Finally, netburst is dead!

[TheRod]

<A HREF="http://babelfish.altavista.com/babelfish/trurl_pagecontent?lp=de_en&trurl=http://www.golem.de/0505/37996-3.html" target="_new">http://babelfish.altavista.com/babelfish/trurl_pagecontent?lp=de_en&trurl=http://www.golem.de/0505/37996-3.html</A>

We all wanted dead, it will be soon!

-
GA-K8NF-9 / <b><font color=green>Athlon 64 3200+</font color=green> @ 3800+</b>
Infineon DDR400 (CL2.5) 2x512Megs
<font color=green>GeForce 6600GT 128Megs</font color=green>
<A HREF="http://www.getfirefox.com" target="_new">Firefox</A>

 

[Ned_Flanders]

Gelsinger: I build nevertheless no Fabs! (laughs)

We do not throw Netburst away.

______________
Welcome to my Shed of Pleasure

 

[Mephistopheles]

It was already dead, they'll just bury it.

BTW, they've stated that the next architecture will inherit a lot of netburst peculiarities - and as far as I know, the good ones. For instance, the next architecture will feature HT too.

It has also been stated officially that many features from dothan/centrino will also be present...

 

[TheRod]

Intel is putting Dothan and NetBurst in the mixer... To get a better core! I think they want to compete in the big league with AMD. It's good to know that Intel is taking exprience... We might see a real competition to AMD in a couple of years!

-
GA-K8NF-9 / <b><font color=green>Athlon 64 3200+</font color=green> @ 3800+</b>
Infineon DDR400 (CL2.5) 2x512Megs
<font color=green>GeForce 6600GT 128Megs</font color=green>
<A HREF="http://www.getfirefox.com" target="_new">Firefox</A>

 

[Mephistopheles]

It'll probably take them a full 18 months from now (2H06) to truly make up for the very bad state they're in right now. By then, they'll have been nearly defenseless for a looong time.

I'm a bit disappointed at AMD, however, for not playing a more agressive game. I mean, the top of the line chips are still 130nm! What's up with that? Why don't they launch a 4200+, 2.6Ghz Venice all the way? Where's the 2.8Ghz FX-57? As of yet, venice <i>did not enable any new speed grades!</i>

Launching an even higher-rated CPU would probably even make our friend Joe Sixpack interested. In my opinion, AMD should be playing more agressively...

 

[slvr_phoenix]

It's actually kind of a shame that Intel is killing Netburst. I mean the concept itself had some potential. It was just the way that Intel kept making things worse and worse instead of better that was the real problem.

That aside, the higher IPC/lower clock is better for power consumption and does allow for lower cache latency to be used. So there's a lot of good potential available should Intel do things right. It's just a shame that based on the Pentium 4's history we can't really trust Intel to do it right.

I have to agree though that it's kind of a shame that AMD isn't pressing their leverage right now. It's probably for some stupid reason like they couldn't ship the quantities necessary yet if they did.

Though I have the sinking feeling that were AMD to push hard on Intel, it'd wake a sleeping beast that has enormous resources at hand and could well suffer a temporary loss to push enough of an advance to totally dominate once more. So AMD may just be letting the sleeping dog lie and be content with nibbling away marketshare bit by bit without causing a major upset. In the long run it may even be the right business plan to use.

<pre><font color=purple><i>Jesters do oft prove prophets.</i> -Regan in
King Lear (Act V, Scene iii) by William Shakespear</font color=purple></pre><p>@ 187K -> 200,000 miles or bust!

 

[TheRod]

So AMD may just be letting the sleeping dog lie and be content with nibbling away marketshare bit by bit without causing a major upset. In the long run it may even be the right business plan to use.

I agree, AMD can't just throw 200$US 2.8GHz dual-core CPU in the market and expect Intel to not do anything... AMD is not financially STRONG. They play safe, they launch products that are (usually) faster/better than Intel in most segments at competive price to make sure they will make money every time they will sell a CPU.

