsonDR]

I just felt that maybe it was in part the lack of personal consideration as opposed to emotionally misguided information.

Agreed. In fact, it can certainly be a combination of both.

I do not need to reiterate AGAIN why LGA775 is not well suited to dual cores.

Actually you do, because I countered all your previous reasons.

LOL

You're so stubborn it's unbelievable. With all the facts right in front of you, you still won't admit that your precious chipmaker made a mistake, and that their dual-core solution is a mess compared to AMD's. That's just moronic from any standpoint. AMD's Athlon X2's are faster, more streamlined in design, run cooler, draw FAR less power, are more overclockable .. and the boards they are on won't require to be REPLACED to add one.

LOL again. Did the close to that particular discussion remind anyone else of tricky telemarketers/salesmen? They start out with conversation to get you talking, and then less-than-subtly move into their sales pitch. At this point you realize they were never really listening to you at all and hang up.

 

[Scali]

hehe, proudly made is dresden, germany!

though, no offense intended, Erasmus is not the first one to think that you are very biased, even if you say you aren't. that's all past history though

Not the first no... but afaik all people who said it, had a rather extreme slant towards AMD.
As I say, I'm neutral. The problem is that I may be the only neutral person on these boards, because most people here only know x86, and either like Intel or AMD (usually because that is what they bought). As I said before, I don't like x86 at all. But at least I can see the good and bad points of both brands. It's just hard for these biased people to realize that their 'opponent' can have good points aswell, so when I bring those up, I am considered biased towards the other party.
Also, as a developer and experienced assembly programmer, I probably have a much clearer view of what a CPU can actually do, since I don't have to rely on other people's code and compilers to benchmark the CPUs. The same goes for videocards, which became horribly painful when the GeForce FX arrived. People like me would realize immediately that it was a dud, but most people had such strong faith in NVIDIA, that they had to find the cause of the bad performance somewhere else.

 

[Scali]

LOL again. Did the close to that particular discussion remind anyone else of tricky telemarketers/salesmen? They start out with conversation to get you talking, and then less-than-subtly move into their sales pitch. At this point you realize they were never really listening to you at all and hang up.

Yea, it certainly felt that way to me. I have repeatedly said that AMD's dualcores were better, but still he claimed I said they were bad, and avoided the argument about the sockets completely, and instead started going for the ad hominems.