Watch out Intel, there's an uprising in Barcelona...

[Biscuit]

Is there a particular reason for not wanting to use Intel?

Had bad experiences with crappy pentiums in the early days and never really been dissapointed with AMD. I guess i prefer the way they work & im a bit of a fanboy

[Blitzen]

You think?

Its a fact

[funnelhead]

Intel stuck the quad to £165 just to make sure anyone thinking about a new system didnt hang aroung long enough to consider Barcelona.

but i dont think it was out of fear, just spite

[redlight]

Its a fact

Vistas "fine" its a fact.

[badass]

well i for one am very happy someone finally pointed out, and maybe it should have been in bold, that the power ratings don't include northbridge mem controller on intel systems, which is what a pretty large portion of the northbridge power consumption is down to on intel systems.

When we all went ath 64 suddenly chipset cooling was really not at all important, on nf2, and intel p4 setups cooling the norty was massively important, as it is for the very highest fsb clocks on core 2 duo systems.


I have to say i was a little surprised to see 3Ghz Barcelona's demo'd that should see the light of day(i would assume as a FX branded cpu sometime in nov/december) this year. i've been guessing that firstly, you have to ignore the speed's opto's demo at, like the K8 opto's demo'ing at 1.4Ghz shortly before release. I would assume that desktop will be launching with chips probably up to 2.5Ghz, and people have been saying for a while, and it looks entirely possible, that a 2.5Ghz could beat a 3Ghz kentsfield, probably even their 3.33ghz penryn that was benched on hexus. I can't see a penryn beating barcelona. as for are intel worried? why else would you annouce expected release speeds for server chips on the next core, because you think you know roughly what speeds AMD are bringing out , but refuse to post info on speeds of any of the desktop range, due out after barcelona. Looks to me, and the world, like Intel don't know how AMD will compete, and even think they might have to pump up their speeds to compete.

I would assume that AMD will simply be faster, Intel will play the numbers game, as in, we have 545 fab's, AMD have 4, we'll sell our quads for a number you simply can't compete with. Now, some people buy for the pure performance and damn the money, most people buy on value.

the other problem is, jumping to 2 cores , then 4 put us massively ahead of the software need for power. there just aren't any games that will show a big enough difference between a £60 dual core and a £600 quad core at the moment, or for a while. since the home market is mostly gaming + internet + movies, gaming is where you want power.

Thats a whole lot of Pro AMD hope based on no evidence. There is nothing in the Public domain, including AMD roadmaps to suiggest anything like those kind of speeds next year, let alont this year!

[Noxvayl]

Thats a whole lot of Pro AMD hope based on no evidence. There is nothing in the Public domain, including AMD roadmaps to suiggest anything like those kind of speeds next year, let alont this year!

You will find that history tends to repeat itself and the basis of his claims are from the similarities of the K-10 launch to the K-8. Remember there have been no significant time consuming changes to that architecture for more than 2 years now, what could they have been doing in their extra time????

K-10 has been developed very carefully and will no doubt kick any Quad of Intels simply because of its true native quad die(the cores are linked on the die itself not through the bus like Intels). Also along with the CPU architecture having a major overhaul AMD are launching a new chipset of there own with HT3.0 which will run considerably higher than Intels 1333Mhz FSB which logically means it would be better being able to communicate with the peripherals faster.

Since the K-8 launch AMD to most have become complacent but behind the scenes without anyone knowing or suspecting they were working on the K-10. As with any new architecture you should always expect it to be the best, when Intel first brought out the Pentium family it was exceedingly great and then AMD brought out K-8 which kicked its arse, then Intel brought out C2D which whooped AMD and now its going swing back.

To those of you looking forward to the Nehalem and Penrhyn from what I've heard there are no major architectural changes from the C2D family just a drop in manufacturing process which allows more cache on the die and more instructions, they won't be the step up that C2D was and the performance(15-30% extra) I expect to be close to if not below the Barcelona family from AMD.

[Funkstar]

Since the K-8 launch AMD to most have become complacent but behind the scenes without anyone knowing or suspecting they were working on the K-10.

So everyone thinks once a processor generation is released, the designers just kick back and relax? Do they go paint their boats wile they wait for the call from their boss to get working on the next one when Intel are kicking their ass with the Core2Duo?

K-10 will have been in development since the K-8 was released, elements of it have probably been in R&D since before the K-8. The K-11 or K-12 will also have a development team of some desription working on it just now too. Just like Microsoft has a team working on Vistas replacement.

[kalniel]

K-10 has been developed very carefully and will no doubt kick any Quad of Intels simply because of its true native quad die(the cores are linked on the die itself not through the bus like Intels). Also along with the CPU architecture having a major overhaul AMD are launching a new chipset of there own with HT3.0 which will run considerably higher than Intels 1333Mhz FSB which logically means it would be better being able to communicate with the peripherals faster.

What an interesting statement We'll see soon enough, but I doubt just being native quad is nearly enough to take the performance crown - I think Intel will still have the fastest chip.

To those of you looking forward to the Nehalem and Penrhyn from what I've heard there are no major architectural changes from the C2D family just a drop in manufacturing process which allows more cache on the die and more instructions, they won't be the step up that C2D was and the performance(15-30% extra) I expect to be close to if not below the Barcelona family from AMD.

I think you need to read up a bit more - even Hexus' own article on the two proves that wrong! Nehalem is a completely new architecture, and the versions with onboard memory controller are most likely to blow away all that came before it.

[Sinizter]

AgentK obviously loves AMD to bits to be so filled with optimism.

