AMD Down 4 the Count? Future Map I vs A

[duby229]

Can you even imagine a dual core ULV AMD processor based on 65nm! Sub 30 watts perhaps! Passive cooling anyone...



AMD really needs new FABS and fast. They're currently selling every chip they make, including the ones out of Chartered. This Wednesday will be an interesting earnings call.

And with Fab36 about to come online, it will be interesting to see how the transition to 65nm, and the extra capacity it brings with it, will increase market share....

I'm thinkin about 30% global market share by the end of 07. Just a guess based on capacity, and demand.

 

[enumae]

And with Fab36 about to come online, it will be interesting to see how the transition to 65nm, and the extra capacity it brings with it, will increase market share....

Sorry, but how do you see Fab 36 helping AMD gain market share?

AMD is capacity constrained now, they will need Fab 36 to transition to 65nm and fast.

Also Fab 30 conversion to 65nm doesn't happen until Fab 36 is fully converted, which is scheduled for the middle of 07.

AMD will have a very hard time regaining/holding onto market share next year.

Remember that all of Intels 90nm low end chips will be converted to 65nm Core based processors by Q207, and all of the Core, Core Duo and Core 2 Duo based processors are beating 90nm K8 in there equivalent segments.

Then factor in all new highend parts from Intel will be 45nm exiting 2007.

Just to be clear... all comments made above are for Desktop and Mobile segments.

I'm thinkin about 30% global market share by the end of 07. Just a guess based on capacity, and demand.

That is optimistic IMHO, they will probably get to about 30-35% server due to architecture superiority in the server space (4P and above), desktop and mobile will be very hard areas in which to gain market share, relating to comments above.

 

[duby229]

Like I said it's just a guess. I may end up being very wrong.

Look, Fab36 will have more capacity then Fab30. So even with Fab30 going partailly down for the transition, they will have more capacity next year then they do now. By the end of next year Fab 36 will be in full production, Fab30 will be partialy transitioned, with at least some output, and Chartered will have at least some output. Any way you look at it, by this time next year AMD will have considerably more capacity then it does now.

The average consumer doesnt know two licks about what procesor performs better, they only see that this computer has wireless and a dvd burner, and it costs less then that one over there..........

 

[enumae]

The average consumer doesnt know two licks about what procesor performs better, they only see that this computer has wireless and a dvd burner, and it costs less then that one over there..........

I Completely agree the average user has no clue, price is the main factor.

 

[sethk]

What I'm wondering is how much of an improvement in IPC there will be on the move to the newer AMD cores.

AMD seems confident enough in keeping the clocks down, and I've read a couple of articles about the core formerly known as K8L (lots of tweaks and improvements, but still a 3 issue core.) At least the 4 core chips will be a real 4 core and not two dual cores glued together, which might give it a boost when combined with the HT upgrades. Even the dual core next gen AMD chips should be interesting with their 2MB L3 cache - hopefully they are clock for clock competitive with Intel. While this roadmap is interesting, it doesn't really tell us where we'll end up performance-wise, and that's assuming that the map is even semi-accurate, which given Intel's past roadmap history is probably too much to assume. How long after they promised a 4Ghz Netburst did they finally admit that the mhz war was over?

I can't believe that AMD will just hand over the performance crown for close to two years to Intel - it would lead to a serious erosion of the enthusiast base that led them to where they are now. If it came to it, I think they'd go back to competing on price like they did before the hammer architecture debuted. I see AMD continuing to provide either a compelling value or performance argument for their chips, even if they cant offer both.

 

[sethk]

Oh yeah, and what is the deal with the Windsor 5400+ and the Brisbane 5000+ (other than the process, the cache and clocks are the same according to that roadmap - is that a mistake?) Will there be a drop in IPC with the process shrink? I would hope it would be the other way around.

 

[harpoon]

I can't believe that AMD will just hand over the performance crown for close to two years to Intel - it would lead to a serious erosion of the enthusiast base that led them to where they are now. If it came to it, I think they'd go back to competing on price like they did before the hammer architecture debuted. I see AMD continuing to provide either a compelling value or performance argument for their chips, even if they cant offer both.

What percentage of AMD's marketshare is made up by enthusiasts? I suspect it's a lot higher than Intel's, mainly because in the past only 'tech geeks' know of AMD, though I'm sure that is slowly changing with Dell now selling AMD rigs.

If AMD can't compete on performance they are in trouble. Asking them to lower prices to compete is asking them to bleed money. Not good when AMD has just spent $5+ billion acquiring ATi.

