I honestly think that 4S is reaching the end of the road. When you start putting so many cores into a single package, who needs 4P? 4P is already such a narrow market... The increasing power of 2P will put even more pressure on this already niche market!
Consider this for a moment: If Intel adds Dempsy-style internal bus to quad-core Penryn, then the 4 cores will look like just a single load to the FSB, and arguably, Intel will be able to pack two of these qaud-cores onto the same package (if the package has enough space). So, in theory, a Dempsy-style internal bus will allow Intel to have 8-core chips by the end of next year. If this happens (I understand that that is a big if, but Intel is desperate to claim sustained leadership, thus who knows), we are looking at 2P systems with 16 cores!! Going forward, rumors are that Intel will reintroduce its hyperthreading. That will put 32 logical processors on a 2P system. This is bound to make the 4P market ridiculously niche. So what does that mean? Does Intel really need the CSI? Afterall, the dual FSB is more than sufficient for the 2P market...
In my opinion, it does need CSI. The problem is not 4P or 8P--that is a dying breed. The real problem is Torrenza--AMD's ability to couple third-party processors with its own. At IDF, Intel announced that it will open up its FSB to third parties. Pardon my french, but who gives a f&*^ing @#$%? Why would anyone want to put their co-processor on an FSB that Intel is always in a hurry to upgrade? With HT, you can arguably negotiate the different links at different speeds, and hence the third parties do not have to upgrade their HT logic to keep up with AMD. On the other hand, since the entire FSB system will be limited by the speed of the slowest component, third parties have no choice but to run with Intel or be rendered obsolete. And that is exactly what they don't want. Thus, if Intel wants to provide a Torrenza-like capability, it needs a point-to-point interconnect that can be negotiated independently. FSB just won't cut it.
Does Intel need to provide a Torrenza-like solution? Frankly, I don't know. Today there are not many co-processor applications where the co-processor has to interact with the main processor on a clock-cycle-by-clock-cycle basis. But arguably, that is because, presently there is no technology that allows a co-processor to interact with the CPU that closely. AMD's Torrenza will make that possible for the first time--and who knows, it might even catch on? Intel cannot afford to ignore Torrenza, that's the bottom line. And that is why, it needs a cache-coherent point-to-point interconnect solution. Maybe it's CSI, maybe it's something else. But they need one for sure...
Subscribe to:
Post Comments (Atom)
27 comments:
Since you need a lesson in just about everything, let me teach you.
Wow, 8-Core CPU's? Jeez, how fun!
Only 2-Sockets? Really? Wow, that's great!
Oh but, wait, you forget...co-processors need a socket too :( -- bye bye goes your stupid idea.
As you don't know, Socket 940 has Co-Processors ALREADY that fit into another Socket on 2P, 4P & 8P (you probably didn't know 4P boards upgrade to 8P, as there is no single motherboard w/ 8-Sockets on it) and Socket 1207 increases this performance by utilizing HT 3.0.
Intel & CSI is a very strange road, one that is riddled with ignorance and much arrogance on all parts. Will it be released? Who knows, what's known for sure is Torrenza will be the DK for over a year until Intel can respond (if they can).
2-Socket will never be king over 4S or 8S. You say Intel's bringing back HyperThreading? Maybe you lost the memo that said "HyperThreading reduces performance on most server apps" -- but hey, not that important, right?
If you can have 8-Core CPU's, I want a 4-Socket CPU w/ 2 of those and 2 Co-Processors for Multi-Media which give me 1000x more performance than a 2S. Can Intel give me that? No, but AMD-ATI can.
2-Socket will never be king over 4S or 8S.
King in what way? Obviously not performance, but they make up a huge majority of the sales.
C2D was trying to say that 2S will be king over 4S+ by having 16-Cores, thus implying 2S can outperform 4S+, and I had to correct him.
"This is bound to make the 4P market ridiculously niche."
That does not imply that 2P is faster than MP, it implies that with multiple cores, MP is going to become even more niche than it already is.
Current 4P systems have 8 cores max. A 2P system has 8 cores coming this Q406. Add 1-2 years, with 45nm, Intel should bring 8 core chips per socket. In a few months, a 2P would have same amount of cores as 4P, a few years from now, a 2P system would have more cores than a 4P system of today.
http://www.aceshardware.com/forums/read_post.jsp?id=120066919&forumid=1
I've found the Bunny from the Inquirer!
His posts say 65nm K8s for sure.
http://www.aceshardware.com/forums/read_post.jsp?id=120066245&forumid=1
9.11.06..