[dfour]

Will Chipzilla be worried, probabley not, will they be watching almost certainly yes. Will AMD be sucessful in powering past chipzilla?? I certainly hope so or at a minimum be able to compete on a level playing field.

We NEED AMD to be competative to keep the market alive. AMD are in to deep and the only competition to chipzilla. What would happon if AMD falls??? It will be worse than the OS market with only a few minows scavenging for scraps

Maybe if AMD hadnt purchased ATI this state of affairs wouldnt be as far gone as it is but AMD needed a graphics partner to compete againts int*l with its intergrated graphics setups and its ties with nvidia.

The answer is 'yes' im a fanboy of AMD and ATI

[DR]

AMD's biggest partner pre aquiring ATI was infact NVIDIA. So I don't think Intel/NVIDIA working together is valid cause for AMD to splash out $5B the fact of the matter is they had to do something to move to a platform approach. AMD wanted to grow and were do dependant on their partners - SiS, VIA, NVIDIA, ATI to make chipsets to help them complete the platform.

AMD's issue was R600 was late, and didn't dominate. They have had to go back to the previous AMD model of competiting on price, rather than having the faster technology which just "sells"

Intel have also changed and become a lot more dynamic - attacking AMD on price point and technology and doing a pretty good job of it.

I agree though - we do need AMD and we do need them to be competitive, also Intel want them to be around and to be competitive. I've spoken to people inside Intel who are want AMD to attack them so they build better products.

There is no doubt that the 'system' of choice - currently, is Intel. Let's hope AMD's claims/promises can be delivered upon

[Kumagoro]

I think I have read someone where, PC pro perhaps, that a major mobo company stated that the
proper 4 core amd doesnt show up real performance benefits. I also read somewhere that AMD reckon
their 2 Ghz version is worth intel at 2.3 GHz.

[Kairiaku]

Personally, I'm keeping an eye out...

[Noxvayl]

I think you need to read up a bit more - even Hexus' own article on the two proves that wrong! Nehalem is a completely new architecture, and the versions with onboard memory controller are most likely to blow away all that came before it.

Sorry that was meant for the Penryn family only as the Penryn is an upgraded C2D... And for the onboard memory controller AMD already have that so that should only take away the advantage AMD have with memory intensive programs

The nehalem should improve on what AMD release in the Barcalona but remember its only coming out in 2008 H1 but with the Penryn only out at the end of 2007 its most likely to be nearer the end of 2008.

With the nehalem comes 8/16 cores but with what we all perceive to be a feeble attempt from AMD to get a QUAD out there(socket F) 2 AMD quad cores will then become the first octacore platform which will obviously take the performance crown(not having anything capable of matching 8 cores yet) but will come with a heck of a price tag.

I know I'm an AMD fanboy but there is nothing you can do about it...

Edit: Here is something I read when trying to find out when the Penryn and Phenom CPU's were being released and found it to be remarkably true... makes you think doesn't it:

"AMD lead in the use of DDR memory and better core architecture i.e., on chip L1 and L2 cache. What did intel do? it kept trying to increase MHz. Intel has now copied everything AMD did and is also about to try to copy amd's use of on-chip memory controllers. The problem for AMD is that Intel is copying everything and doing a better job of selling it. Along with anti-competetive (illegal) deals with OEM's, Intel has had a great deal of success. I think the Phenom chip will once again leapfrog the Core 2, but if AMD doesn't think ahead, Intel will copy cat and repeat history. Only next time, AMD may not survive and we will be left with Intel--and then where will they get great ideas for technological advances?"

[kalniel]

I would hardly call the shift to Core architecture 'trying to increase MHz' - that's a very misguided article whereever you found it - it's also completely liable

Saying Intel are just copying AMD is also really quite a blunt and misleading idea - phenom is copying core's shared cache, core still has 4IPC which AMD are struggling to copy. AMD have innovated in some, obvious, areas, but we've got to be honest here and say that Intel have probably innovated in more - they're just less obvious ones because the average punter (including myself here) doesn't understand the finer details of instruction enhancements etc. whereas a bulky concept like HT or integrated memory controller is easier.

[Noxvayl]

I would hardly call the shift to Core architecture 'trying to increase MHz' - that's a very misguided article whereever you found it - it's also completely liable

Saying Intel are just copying AMD is also really quite a blunt and misleading idea - phenom is copying core's shared cache, core still has 4IPC which AMD are struggling to copy. AMD have innovated in some, obvious, areas, but we've got to be honest here and say that Intel have probably innovated in more - they're just less obvious ones because the average punter (including myself here) doesn't understand the finer details of instruction enhancements etc. whereas a bulky concept like HT or integrated memory controller is easier.

The Mhz reference was to Pentium4 CPU's not the core architecture.

In the hexus article you referred me to on the Nehalem it even states and I quote: "Performance, predictably, should be good, yet it may become difficult to distinguish between it and AMD's offerings: both CPU companies appear to have arrived at the same design conclusion. Is Intel the new AMD, then? " which clearly implies both companies are doing the same thing which one side would refer to as copying.

You can't escape the fact that things have been "copied" and AMD most certainly could of done the same with the likes of hyper threading but didn't. The changes Intel have made that you refer to are mostly software where as what my quote referred to was hardware/architectural. Software is a one way street AMD and Intel have pretty much the same offerings differing only slightly, its architecture that has always separated the companies and it seems all that AMD have said they are doing Intel are doing as well.

For all we know that could of been done independently of what each other were doing meaning that the conclusions being the same could of been coincidence but then again how many of us believe in that.

Because I like AMD I would perceive it to be copying and because you like Intel you would say its coincidence but either way it happened and AMD + Intel are becoming much more similar...