 

[(cf)Eclipse]

The problem is nobody really makes a LV CPU... not even intel. They basically get lucky with them. They'll bin the parts out to see which ones are capable of it... and market those as such. But for them to ONLY offer those LV and ULV parts it would cost them a fortune in wasted parts. Could they have a sku of LV quad.... sure, but they would also have the 150+ Watt ones. Thats not even mentioning the supply issues that AMD already have..... so those parts would be EXTREAMLY difficult to obtain on the market.

you might be surprised at how much can be done to make a chip hotter/colder before production. there is such a thing called oxide gates, and by varying the tickness, you vary the current leakage, and voltage requirements...

 

[duby229]

you might be surprised at how much can be done to make a chip hotter/colder before production. there is such a thing called oxide gates, and by varying the tickness, you vary the current leakage, and voltage requirements...

It's real hard to get right on mass production tho... That is the whole reason why IBM is working on dual gates using carbon gate pads....

I dont know if the gates will use some sort of metal oxide or if they will be silicon, but either way they claim that it will work well if they use carbon gate pads.

 

[SLee]

Can you even imagine a dual core ULV AMD processor based on 65nm! Sub 30 watts perhaps! Passive cooling anyone...

What's so special about 30W, Merom does it at 1.25v. Conroe could probably do it at LV at 2.4GHz.

 

[Poncho]

What's so special about 30W, Merom does it at 1.25v. Conroe could probably do it at LV at 2.4GHz.

Don't forget the 45 WATT Xeons running @ 2.33/1333.... those are some BADASS processors. I'm VERY close to building a passively cooled workstation based on them.

As for AMDs "increased" capacity with FAB 36.... you have to realize that with their transition to 65nm their supplies with be reduced. I don't see this FAB as their saving grace in capacity. I would be AMAZED if they hit 35% market share in desktop and mobile. They may hit that if you factor in MP servers.... but Intel is becoming competitive in that make again next year and are already competitive in the DP market.... don't forget the blade market that will include 50WATT quad core next month. Yea... 35% not gonna happen.

 

[duby229]

I'm saying possibly 30% I never mentioned anything about 35%.....

Anyhow, you may be right even about 30%.... I really dont know.

 

[Poncho]

I'm saying possibly 30% I never mentioned anything about 35%.....

Anyhow, you may be right even about 30%.... I really dont know.

You're right... for some reason I thought I saw 35%. My bad... My opinion still holds @ 30% though. It's possible... but VERY unlikely.

 

[Donnie27]

You're right... for some reason I thought I saw 35%. My bad... My opinion still holds @ 30% though. It's possible... but VERY unlikely.

If AMD has 20% they'll be happy LOL!

I see it similar to these guys.

Hexus.net

The Barcelona will also allegedly include a HyperTransport 2 interface, which will give it a distinct advantage in multi-socket arrangements – like 4x4. Looking towards the end of next year, AMD also has HyperTransport 3 on the roadmap, and a new connection called Torrenza. This will allow peripherals (ATI graphics cards perchance?) to be connected directly to the processor via a bus with much more bandwidth than even PCI Express 2.

So although Intel is on top for now, AMD still has a few cards yet to play. These may not amount to the same kind of winning hand the company had for the last couple of years. But to say AMD is done for, and has failed to capitalise on the Athlon 64’s success, is simply short-sighted. The market moves in phases, and there are bigger gameplans than just having the fastest gaming chip around.

Rocky Relationship with AMD

It's been a challenging few months for AMD. It launched its AM2 socket for desktops and workstations in May, but the move to DDR-2 hasn't come with any great performance leaps. That in itself wouldn't be too much of a problem, but Intel delivered on its promise of opening a can of whoopass with its new mArch, letting Core 2 desktop chips loose just a month after the release of AM2.

Hehehe!

 

[(cf)Eclipse]

i'll be sure to rub in your face just as much when K8L comes around...

 

[Poncho]

i'll be sure to rub in your face just as much when K8L comes around...

At which point Intel will have upped their performance as well. Towards the end of next year you should start seeing 45nm as well as increased FSB with Bearlake chipset and let's not forget built in DX10 capapblilty with bearlake. Unless K8L increases performance 30-40% over current AMD cpus they've got a lot of catching up to do. And that's just in the desktop market. In the MP market you'll see a VERY competitive platform using NGMA and quad FSB around the same time as well. About a year after that you should see a HUGE can of whoopass in the DP server market. It's all about catch up with AMD at this point. Maybe they'll do it... but they are at a bit of a disadvantage with their problem ridden transition to 65nm and limited capacity. I bet by next year you'll see desktop/mobile marketshare decline back down to 15% or so.

