New about Hammer

[castle]

I wouldnot say copper is some tech that advanced compared with 0.13. It is true that AMD adapted it earlier than Intel, but at what cost? Copper is extremely dangerous in this industry and great caution is needed. Once production line in Dresdon was shutdown for copper comtamination. For something like this I donot want to be the 1st to try. Same arguement for SOI.
If you want to compare which company is faster in new process convertion, please note: Intel finished copper and .13 together and started mass production in summer 2001 in multiple fabs; AMD did copper in 2000, and not yet completely finish its 0.13 in one fab.

 

[zengeos]

Oh brother.

Each company is ahead in some ways and behind in others.

Why is that?

Because each company has it's own priorities to meet it's particular production roadmaps.

INTC has decided that A, B and C need to be done before D.

AMD has decided that A, C and D need to be done and THEN move on to B.

Saying AMD is behind or saying Intel is behind.. or saying either is ahead is totally irrelevant.

Yes, Intel is a larger company (4-5x bigger) with commensurately more engineers, designers and personnel in general. This certainly gives Intel some advanatages. However, AMD, for a much smaller company has managed to maintain general development and production process parity with Intel; enough to remain competetive and maintain marketshare, at least.

Finally, while Intel has far more engineers to develop their processes and chip designs, AMD has managed to garner more patents by far than Intel.

Mark-

<font color=blue>When all else fails, throw your computer out the window!!!</font color=blue>

 

[zengeos]

Oh brother.

Each company is ahead in some ways and behind in others.

Why is that?

Because each company has it's own priorities to meet it's particular production roadmaps.

INTC has decided that A, B and C need to be done before D.

AMD has decided that A, C and D need to be done and THEN move on to B.

Saying AMD is behind or saying Intel is behind.. or saying either is ahead is totally irrelevant.

Yes, Intel is a larger company (4-5x bigger) with commensurately more engineers, designers and personnel in general. This certainly gives Intel some advanatages. However, AMD, for a much smaller company has managed to maintain general development and production process parity with Intel; enough to remain competetive and maintain marketshare, at least.

Finally, while Intel has far more engineers to develop their processes and chip designs, AMD has managed to garner more patents by far than Intel.

Mark-

<font color=blue>When all else fails, throw your computer out the window!!!</font color=blue>

 

[chuck232]

What I say is that i really doesn't matter who comes out with what first. As long as it benifits us in the long run I'm happy. Also it shouldn't matter who is more advanced right now. For example AMD may be ahead of Intel, but it seems like Intel still has the consumer base, which AMD does not.

My firewall tastes like burning.

 

[Matisaro]

wouldnot say copper is some tech that advanced compared with 0.13. It is true that AMD adapted it earlier than Intel, but at what cost? Copper is extremely dangerous in this industry and great caution is needed. Once production line in Dresdon was shutdown for copper comtamination. For something like this I donot want to be the 1st to try. Same arguement for SOI.
If you want to compare which company is faster in new process convertion, please note: Intel finished copper and .13 together and started mass production in summer 2001 in multiple fabs; AMD did copper in 2000, and not yet completely finish its 0.13 in one fab.

Copper is more of a tech advancement than .13 given the reasons you mentioned aboved(also dresden has never shut down due to copper contamination to my knowledge, please give me links to where you read this).


Furthermore, amds transistion to .13 seems to be on track and very closely behind intels as evidenced by the many hammer samples already on the market, its clear that amd can make .13 chips, and there has been information given to us by texas which explains why the tbred is late, the first run of the core had issues which have been cleared up, amds .13 process itself is sound IMO.

The Cash Left In My Pocket,The BEST Benchmark

 

[FatBurger]

<i>Chuck232 says:</i>
Remember when Intel said that 3GHz P4 wouldn't be out till at least Q103? Now they're saying Q402... looks like they're kinda worried.

That's strange, because I remember hearing from more than one source around January that 3GHz would be here by Christmas. Interesting how selective memory works :tongue:

<i>Chuck232 says:</i>
That really gives some hope that 1Mb L2 cache is on the way.

Or perhaps an L3 cache? They'd definitely have the room, though they might opt for more dies per wafer instead.

<font color=blue>Hi mom!</font color=blue>

 

[chuck232]

That's strange, because I remember hearing from more than one source around January that 3GHz would be here by Christmas. Interesting how selective memory works

Well I don't know how reliable inquirer.net is but <A HREF="http://www.the-inquirer.com/10040205.htm" target="_new">here</A> you go. Yeah it is pretty interesting how selective memory works.... :tongue:

My firewall tastes like burning.

 

[FatBurger]

That's ok, US News had 3GHz around Christmas, I'd call them a bit more reliable than the Inquirer (but less of a tech publication, obviously).