Since AMD can't beat Intel in volumes and there is a ton of people who stil thinks that AMD sucks, AMD would not benefit much from selling 2.8GHz or 3.0GHz chips at lower costs right now. Their sales would not climb that much and AMD would probably get their benefits too much.

As you said, a few percents each year is better than 100% in 1 quarter that would vanish in the long run.

With emerging markets, like China and India. I think AMD have an opportunity to steal markets from Intel there. I would be surprised to hear agreements news about Lenovo and AMD in a near future.

-
GA-K8NF-9 / <b><font color=green>Athlon 64 3200+</font color=green> @ 3800+</b>
Infineon DDR400 (CL2.5) 2x512Megs
<font color=green>GeForce 6600GT 128Megs</font color=green>
<A HREF="http://www.getfirefox.com" target="_new">Get Firefox!</A>

 

[Spitfire_x86]

With emerging markets, like China and India. I think AMD have an opportunity to steal markets from Intel there. I would be surprised to hear agreements news about Lenovo and AMD in a near future.

I'm not very optimistic about AMD's future in Indian subcontinent. In Bangladesh, AMD is very disliked by most PC sellers. These clueless idiots build AMD system with cheap stuff and put the blame on AMD for all of the consequent problems. People know that computer means Pentium "X", P4 > P3 > P2, higher MHz = better. And if someone is curious about AMD then these PC sellers try their best to convience them with all their lies that, AMD = bad & headaches. I guess if I owned a PC shop, then my AMD sales : Intel sales ratio would be somewhat like 1:10, even if I tried my best to push AMD.

Overpricing/lower availability of AMD mobos, (particularly nForce mobos) is also a big problem. Can you believe that an MSI nForce2 IGP mobo (without MCP-T) still costs $125?? Other AMD mobos are not that badly overpriced, but their price is far away from justified.

I don't know much about the Indian situation, but I can safely assume that it's not much different.

------------
<font color=orange><b><A HREF="http://www.mozilla.org/products/firefox" target="_new">Rediscover the web</A></b></font color=orange>

 

[TheRod]

Humm... Good points! The situation you describe reminds me of the old AMD days (486-clone and K6), PC shop were usually putting together a lot of cheap components around the AMD Processor. This situation really hurts the AMD image at the time.

For my part, I push AMD a lot in my entourage except for Laptops.

-
GA-K8NF-9 / <b><font color=green>Athlon 64 3200+</font color=green> @ 3800+</b>
Infineon DDR400 (CL2.5) 2x512Megs
<font color=green>GeForce 6600GT 128Megs</font color=green>
<A HREF="http://www.getfirefox.com" target="_new">Get Firefox!</A>

 

[endyen]

It's a pity you know. Oh, it's fine for amd fanboys (like me) because it keeps intel in bad shape for a while longer.
Truth is that the P4c was the netburst chip that started to make sense. The mem bandwidth was finally catching up to the chip requirements.
Yea, scotty was a screwup. If they had put northwood onto 90 nanos, they could have gained more speed, the way they should, with a 233 mhz fsb. Maybe even 266, with FDsoi. That would have them well past the 4 ghz, closing on 5. With 65 nanos, DDR2 would start to kick in big time on a 300mhz fsb. That is where Amd would have some serious problems keeping up.
Oh well, Intel's basic conservative approach has held tech back again.

 

[P4Man]

>I mean, the top of the line chips are still 130nm! What's up
>with that? Why don't they launch a 4200+, 2.6Ghz Venice all
>the way? Where's the 2.8Ghz FX-57? As of yet, venice did not
>enable any new speed grades!