 

[duby229]

At which point Intel will have upped their performance as well. Towards the end of next year you should start seeing 45nm as well as increased FSB with Bearlake chipset and let's not forget built in DX10 capapblilty with bearlake. Unless K8L increases performance 30-40% over current AMD cpus they've got a lot of catching up to do. And that's just in the desktop market. In the MP market you'll see a VERY competitive platform using NGMA and quad FSB around the same time as well. About a year after that you should see a HUGE can of whoopass in the DP server market. It's all about catch up with AMD at this point. Maybe they'll do it... but they are at a bit of a disadvantage with their problem ridden transition to 65nm and limited capacity. I bet by next year you'll see desktop/mobile marketshare decline back down to 15% or so.

Quad GTL+ eh?

If that happens and the boards are actually manufacturable... I think I'll faint. You'd need what? 30 PCB layers?....

As far as market share, it will go up. No doubt about that. The question is simply by how much. AMD will have more capacity next year. The question is how well they will be able to leverage that.

 

[Poncho]

Quad GTL+ eh?

If that happens and the boards are actually manufacturable... I think I'll faint. You'd need what? 30 PCB layers?....

I'm assuming you mean Quad FSB? If so... I should have a system in my hands within the next couple weeks. Not sure of ALL the details on it... but it's an MP platform so the normal contraints that you may see on DP and Desktop aren't there. Though I doubt it's 30 layer.

 

[duby229]

I'm assuming you mean Quad FSB? If so... I should have a system in my hands within the next couple weeks. Not sure of ALL the details on it... but it's an MP platform so the normal contraints that you may see on DP and Desktop aren't there. Though I doubt it's 30 layer.

I was just jabbin at ya

Really tho, 4 64 bit busses? That will be one hell of a lot of traces.

 

[tsuehpsyde]

At which point Intel will have upped their performance as well. Towards the end of next year you should start seeing 45nm as well as increased FSB with Bearlake chipset and let's not forget built in DX10 capapblilty with bearlake. Unless K8L increases performance 30-40% over current AMD cpus they've got a lot of catching up to do. And that's just in the desktop market. In the MP market you'll see a VERY competitive platform using NGMA and quad FSB around the same time as well. About a year after that you should see a HUGE can of whoopass in the DP server market. It's all about catch up with AMD at this point. Maybe they'll do it... but they are at a bit of a disadvantage with their problem ridden transition to 65nm and limited capacity. I bet by next year you'll see desktop/mobile marketshare decline back down to 15% or so.

You can honestly tell me that Intel's quad quadcore implementation (16 cores for those following along) is going to keep up with, and even beat AMD's K8L quad quadcore platform with direct connect architecture? I really, really don't think that'll be the case, reguardless of how many front side busses Intels slaps on. HT3.0 will slaughter it.

As I've said many times, I believe AMD is much more interested in securing their server marketshare dominance over Intel, and will make sure they keep it. If in the process this technology puts them in the forefront on the desktop market, all the better.

 

[Poncho]

I was just jabbin at ya

Really tho, 4 64 bit busses? That will be one hell of a lot of traces.

LOL... forget the traces, you should see the MCH. This is in a MP server though so the boards themselves are fucking HUGE... real estate isn't that big of an issue when it comes to MP, they are defenitly less cramped when it comes to components. Pretty cool stuff... wish I worked more directly on those projects. Oh well.... shoudl be cool to take a look at it when I get it.

 

[Poncho]

You can honestly tell me that Intel's quad quadcore implementation (16 cores for those following along) is going to keep up with, and even beat AMD's K8L quad quadcore platform with direct connect architecture?

Take a look at the DP market segment now with Core arch and dual FSBs. They are VERY competitive in this arena... what makes you think that the same principels won't apply to MP when they go that route? You have to remember that FBDIMMs help A LOT here as well. Intel has always been better in the MP market when it comes to what the consumer wants... their major limitation is the single, shared FSB. This platform will address this somewhat and they'll have it totally figured out with the next one using CSI or a variation of that.

As I've said many times, I believe AMD is much more interested in securing their server marketshare dominance over Intel, and will make sure they keep it. If in the process this technology puts them in the forefront on the desktop market, all the better.