<font color=blue>Hi mom!</font color=blue>

 

[bront]

Actualy, AMD's transition to .13 is done. The Mobile Athlon XPs now feature the T-bread core, and have since early this month.

We're simply waiting for the desktops to be released.

"Search your feelings you know it to be true, I am your... twin sister" - Darth Vader

 

[bront]

Or perhaps an L3 cache? They'd definitely have the room, though they might opt for more dies per wafer instead.

With the memory controler on the CPU Die, I would think the benifits of an L3 cache, or even a larger L2 cache would be lessened. L3 more than L2.

"Search your feelings you know it to be true, I am your... twin sister" - Darth Vader

 

[Kemche]

With the memory controler on the CPU Die, I would think the benifits of an L3 cache, or even a larger L2 cache would be lessened. L3 more than L2.

So are you trying to say if Hammer would to include less L2 cache it wouldn't hurt the Preformance since it has an integrated memory controller. I think hammer is going to have 512kb of L2, can you tell me if it would to have say 256kb of cache how much would it effect if any?

KG

"640K ought to be enough for anybody." - Bill Gates.

 

[Kemche]

Furthermore, amds transistion to .13 seems to be on track and very closely behind intels as evidenced by the many hammer samples already on the market, its clear that amd can make .13 chips, and there has been information given to us by texas which explains why the tbred is late, the first run of the core had issues which have been cleared up, amds .13 process itself is sound IMO.

Hey, if you want to talk about samples then let's please count the samples of P3 Tually were also available long before they released the processor. So are you saying that their .13u is done but they just aren't releasing since they don't have any product that they want to deploy it in? I don't think that's the case.

Oh and the most likely reason Intel probably didn't use copper is because to save cost since copper is littele expensive then Aluminium.

About T-bred core having problems and not the .13u having problem is just totally wrong. Just think about it T-bred is just a die shrink meaning nothing new from Palomino. Only thing new is the process technology, so the only thing that can go wrong making the processor is the process technology which is .13u.

If you still don't believe that AMD isn't haveing problem with their .13u process technology then how do you explain the low yields and clock frequency of the T-bred released for Mobile. Also the T-bred that will be launched on June 10 will only be available in Limited Quentities, so can you please explain why would AMD release something in Limited Quentities if their .13u was done and was having an excellent yields like you think?

KG

"640K ought to be enough for anybody." - Bill Gates.

 

[bront]

I was guessing/theorizing/asking (Hence the words "I would think"). Also, I said that the effects would be lessened, not that they would have no effect whatsoever.

"Search your feelings you know it to be true, I am your... twin sister" - Darth Vader

 

[Matisaro]

Hey, if you want to talk about samples then let's please count the samples of P3 Tually were also available long before they released the processor. So are you saying that their .13u is done but they just aren't releasing since they don't have any product that they want to deploy it in? I don't think that's the case.

You mean the tually samples from intels small RESEARCH FABS, research fabs amd does not have, the .13 samples of hammer come from DRESDEN, amds ONLY production fab.

You are trying to compare apples and oranges.

Oh and the most likely reason Intel probably didn't use copper is because to save cost since copper is littele expensive then Aluminium.

Pulled that one out of nowhere. Amd apparently thinks copper is good, and copper adds quite a bit of scalability to a process. And its not much more expensive than aluminum, or do you think the fact a copper hsf costs more than an aluminum one means amd is paying tons more for copper than it would for alum....lol.

About T-bred core having problems and not the .13u having problem is just totally wrong. Just think about it T-bred is just a die shrink meaning nothing new from Palomino. Only thing new is the process technology, so the only thing that can go wrong making the processor is the process technology which is .13u.

This shows you dont know about the industry, the tbred could have a timing issue caused by its design, a shrink requires an entire new layout for timing issues etc.(everything happens so fast on a microchip that the difference an electrical signal takes to travel between a .2mm line and a .3mm line can cause the whole processor not to work properly.) The fact amd has produced working .13 chips in its MAIN AND ONLY PRODUCTION FAB, is a signal that they are not too far behind intel, and since hammers schedule is so advanced leads me to believe that the .9 transistion would be the same.

If you still don't believe that AMD isn't haveing problem with their .13u process technology then how do you explain the low yields and clock frequency of the T-bred released for Mobile. Also the T-bred that will be launched on June 10 will only be available in Limited Quentities, so can you please explain why would AMD release something in Limited Quentities if their .13u was done and was having an excellent yields like you think?

Your confusing tbred having yield problems with amds .13 process, for the last time they are different things!@!

I explain the low yielding of mobiles(which btw you have no evidence to back up) with the fact that the first run of tbired chips are to be sold as mobiles, and those chips had some timing issues which limit their clockability, the new tbreds coming out(the ones in june) are fixxed according to texas techie's sources.