"No, AMD is not about to put out anything new with any new goodies for over a year. By then Intel will have the Penryn cores out and those WILL have a lot of goodies. Have your mother put some Intel roadmaps on the top basement step when she places your food there."
http://www.aceshardware.com/forums/read_post.jsp?id=120067118&forumid=1
He at least got the Sun/AMD common socket thing right:D
http://www.aceshardware.com/forums/read_post.jsp?id=120066879&forumid=1
"I still expect Conroe to beat it core for core in all cases, but AMD has a chance when the platforms are taken into account, IE >1 core systems."
Intel still ftw in desktops!
http://www.aceshardware.com/forums/read_post.jsp?id=120067743&forumid=1
Says ATI working on raytracing, as is Intel.
http://www.aceshardware.com/forums/read_post.jsp?id=120065810&forumid=1
Claims lots of K8L ES in Taiwan.
http://www.aceshardware.com/forums/read_post.jsp?id=120066243&forumid=1
Anyone doubting Intel's smarts are in for a shock[in regards to phase change RAM]
Well, I see the last reply from where...?
Oh, yes, it's from Tru64.
Why do someone, copy and paste their replies from one blog to another and bolding the same pharse?
Oh, I forgot, it's Intel fanboys.
"I still expect Conroe to beat it core for core in all cases, but AMD has a chance when the platforms are taken into account, IE >1 core systems."
Intel still ftw in desktops!
Yeah, Conroe and Kentsfiled will beat AMD K8L when it's compared apples to apples. That's OLD news. Retards!
I just want to say two things: I never said 2P will outperform 4P, if the processors are equivalent. However, when 2P servers of today start providing power of 4P servers of yesterday, the demand for 4P goes down, because the performance requirement on the server side does not go up by that much. At some point, you start wondering, what is better, a 4P server or a cluster of 2P servers.
Second, in my post, I pointed out that Intel needs CSI for coprocessors. So pointing out that Coprocessors do not go into today's 2P systems is irrelevent.
Only 2-Sockets? Really? Wow, that's great!
Oh but, wait, you forget...co-processors need a socket too :( -- bye bye goes your stupid idea.
BTW Mike, 8C 1P systems themselves will reduce the need for 4P systems. However, they still will not be able to sport any coprocessors, since no one will build those for FSB. The whol point of my original post was this.
C2D, just stop posting, you don't know wtf you're talking about.
C2D, just stop posting, you don't know wtf you're talking about.
MMM back to namecalling (surprise!!).
"MMM back to namecalling (surprise!!)."
That's namecalling? Wow...guess I've been out of 6th grade too long. Tell me, how is integrated math for ya?
http://www.xbitlabs.com/news/cpu/display/20060731233200.html
First they claim 4x4 will be 80% better at Cinebench.
http://www.amd.com/us-en/assets/content_type/DownloadableAssets/PhilHesterAMDAnalystDayV2.pdf
Here claims of 60% better "performance per watt". Geez, isn't comparing by clocks confusing enough?
http://www.hkepc.com/bbs/itnews.php?tid=678736&starttime=0&endtime=0
Then here claiming 40% improvement.
http://www.vr-zone.com/?i=4109
If AMD it's true that it's 40% better than K8, which should be 17% better than Conroe... A 2.9 Altair should theoretically be at best case better than a 3.4 Conroe. Yorkfield should bring that down[even possibly gaining with all the new features], and VR-Zone reports that these will come in 3.46-3.73[nullifying the per clock gains of Altair] parts.
PS Mike, condescending, name calling, all the same;)
Will the author of the last reply, please stop posting links to other "websites", telling others that Core 2 Quad is far better than K8L, blah, blah, blah.
If you want to show me that Core 2 Quad is better, then show me a benchmark testing all by yourself, because I don't want benchmarks from Intel paid pumpers. (esp. TomsHardware)
More to the point, please do not focus on the figures announced by Intel or AMD, these figures are only PR strategies.
They do not reflect the same results when the final products are out.
Lastly, if you want to proofread all AMD documents and slides please do it yourself, but don't show your findings to us, nobody wants to see them.
What a retard.
Where have I claimed that Core 2 Quad[Kentsfield] was better than K8L[not denying the possibility either]
http://news.zdnet.com/2100-9584_22-6041120.html
Intel claimed months before Conroe release that clock for clock, Conroe would be 20% better than K8. And that's what we got.
The whole world does not revolve around you, just because you don't care that AMD is constantly contradicting itself does not mean others don't ;)
Yeah, yeah, Intel always have very "innovative" projects worldwide, however there's no sign of final product(s). Yeah, vapourwares always exists.
Your reply:
"If AMD it's true that it's 40% better than K8, which should be 17% better than Conroe... A 2.9 Altair should theoretically be at best case better than a 3.4 Conroe. Yorkfield should bring that down[even possibly gaining with all the new features], and VR-Zone reports that these will come in 3.46-3.73[nullifying the per clock gains of Altair] parts."
