AMD VS Intel (Intels and AMD's Trunk Cards)

i cant seem to find it now but i read somewhere that amd was planning in the near future to slap 2 quad core cpus together and make something that is every bit oct core as clovertown/kentsfield is quad core im probably wrong but i think i read this on the inquirer probably
1-2 months ago could of been somewhere else anyone else read anything like this?
 
Serge84 said:
Intel = Must show off performance ES's ASAP. Make Hype known with benchmarks. Keep ppl waiting for 9months. Conroe beats AMD for a year. K10 comes out. Intel needs wolfdile and yorkfeild to stay conpeditive. Cus a 2.5ghz K10 quad star processor beats anything intel currently has by 40% witch is a 3ghz xeon still Conroe core based. Wolfdile at 4ghz and yorkfeild at 3.73ghz come out the same time K10 does... Funny how its the SAME TIME... Kuma is clocked at 2.9ghz the 4ghz equivlent of wolfdine and Agena is clocked at 2.6ghz the 3.4 to 3.73ghz equivlent of yorkfield.

Wolfdile and yorkfield are Intels trunk card to delay AMD's 40% dominance.

http://www.mikeshardware.co.uk/RoadmapQ307.htm

AMD = Keep quite, show NO benchmarks until release date has been met. Make NO HYPE. Keep customers guessing and waiting for a year. Only sell truth with worlds never backing up their clams. They never showed benchmarks of K8 until it was released on production, they just said it was going to make intel 2nd best for a few years. Did they lie about it? Nope, killed netburst until Conroe came along. AMD does the same. So is this 2003 all over again? Sure it IS!!!

I'm impressed with AMD. But this will force AMD to show off its trunk card very soon in the game. According to AMD... HKEPC and mobo manufactures state 3.5ghz capable K10's after 2008. AMD must tell mobo manufactures K10 specs to prepare for K10 compatiblity with current gen and next gen sockets to maintain compatiblity stability.

http://badhardware.blogspot.com/2006/10/amds-k8l-revealed-in-cray-rainier.html
Click to expand...

LOL really? AMD no hype? I could swear you were using one of their "hypes" in your very own post despite the fact we still have not seen any real proof of these. :rolleyes: They might as well be faster but guess what it is right now? HYPE

Why don't you just hang a fan boy flag on your signature?
 
merlingt1 said:
LOL really? AMD no hype? I could swear you were using one of their "hypes" in your very own post despite the fact we still have not seen any real proof of these. :rolleyes: They might as well be faster but guess what it is right now? HYPE

Why don't you just hang a fan boy flag on your signature?
Click to expand...

You fail to really understand do you about what I go for, matx is my thing and intel has no game there yet. AMD doesn't need benchmarks out in the opened to be beleaved. There reputation is legendary for wipeing teh floor with intel for the past 3 years until C2D came out at the 2ndH of last year. It will only get a year of dominince. Funny how you say there is not real proof when we have the reviewer admiting there are real ES's out getting benchmarked all over the place with these scores. AMD admits this all over the net.

Quit lieing out of your words, because you know for a fact your in denial as a intel fallower, for every fact you say is wrong with every word you type your just afraid of the truth and progress, that... that law every 2 years cpus would gane double the performance still holds. Why are you afraid of change so much? Go buy whatevers best at the time, looks like you'll just have to upgrade soon again. Your system will be opsolete again in next month, wile I can just put a opteron quad into my comp. Oh and this time I'm going all high end, hell I'll buy a AM2+ mobo as well just cus I have the ability to afford it now.

When the time comes then lets bring it on with the real benchmarks and see how your C2D or Q stands up to my rig with a 1270SE OCed with a AM2+ mobo and a HD2900XTX in my little super box. That will be one hell of a upgrade nomatter how much faster K10 is. But yeah we'll see in the end really who's right and who is not. Just don't make me say I told you so. But I'd put a wager on that. AMD tends to wait for its release rather then getting you all excited with ES's like intel does. AMD's stratagy is not about that, its about giving you what they say they will. AMD said we would get 65nm cpus with much lower power ratings and heat. Thats what we got. Now AMD says with K10 we will get 50%+ performance over intel, not current AMD stuff. So there for being 70% over K8 and 50% over C2Q.