The Cash Left In My Pocket,The BEST Benchmark

 

[FatBurger]

<i>bront says:</i>
With the memory controler on the CPU Die, I would think the benifits of an L3 cache, or even a larger L2 cache would be lessened. L3 more than L2.

I agree, but we were talking about Prescott, not Hammer.

<i>Matisaro says:</i>
Pulled that one out of nowhere.

...and then you agreed with him.

<font color=blue>Hi mom!</font color=blue>

 

[Matisaro]

In reply to:
--------------------------------------------------------------------------------

Matisaro says:
Pulled that one out of nowhere.



--------------------------------------------------------------------------------


...and then you agreed with him.

Pulled that one out of nowhere. Amd apparently thinks copper is good, and copper adds quite a bit of scalability to a process. And its not much more expensive than aluminum, or do you think the fact a copper hsf costs more than an aluminum one means amd is paying tons more for copper than it would for alum....lol.

Obviously the lol at the end was missed, copper is not more expensive than aluminum in any signifigant capacity, his theory was bullshit and I thought the lol at the end(where I mocked his idea based on the fact he probably went by the price of copper over alum as a basis, and laughed at his flawwed reasoning)

But apparently I have to be really specific lately in my intent heh, I will try from now on, instead of hinting the person is full of [-peep-] to straight out and tell them that their idea about cost being a reason for non adoption of copper is bullshit and they are a moron for even suggesting it.

If that makes you happy oh burgermeister.

;-)

The Cash Left In My Pocket,The BEST Benchmark

 

[thelastguardian]

but i think all of you will agree that copper is a very toxic substace and need to be more carefully to handling around and dispoise of.

 

[thelastguardian]

thus making it a little bit more expensive, and copper is much softer than alu. so harder to cut (i guess)

 

[FatBurger]

And its not much more expensive than aluminum

So it's a "little more expensive", like he said.

<font color=blue>Hi mom!</font color=blue>

 

[Kelledin]

all of you will agree that copper is a very toxic substace

Far less toxic than lead, which is a principal component of solder.

How toxic is aluminum, btw? I honestly don't know.

copper is much softer than alu. so harder to cut (i guess)

Hmmm...copper is softer...so it's harder to cut? You might want to rethink that a bit.

Aluminum burns more easily than copper (anyone who welds tubing for refrigeration will tell you this), and even light friction overheating can cause aluminum to melt and "smear" (any mechanics seen what aluminum rods do to a crankpin when run without oil? ). This makes cutting aluminum a rather more delicate task.

Also, aluminum's greater tendency to burn makes it rather more toxic to work with (fumes get produced primarily when something burns or evaporates, not when it melts).

<pre>We now <b>return</b>(<font color=blue>-1</font color=blue>) to an irregular program scheduler.</pre><p><P ID="edit"><FONT SIZE=-1><EM>Edited by kelledin on 05/20/02 06:15 PM.</EM></FONT></P>

 

[FatBurger]

I think he meant that since it's softer, a blade wouldn't be able to grip against it as easily. Kind of like cutting through slightly melted butter instead of firm butter. Harder to get a clean cut.

I don't know if that would hold true in a FAB though.

<font color=blue>Hi mom!</font color=blue>

 

[Kemche]

Sorry, I confisued everyone on this copper issue, All I ment to say was copper is probabaly little more expensive like pennies more expensive then "Aluminum". And since Intel sells more CPU each pennies saved would help big time on their bottom line.

For example AMD has 20% market share and they sell 8 Million CPU's per Quarter. Therefore Intel probably sell 40 Million. If they save just a Nickel on each they would save 2 Million per Quarter.

I am not saying that's the reason they didn't adopt copper, I was just saying copper is little expensive.

KG

"640K ought to be enough for anybody." - Bill Gates.

 

[thelastguardian]

Fatburger comes to the rescue wahhhhhhhhh
taht was my guess anyway but copper's price ahd been all time low from all i no, and alu is not doing much better either.

I know what you did last summer when you install amd, and now its time for punishment!! -Intel

 

[zengeos]

For example AMD has 20% market share and they sell 8 Million CPU's per Quarter. Therefore Intel probably sell 40 Million. If they save just a Nickel on each they would save 2 Million per Quarter.

Actually, Intel produces between 30 and 32 million CPU's per quarter by your calculations of AMD's 8 million being 20%

(32 million is 4x 8 million. 1 + 4 = 5, which, multiplied by 20 = 100)

Mark-

<font color=blue>When all else fails, throw your computer out the window!!!</font color=blue>

 

[Matisaro]

So it's a "little more expensive", like he said

A: I havent conceded the fact it is even more expensive, I have no hard data.
B: he was infering it was soo more expensive intel refused to touch it by choice, and furthermore he infered that intel "let" amd get cu ic first.

Which was the reason I countered his post in the way I did.

The Cash Left In My Pocket,The BEST Benchmark