I agree with you here... AMD is hitting server markets VERY hard, though Intel isn't sitting on it's ass here either. They see this and are fighting back. Server is #1 priority over in the blue camp.... and if they are on the right path IMHO from everything I've seen. So long as the powers that be don't fuck it up that is. LOL.

Desktop... that's a whole other can of worms. With Conroe, using the "shitty" FSB, beating the shit out of AMD in just about EVERY aspect you are really starting to see the differences between Server and Desktop... and what it takes to win those areas. IMC isn't needed in desktop... where it is needed in server.

 

[tsuehpsyde]

Not much to say about the first part, since we pretty much hit dead even on opinions, with a slight agree to disagree point there.

Desktop... that's a whole other can of worms. With Conroe, using the "shitty" FSB, beating the shit out of AMD in just about EVERY aspect you are really starting to see the differences between Server and Desktop... and what it takes to win those areas. IMC isn't needed in desktop... where it is needed in server.

I wouldn't exactly calling it 'beating the shit' out of. AMD's K8 architecture (read, OLD architecture) has been 'beating the shit' out of NetBurst for what, 3 years? Intel finally pulls their head out of their metaphorical ass and puts out C2D, a beefed up Pentium3 after they finally kicked out the marketing gurus and let the nerds do their thing. Even so, at multitasking the C2D isn't as badass as everyone seems to believe. In single threaded apps yes, C2D will stomp all over AMD's three year old K8 platform (surprising? not really) but this can mainly be attributed to the assloads of extra cache, and the FSB not being tacked since it's only using one thread. Start splitting up that FSB and cache, and C2D loses a little steam.

Then, AMD is still waiting to reply to C2D. K8 is losing out, I'm not denying that, but K8L is going to even up things nicely. With HT3.0, shared L3 cache and independant L2 cache per-core, as well as the performance upgrades and die shrinks, we will at that point have a fair comparison, and multible FSBs or not, I fully believe when it comes to multi-processor servers, Intel is going to get it's ass handed to it on a silver platter. If K8L scales as well as I hope it does on 65nm, then it'll just as easily keep up with C2D in the desktop realm if not beat it. Chances are solid at the very least, they'll be neck and neck, one or the other winning by a margin of percentages....plus, K8L is simply a tweaked K8, it's not even a completely new platform like C2D.



Ninja Edit: C2D is not a new platform, just a "tweaked" P3...but for a fair comparison, AMD diddn't put out a piece of shit, then go "back" to K8.

 

[sethk]

Intel finally pulls their head out of their metaphorical ass and puts out C2D, a beefed up Pentium3 after they finally kicked out the marketing gurus and let the nerds do their thing. Even so, at multitasking the C2D isn't as badass as everyone seems to believe. In single threaded apps yes, C2D will stomp all over AMD's three year old K8 platform (surprising? not really) but this can mainly be attributed to the assloads of extra cache, and the FSB not being tacked since it's only using one thread. Start splitting up that FSB and cache, and C2D loses a little steam.

I hate to be in a position of defending Intel, but the C2D architecture is fast for a lot of reasons, not just the extra cache. The C2D has much more to do with the PIII than the Netburst arch. but it has made some real strides in many areas, including instruction concurrency, number of issues / clock, cache structure and logic (not just size), branch prediction logic, etc. Bottom line, it can execute more instructions per clock than K8, and 4MB of cache on a dual core K8 wouldn't start to put a dent in its lead. To my point, look at the performance differences in the 2MB vs 4MB C2Ds. Slight drop but not a precipitous drop. The next gen AMD cores should do a lot to improve IPC, and the fact that AMD is slapping some L3 on there can't hurt either.

 

[The-One]

I hate to be in a position of defending Intel, but the C2D architecture is fast for a lot of reasons, not just the extra cache. The C2D has much more to do with the PIII than the Netburst arch. but it has made some real strides in many areas, including instruction concurrency, number of issues / clock, cache structure and logic (not just size), branch prediction logic, etc. Bottom line, it can execute more instructions per clock than K8, and 4MB of cache on a dual core K8 wouldn't start to put a dent in its lead. To my point, look at the performance differences in the 2MB vs 4MB C2Ds. Slight drop but not a precipitous drop. The next gen AMD cores should do a lot to improve IPC, and the fact that AMD is slapping some L3 on there can't hurt either.

Do you happen to know the percentage difference between the 2MB C2D and the 4MB?
I haven't seen an article yet testing them against eachother.


Thanks,
The-One

 

[Poncho]

Do you happen to know the percentage difference between the 2MB C2D and the 4MB?
I haven't seen an article yet testing them against eachother.