Be that way about it, beleave what you want, I don't care, the truth will find us all very soon indeed. lmao

If they manage some OCing merical with wolfdile or yorkfield over 7ghz and AMD can't go much higher then 3, then why not I'll go for whatever performs the best, but I got to have a Matx setup, most of the reason I'm with AMD was cus they made the best matx ocers. C2D really doesn't offer the performance boost I was hoping for specially for matx setups they have no good stuff there even in the fatality crap. Still can't go over 400FSB so AMD it is. So its very doultful you'll see me with another intel setup besides what I have now. Yeah didn't know that did you, I have a E6400 and its not impressing me. matx selections are crap so a nogo with much of anything in a OC that would be worth it. I will not buy a ATX board when I will never use 2 pci-e or that many pci slots. Too big for my tastes. I like small rigs that can pack a punch so hard it can level a ATX system like that. Smaller is better, not the other way around. Why pay more when I can get the same features in a matx thats just as good as a ATX board.

Besides I like the small rig look. When you go to a lan party they are like, lol what in the hell is that piece of crap poor excuse for a system you got there. Ppl don't expect a small rig to be so powerful. I get the responce, what in the hell is that, how in the Frick are you getting that much FPS from a crap comp like that oh man that game looks so good, why is it so responcive what monster parts do you have in there? Because they are deceved by its small size. They check the connections and under the table to see if a really big computer is stuck under there faking them out, because they can't beleave such a small box is outperforming like a box 3 times as big rig. They are like OMG its a super lil box. When I'm out performing a rig thats 3x as big as mine and worth 3000 bucks when mine is only about $1000, and get the looks like that, you know you hurt there pride. <Grin> You never expect that from lil rigs.
 
Serge84 said:
... that law every 2 years cpus would gane double the performance still holds. ...
Click to expand...

I believe you're referring to Moore's Law. It has NOTHING to do with performance... it is the transistor count, but yes, you make transistors smaller, you can fit more in the same die space (or less) and theoretically gain performance.

But Serge you're missing the real point here with everybody pointing it out... These are HYPOTHETICAL numbers AMD is pooping out here... There isn't anything substantial to these claims. I recall a while back when AMD 'demoed' their Barcelona... they didn't even let anyone close to sneeze on it, how can I see if it's real if I can' sneeze on it? Huh? LOL anyway...

I'm sorry to hear you think everybody is in bed with Intel just because they're calling YOU a liar. Sorry.

Granted Intel does hold the Performance Crown right now, it could change with K10, or it could FLOP like 4x4, ahem... Intel is just kicking AMD while they're down... they're getting ready to release 45nm and soon 35nm... and AMD? BARELY to 65nm.

Anyway... where was I? My head hurts from reading this thread.

-Eric
 
lord_farfig said:
they're getting ready to release 45nm and soon 35nm... and AMD? BARELY to 65nm.
Click to expand...

Sorry dude but that is BS.... Intel will have 45nm out soon enough, but 32nm? I dont think so... AMD will likely not be any more then 6 months behind Intel on the transition to 45nm at the lates, and as far as 32nm, that wont happen till 2009-2010 at the earliest. At this point in time it is likely Intel will get to that node first based on historical evidence, but by how much of a lead time is not known.
 
empoy said:
do you expect us to believe this? :D
Click to expand...

If I remember correctly AMD mentioned on their conference call that they'd have 45nm available to customers mid-2008 (so not 6 months, probably more like 8 months). Process size isn't the be all and end all of performance (as was evidenced by A64@90nm and P4@65nm) either way, but we'll see how things clock on K10 as the 65nm process matures.
 
:) it really disappoints me that when AMD was on top with the a64 939everyone jumps on the bandwagon talking about how the underdog brought out a real gem to stick it to chipzilla and then months later chipzilla breaks out the c2d and the people who months before claimed to be a diehard AMD *!*!*!*!*!*!*!*!*!*!*!*!*!*!*!*!*!*!*!*!*!*!*!*! buy the c2d