If memory serves, AMD has said like >6 months ago that initial 90nm products would be not be geared towards higher clocks than 130nm ones. It would need another metal layer and stepping to unleash the clock potential of 90nm (tbred A anyone ?). At least I have always expected first 90nm to be mediocre clockers for that reason, if anything I'm pleasantly suprised by both the low power, and quite reasonable overclocking headroom. It will be interesting to see how much lowpower AMD has to sacrifice for the higher clocks.

>BTW, they've stated that the next architecture will inherit
>a lot of netburst peculiarities - and as far as I know, the >good ones.

yes, with all the low hanging fruit gone, all cpu mfgs are picking and mixing the best idea's that they have not implemented yet themselves, and these get increasingly rare. Won't be long until intel goes ODMC and P2P, and neither will it take AMD more than a few years to implement a trace cache and µops fusion. Dunno, but less than 5 years from now, I don't think there will be a big design difference anymore between any high end, general purpose CPU. They will probably all look like "K8 evolutions" with trace cache and µops fusion and possibly vastly improved FP units

>For instance, the next architecture will feature HT too.

Did intel really confirm HT in Merom ? I never heard that, but would be quite interested to learn. Got a link ? I like the idea behind HT a lot, but I doubt it makes a lot of sense in a multicore future.

>Launching an even higher-rated CPU would probably even make
>our friend Joe Sixpack interested. In my opinion, AMD should
> be playing more agressively...

Not many Joe Six packs are interested in a $1000 4800+, let alone anything even faster and more expensive.

= The views stated herein are my personal views, and not necessarily the views of my wife. =

 

[silverpig]

">Launching an even higher-rated CPU would probably even make
>our friend Joe Sixpack interested. In my opinion, AMD should
> be playing more agressively...

Not many Joe Six packs are interested in a $1000 4800+, let alone anything even faster and more expensive."

That's it right there. Joe Sixpack can't afford a 4800. The people who can are into workstations, and those benefit more from dual cores at lower frequencies and multi cpu systems. Even "hardcore" users will be hard pressed to need anything past a 3500 right now. It's really just bragging rights to get 152 fps as opposed to 143.

s signature has been formatted to fit your scr

 

[apache_lives]

from what i heard HT only works properly on Netburst cpu's because of the long pipelines, and second - didnt intel say HT was an evolutionary step between uni and multi-threaded programming to help kick the market in the "right" (intels future) direction?

im wondering if intel will follow suit with AMD and integrate the memory controller into the cpu package - was that AMD's original idea or was it IBM's idea or something?

 

[Ned_Flanders]

Because AMD intergrated their memory controller into the die itself, how come the price of AMD chips aren't significantly higher? Are they running at a loss?

I beleive it is all to do with die space - you can put what the hell you want in your silicon, the only cost factor is the amount of silicon you use. AMD and Intel have relitavly similar die areas.



______________
Welcome to my Shed of Pleasure

 

[P4Man]

>from what i heard HT only works properly on Netburst cpu's
>because of the long pipelines

HT is just a way to extract more threal level parallellism that most likely gives a slightly better potential benefit with a deeply pipelined CPU because of the higher cache miss penalty, but in theory,it could benefit any cpu, regardless of pipe length. Its a neat idea, that doesnt cost a lot of transistors, and could benefit a K8 or Dothan like architecture as well.

>m wondering if intel will follow suit with AMD and integrate
> the memory controller into the cpu package - was that AMD's
> original idea or was it IBM's idea or something?

Such ideas are never truly original. AMD didnt "think" of it first, neither did Intel think of hyperthreading first. usually, such concept are 'invented' decades earlier in academic circles, and most often implemented in silicion first in specialty chips or risc. HT afaik, intel was first with an actual commercial implementation. The concpet of ondie memory controllers has been around for a very long time (Timna, Alpha EV8 and probably half a dozen other riscs). As always, the major issue is not the idea, but a succesfull implementation.


= The views stated herein are my personal views, and not necessarily the views of my wife. =

 

[apache_lives]

The market leader always defines prices does it not? Amd's would cost more but intel defines the overall prices.