Thanks,
The-One

IIRC it's less than 10% for most benchmarks, though I may be wrong. I don't pay much attention to the 2mb versions..... since 4 is > 2.

 

[RobbieV]

Do you happen to know the percentage difference between the 2MB C2D and the 4MB?
I haven't seen an article yet testing them against eachother.


Thanks,
The-One

Anandtech did a comparison. http://anandtech.com/cpuchipsets/showdoc.aspx?i=2795&p=4

Some apps show very little change with the extra cache, others show an 8-10% performance gain with it. There was at least one other website that did a comparison as well but I don't recall which one it was.

 

[{NcsO}ReichstaG]

Intel is not Microsoft and is smart enought to avoid all the legal woes MSFT went through.

wtf are you talking about? Intel has gone through lots of "legal woes" and antitrust lawsuits.

 

[kill4killin]

Intel: 1
AMD: 0

WTF happened to AMD, they were doing SO well for like the Q4 05' and Q1 06' :'(

 

[savantu]

*intel stockholder warning*

I dunno... AMD has soooo much more money today than it has ever had before (not to mention tons of OEM and corporate support!). People today actually know who AMD is!! Same couldnt be said 4 years ago...
...

Actually AMD has more debt now than ever before and the next quarters are anything but nice. ( DELL also erodes their margins badly )

 

[savantu]

...

I don't think it is a matter of AMD running out of steam, it is a matter of AMD having reasonable expectations of the future. Did you all forget what happened on the Conroe launch already? We were told unofficially (just like this roadmap) that there would be an XE @ 3.2GHz with a 1333MHz bus. We are still waiting for that.

Sorry, but this roadmap seems unreasonably skewed to me. It sounds like Intel is simply making big claims to impress customers. From the comments in this thread, it sounds like is it working....

Stop for a moment and think : is Conroe at the limit of its scaling or is it being held back artificially ?

3.5GHz is the norm for Conroe while in stock it reaches only 2.93GHz.
It is fairly obvious that Intel has a huge ace up its sleeve : if need be a 3.33GHz/1333 XE is here.

You mention 1333 FSB.I assume you have no clue that
1.Woodcrest is using it for over 4 months
2.Less than a month from now Intel will launch CLoverton which guess what is QC and uses a 1333MHz FSB.

From the above it is obvious that Conroe XE has a lot of leverage to go up both in frequency and FSB.It is a matter of "we don't need it yet" not " we can't do it".

Conroe XE was expected to ship with a TDP of 95w.Actually the XE6800 has a 75w TDP.It is fairly obvious Intel held it back simply because there was no answer from AMD.

I expect Intel to launch a 2.93GHz E6800 (~$750) and a XE part 3.2/1066 or 3.33/1333 in november along with the 2.66GHz Kentsfield.

From then on they have a full year to get an extra 500-800MHz and a new process to use.I fail to see how 3.7-4GHz is unreasonable for a 45nm Penryn.I bet they could get a 3.66GHz/1333 part out of the 65nm process in the original 95w TDP for XE.

 

[visaris]

savantu,

Even if you want to fudge my numbers and raise Intel's starting clockrate, you see that Intel claims they will be able to raise clocks by around 1GHz in around 3 quarters. Sorry, I hate to ruin your dream, but it just isn't going to happen. If you are that confident, why don't you bookmark this post and put some money where your mouth is? I'd love to make some free cash. Perhaps I should remind you, before taking your money, that OC headroom does not directly translate into stock clock headroom; not even close.

----

Ah, and if I'm wrong about Clovertown's FSB, please provide a link. Everything I've been reading for a long time now has stated that Intel is trying their best to get the 1333MHz bus to work by release, but so far it wasn't looking good. Do a google search for "clovertown fsb" and you'll see hundreds of links like the following:

"Those 110 watts refer to a 2.67 GHz version of Clovertown, which apparently will be the fastest version offered at launch"
"Clovertown's FSB speed will drop from 1333 MHz in Woodcrest to 1066 MHz"
-- and so many more --
If I'm wrong, please post a link.

 

[savantu]

savantu,

Even if you want to fudge my numbers and raise Intel's starting clockrate, you see that Intel claims they will be able to raise clocks by around 1GHz in around 3 quarters. Sorry, I hate to ruin your dream, but it just isn't going to happen. If you are that confident, why don't you bookmark this post and put some money where your mouth is? I'd love to make some free cash. Perhaps I should remind you, before taking your money, that OC headroom does not directly translate into stock clock headroom; not even close.