Then everyone starts saying zomg AMD is done for wtf are they doing migrating to AM2 and buying ATI AM2 is inevitable when you need to upgrade to ddr2 u need a new cpu with the memory controller and a newmotherboard for ddr2 slot so deal with it imo AMD should cut prices on 939 cpus more than the AM2 but what technological advances do AM2 offer other than the 65nm tech which hasnt really effected performance so they didnt sandbag anyone with AM2 and im sure they will see good returns when ATI bounces back and

everyone keeps camplaining about how AMD is late with their releases
1. as i recall the C2D has dominated since the market Q3 06
2. the A64 X2's dominated the market for what Q2 04-Q3 06
3. AMD is expected to release desktop quad core in Q3-Q4 07
so what C2D dominates the market for just over a year how fast do u really expect AMD to be capable of bouncing back a corporation that has the resources and means to do something that is in their best interest they will do so

not only that but the c2d doesnt even dominate all aspects a 60$ brisbane which im confident to say can handle any games out today more than adequately

and then theres those people who show market statistics of AMDs losses that are basically saying oh AMD's recent losses basicaly=the cost of the purchasing of ATI so basically they could of almost broken even if they didnt buy ati i think they probably used the money from that time period when they spanked intel with the x2's to finance ATI this is when they arent the top performing company id say thats pretty damned good for an industry as R&D intensive as cpu manufacturing so what if they had losses imo if they really were in serious financial trouble we would see them laying off employees

anyone who thinks AMD is going under doesnt understand this industry

btw im still running my 3000+ Winchester and x850xt and still loving it:)
 
I'm still fully expecting AMD to lay off a bunch of ATi employees... I'm also still expecting AMD to spin off a few divisions that ATi has that AMD doesnt need.

But both of these are long term plans that will take time to implement. So we'll see how it goes.
 
Serge84 said:
Besides I like the small rig look. When you go to a lan party they are like, lol what in the hell is that piece of crap poor excuse for a system you got there. Ppl don't expect a small rig to be so powerful. I get the responce, what in the hell is that, how in the Frick are you getting that much FPS from a crap comp like that oh man that game looks so good, why is it so responcive what monster parts do you have in there? Because they are deceved by its small size. They check the connections and under the table to see if a really big computer is stuck under there faking them out, because they can't beleave such a small box is outperforming like a box 3 times as big rig. They are like OMG its a super lil box. When I'm out performing a rig thats 3x as big as mine and worth 3000 bucks when mine is only about $1000, and get the looks like that, you know you hurt there pride. <Grin> You never expect that from lil rigs.
Click to expand...

What un-developed 3rd world country do you come from? SFF cases have been popular for 5 years at least.

I will not buy a ATX board when I will never use 2 pci-e or that many pci slots. Too big for my tastes. I like small rigs that can pack a punch so hard it can level a ATX system like that. Smaller is better, not the other way around.
Click to expand...

So you can cut a corner off a post-it note and cover your wang? It's actually refreshing to see someone not try to over inflate their e-penis......but I'm pretty sure the chicks hate it.

BTW anything posted from fudzilla will NOT be taken as proof around here. Hack into Kyle or Chris' computer and show us their benchmarks and then we'll believe you.
 
Doesn't anyone notice that these 'benchmarks' are all synthetic benchmarks...
More to the point, they are specfp and Sandra... benchmarks that Athlon X2 even beats C2D in *today* (advantage of the integrated memory controller).
However, Athlon X2 cannot convert this into ANY real-world advantage whatsoever.

Therefore I'm not surprised if the same goes for Barcelona... After all, the architecture is based on the Athlon X2, so it's probably good at the same things that X2 is good at.
They probably 'conveniently' picked these particular benchmarks because they gave the most impressive results.

So I'd like to see some real-world applications, like Intel has demonstrated with Penryn, and see how Barcelona does there.
 
Any architectural benefits of K10 will be mitigated by intel's 45nm process. Higher clocks (like someone on pg. 1 they could increase them by almost a full ghz across the board if they wanted), lower power consumption and lower costs as well from the 45nm wafers. Now factor in all the $billions intel has in the bank, they can just keep up this brutal price war indefinitely while regaining market share.

By the time AMD gets to 45nm (if they don't go bankrupt in the process), Nehalem will be ready with the integrated memory controller. Then intel goes 32nm for another significant IPC advantage and whatever oem partners AMD has left jump ship for good (assuming AMD survives this long).

I haven't even mentioned the laptop market where watts rule all and where AMD will have and even harder time being a die-shrink behind. I don't see many oem's buying 65nm turions over 45nm centrino.
 
duby229 said:
Sorry dude but that is BS.... Intel will have 45nm out soon enough, but 32nm? I dont think so... AMD will likely not be any more then 6 months behind Intel on the transition to 45nm at the lates, and as far as 32nm, that wont happen till 2009-2010 at the earliest. At this point in time it is likely Intel will get to that node first based on historical evidence, but by how much of a lead time is not known.
Click to expand...

Intel laid out their plans to launch the next smaller sized every 2 years. 65nm will be 2 years old when? 45nm launches this year. This time in 2009 for 32nm and before 2010. Who knows though, they could fail just like anyone else.

It is real funny how some take AMD's word while these same folks questioned tests done by Geeks all over the web on Conroe;)

Duby said:
AMD will likely not be any more then 6 months behind Intel on the transition to 45nm at the lates,
Click to expand...

In your wildest dreams only.

1. as i recall the C2D has dominated since the market Q3 06
2. the A64 X2's dominated the market for what Q2 04-Q3 06
3. AMD is expected to release desktop quad core in Q3-Q4 07
Click to expand...

A64 X2 didn't ship in Q2-04;) AMD shipped A64-X2 4200 and 4800 plus in Q2-05 during the month of May. Example.

AMD's official Launch.
 
Considering that Intel wont have volumes till q1-08 at the earliest, anything released before then will be paper.

I figure 6 months between volume releases. If Intel launches 45nm this, get used to waiting for it.
 
I love the OP's posting prices for X2 6000+ as of April 9 and trying to bash Intel for having its competing E6700 priced much too high. E6700 has been at the same level for quite some time, while AMD was and has been asking far too much money for inferior products since Core 2 was released.

Not to even bring up power consumption, which completely destroys an already-lost argument.
 
lord_farfig said:
<removed>
Click to expand...

I know this word has been outlawed... But this is flat out trolling.

This is the AMD forum dude.. Your are bound to run into more then a few people where comments like this will offend. It doesnt matter how many Celerons have been sold. We know Intel is a bigger company... We are not stupid. But that doesnt somehow make them better...
 
Whatever...

I was just trying to point out that AMD is 'saying' that Barcelona will be so much better(it very well could be), which is what OP is saying too... but there isn't any tangible evidence of this, or at least to me or you. Maybe just because of NDAs and whatnot, but I also think the OP was trying to point out AMD is laying low too... but I think it is from all the financial pressure they're trying to keep quiet hoping they'll smash the market with Barcelona.

Sorry, I was just being an ass with the Celeron remark... I know they're NOT better in anyway shape or form, it was a stupid fact.
 
duby229 said:
Considering that Intel wont have volumes till q1-08 at the earliest, anything released before then will be paper.

I figure 6 months between volume releases. If Intel launches 45nm this, get used to waiting for it.
Click to expand...


Please note that you said the same thing about Conroe last year. I saved what you said about Late Sept. and not ramping up until 07.
 
Donnie27 said:
Please note that you said the same thing about Conroe last year. I saved what you said about Late Sept. and not ramping up until 07.
Click to expand...

just curious, can i see? i remember conroe hitting and supply being REALLY tight to the point where it basically was a paper launch until a while later. :(
 
Donnie27 said:
Please note that you said the same thing about Conroe last year. I saved what you said about Late Sept. and not ramping up until 07.
Click to expand...

Yep, your right I was definitely wrong about Conroe. I under estimated availability and performance. It took a while for it to become available, but not as long as I thought.

I think Intel did an awesome job with Conroe. It out did my expectations by a large margin
 
(cf)Eclipse said:
just curious, can i see? i remember conroe hitting and supply being REALLY tight to the point where it basically was a paper launch until a while later. :(
Click to expand...

The supply issue was entirely made up. Intel's bumbling Paul Ortellini screwed up stating shipment rates by saying "weeks" instead of "days" at launch.

Though it was not the hardest launch ever, the chips were accessible in the first week, supply was tight for another week+, then became plentiful albeit at a premium as retailers saw fit to take advantage of the confusion. Heck, almost every commentator ran sprinting into their own mouth.

IIRC, prices dipped at launch then surged way past retail on the higher models. By early August, an e6300+ds3 combo could be had from the Egg for $331 which was straight-up retail. The same appplied to the other C2D chips. Later, when Intel disclosed the amount shipped, it became apparent there had been much ado about nothing, well to a few attentive folks, that is.