----

...

WTF are u talking about ?

It is 1GHz or 30% in 5 quarters with a tweaked core and 45nm .Only from 45nm they gain 20%.I would be very surprised for Penryn not to be launched at 3.66-4GHz.

Btw , word of mouth is for Yorkfield to run at 3.66GHz from the start.

As for Cloverton I expect Poncho to know.

 

[savantu]

wtf are you talking about? Intel has gone through lots of "legal woes" and antitrust lawsuits.

Intel was never declared a monopoly , in fact most of its lawsuits were for patents.

 

[Donnie27]

i'll be sure to rub in your face just as much when K8L comes around...

In all honesty, I hope you can rub it in my Face. I know it is hard for you guys to figure it out but I hope like hell it's killer. Money talks and bullshit walks, that's what us hicks would say. Knowing AMD, if they do "open a can of whoopass" on Intel, will they pull another X2 BS Job with high assed prices? If they're fair, competitive and affordable, my bandwagon jumping ass will be right there to get one.

 

[Poncho]

As for Cloverton I expect Poncho to know.

Clovertown is launching @ 1333fsb with an 80 WATT version running 2.33/1333. Yummy.... and 2.66/1333 being the top model @ ~120WATT.

 

[Donnie27]

savantu,

Even if you want to fudge my numbers and raise Intel's starting clockrate, you see that Intel claims they will be able to raise clocks by around 1GHz in around 3 quarters. Sorry, I hate to ruin your dream, but it just isn't going to happen. If you are that confident, why don't you bookmark this post and put some money where your mouth is? I'd love to make some free cash. Perhaps I should remind you, before taking your money, that OC headroom does not directly translate into stock clock headroom; not even close.

----

Ah, and if I'm wrong about Clovertown's FSB, please provide a link. Everything I've been reading for a long time now has stated that Intel is trying their best to get the 1333MHz bus to work by release, but so far it wasn't looking good. Do a google search for "clovertown fsb" and you'll see hundreds of links like the following:

"Those 110 watts refer to a 2.67 GHz version of Clovertown, which apparently will be the fastest version offered at launch"
"Clovertown's FSB speed will drop from 1333 MHz in Woodcrest to 1066 MHz"
-- and so many more --
If I'm wrong, please post a link.

Hahaha!

Scratches head and thinks, WTF?

http://techreport.com/etc/2006q2/woodcrest/index.x?pg=3

Our dual Woodcrest test system was running at 3GHz on a 1333MHz front-side bus, with 4GB of FB-DIMM memory installed. Here's what CPU-Z had to say about it:

http://stage.tgdaily.com/2006/06/24/woodcrest_benchmarks/

In the end, it doesn't matter if it is running at 800MHz FSB, one or two FSBs as long as it is kicking ass and costs less.

 

[EternityZX9]

BLAH

All this is is speculative bullshit. I don't believe the specs of the K8L cores nor the specs for the wolfdale or yorkfields. It's a year away from now and a lot can change, especially considering since all of this "information" is coming from the vr-zone and hpec lately correct? Yeah...I'll wait until I see direct confirmation from AMD or Intel before I believe these "specs".

Far as we know this information could be called in from Intel's or AMD's marketing department just trying to yank our chain.

Hell, might as well link both these sites to INQUIRER because it seems like they are trying to compete with the Inquirer in the Special Olympics of MOST Random News EVAR to draw attention to themselves.

VR-Zone: "HEY HEY EHY, LOOK AT THE NEW G80 SPECS!!@!@#"

HPEC: "NUB! I GOT INFO THE AMD ALTAIR WILL PWN NEXT YEAR"

VR-Zone: "HEY YOU GUYZ! CHECK OUT INTELS NEW ROADMAP MADE IT MYSELF"

This just in at the INQ: "ATI's R600 will obviosly be moe powrful than the G80...I don't have any concrte prof that this is true, but I'l lambast it acros the internt just beause I feel pike it...moe rumors breakin at 11...or wheever I fell like it."

(Note: spelling mistakes put on the Inquirer section on purpose, if you don't believe this is true, read some of the Inquirers news stories)

 

[MC FLMJIG]

Intel was never declared a monopoly , in fact most of its lawsuits were for patents.

Thanks, qft.

 

[savantu]

Clovertown is launching @ 1333fsb with an 80 WATT version running 2.33/1333. Yummy.... and 2.66/1333 being the top model @ ~120WATT.

Thanks , I knew that , just needed reassurance.