I found a link
 
First off forgive the rambling. But I read an article from IDC stating that the next round of processors was going to use some sort of dynamic processor loading, for lack of remembering the term. Basically (for example) you have a dual core cpu, each core at 2.5ghz. 1 core is maxed out and wants more power. Cpu 2 is sitting around with it's "hand toe" crammed up it's butt doing nothing. With a little bit of FM technology (frickin' magic) they let the single threaded app use the unused 2nd cpu and add it to the 1st; in theory making up to a single 5ghz cpu. Is this going to be a feature of Penryn? Or is this just an inbetween step from conroe to penryn?
 
SpoogeMonkey said:
First off forgive the rambling. But I read an article from IDC stating that the next round of processors was going to use some sort of dynamic processor loading, for lack of remembering the term. Basically (for example) you have a dual core cpu, each core at 2.5ghz. 1 core is maxed out and wants more power. Cpu 2 is sitting around with it's "hand toe" crammed up it's butt doing nothing. With a little bit of FM technology (frickin' magic) they let the single threaded app use the unused 2nd cpu and add it to the 1st; in theory making up to a single 5ghz cpu. Is this going to be a feature of Penryn? Or is this just an inbetween step from conroe to penryn?
Click to expand...

I remember vaguely the same thing was said about Barcelona about a year ago... People were calling it "reverse-ht" I dont honestly know if this is possible, but I cant think of a single good way to do it. Latency between cores would be waaaayyyy too high.

The theory at that time was that all they would have to do is sort of "stripe" the front ends of both cores, and have the decoders work in unison on the same thread. I just dont see any good way to do that. In addition the backends would simply be too far away. If you have data in a register on core0, and it needs to be moved to a register on core1... as far as the cpu is concerned that would take a lifetime... Comparatively speaking.

I'm no expert, but I just dont see how it would work... Of course there are alot of smarter people in this world then me....
 
If amd and intel both have this similar tech in the next round of cpus (assuming they can make it work decently), they'll have sealed the fate of single cores and found the holy grail of multi core cpu advertising.
 
Well, I can imagine that they could basically create a single-core processor, with HT, but with as many execution units as 2 or 4 cores (or even more).

AMD cannot currently do this efficiently, because their cores are physically separated cores, each with their own cache. So you can't easily move data from one core to the next.

Core2 Duo already is a single-core processor, partly, since both cores share a single L2-cache.
This makes it easier to share data between units from different cores. If they can find a way to also make L1 shared in some way, then effectively you have one 'uberpipeline' with a single data store. If you then add HyperThreading logic to dynamically allocate execution units, you'll have a processor that can use all its execution units for all threads it can run at a time.
I'm not sure if L1 is fast enough for this already, L2 probably isn't. So Intel still needs to make quite some adjustments to their design, I suppose. But they're a small step ahead of AMD.

I suppose eventually you'll get a single core with a lot of execution units, capable of running multiple threads. Basically a P4 HT, squared.
HT made perfect sense, but until now they've just put it on regular single-core processors. Now that we can put two or even four cores on a single die, it would make sense to just make it a single ubercore, which uses HT to run multiple threads.
I wonder if that's the idea. Core2 already has almost twice the amount of functional units as the P4 did, even on a per-core level, so HT may already be a good idea, even without merging the two cores into a single pipeline.
 
(cf)Eclipse said:
just curious, can i see? i remember conroe hitting and supply being REALLY tight to the point where it basically was a paper launch until a while later. :(
Click to expand...

Duby said:
Yep, your right I was definitely wrong about Conroe. I under estimated availability and performance. It took a while for it to become available, but not as long as I thought.

I think Intel did an awesome job with Conroe. It out did my expectations by a large margin
Click to expand...

Why pile on Elclipse, he knows what we talked about.
 
Donnie27 said:
Why pile on Elclipse, he knows what we talked about.
Click to expand...

I was responding to you Donnie. You quoted me, and I was responding to you. Your were right about Conroe... I was wrong. It did a lot better then I thought it would.
 
duby229 said:
I was responding to you Donnie. You quoted me, and I was responding to you. Your were right about Conroe... I was wrong. It did a lot better then I thought it would.
Click to expand...

I know, it would have seemed like I was bashing you (pile on) if I posted links to that thread after you already acknowledged it. I'm glad there's been little to no flame-bait here and I was trying to keep it that way:)

Again, Intel, just like anyone else, could slip up and ship later. But the point I tried to make, it would have to be an unexpected slip up because Intel has said things are going great for 45nm.
 
You must log in or register to reply here.
Back
Top