4x4 right?
60%...40%...20%.....0% then -100%....
Oh, do you want to prove that Altair (K8L) has no advantage over the Yorksfield (future Core 2 Quad)?
It seems that both Altair and Yolksfield are not even released and on shelfs yet.
You say that Yolksfield can beat Altair, right?
"Where have I claimed that Core 2 Quad[Kentsfield] was better than K8L[not denying the possibility either]"
There, a "Intel Quad-core" vs K8L. Retard!
Proofs? Justifications to support yourself?
Come on, give me technical details, please, not just speculations from websites and things typed on your own based merely on assumptions.
I can't even trust them. The only thing I can trust is the final products in the market, not that the theories (good example as Pentium 4 EE, yeah ultra deep pipelines + ultra high working frequencies = Ultra high heat dissipation.)
Please, give me proofs, not just website links, retard.
-----------------------------------
Your reply:
"The whole world does not revolve around you, just because you don't care that AMD is constantly contradicting itself does not mean others don't."
Don't be silly, retard.
I care for AMD, and I have a collection of slides of AMD presentations (Do I need to show some, leave me an e-mail address if you want), I do care for the contradictions AMD made, but I won't magnify it a few zillions of time, and shout it loud in the public.
If I find you contradict yourself, and I shout out loud in the public for that, what would you feel? Yeah, maybe you think it's just a reminder of how you're doing, but your errors were announced to everyone and people have a bad impression on you. What would you think?
Yeah, you're surely a genius to work on tabloids, because they do enlarge the errors of some poeple a thousand zillion of times and bring it up as the headline of the day.
Go work in The Inquirer or something, they may have pity on you and let you work as a freelance journalist, oh, you can work with Charlie Demerjian too! Don't forget to sent him a greatful thanks for his truly good looking bunny suit in IDF. ;)
Looking forward to your future replies, retard!
"Yeah, yeah, Intel always have very "innovative" projects worldwide, however there's no sign of final product(s). Yeah, vapourwares always exists."
..I do not understand that nor know what it is in response to.
"4x4 right?
60%...40%...20%.....0% then -100%...."
...AMD gave us the declining numbers;)
Yep, Yorkfield can beat Altair. No certainty, but based on what we've seen from HKEPC and VR-Zone..Charlie of Inquirer says Conroe of 06 should beat Altair of 07[at least on desktop].
http://www.aceshardware.com/forums/read_post.jsp?id=120066879&forumid=1
PS, please stop the flaming;) How am I magnifying it 'a zillion times' by acknowledging it? If you found me contradicting myself, I would watch myself more in the future before spewing BS.
The dual 2.6, 2.8, 3.0 4x4 is still worse than Kentsfield on single threaded.
2.66*1.2[Core 2 advantage over K8] 3.192 > 3.0 max.
On threaded apps, AMD has claimed 80% over dual core. 1*1.8=1.8
Intel has claimed 70% over dual core. 1*1.2[Core 2 over K8]*1.7=2.04
2.04 > 1.8
A weighted average[scores from apps where CPU matters] shows Core 2 20% over K8. Obviously you can't expect Core 2 to play Solitaire 20% faster;) If you use your PC for office/browsing, expect squat.
The HardOCP article compared the E6700/X6800/FX62 with a $400 graphics card that would've limited the results. Not everyone needs/can afford a huge monitor bulky/hot graphics card:) And they talk about real world;)
If you were going to post about how you were going to link it then you might as well have since you just wasted couple seconds typing that. And just because I can find HardOCP does not mean that others are stupid because they can't.
No need to label someone because they can't find an article?
I know HardOCP..I don't see why one needs to resort to name calling ala MMM;)
..Some sites like to bench sites on the 'lowly' 1024x768 to test the 'potential' of the CPU. Some call it synthetic since 'real' gamers have huge LCDs. I see those as showing the power of the PC as well as showing the performance on a resolution that most have. And I was trying to say that $400 is too much for a GPU[as is $500+ for a CPU]. I'd prefer they'd test something in the $250 range.
http://common.ziffdavisinternet.com/util_get_image/13/0,1425,sz=1&i=139042,00.gif
^^Lower resolution result in performance more dependent on CPU.
If a CPU increase does not matter to some folks [office, browsing]..They wouldn't be able to tell that they need anything better nor tell the difference between 3% and 20%;)
If you have a $400 GPU then you can probably afford everything that money can buy, so you're right, lower resolutions won't matter:) But not everyone is a hardcore gamer, and even if one does enjoy the occasional Halo or so, even a $75 GPU can do the job. Using more mainstream parts, [H] would've been able to see better results, as most sites are reviewing at 1024x678.
Link does work, double click on it;)
;)
To get the link to work..
1. Double click on it
2. Ctrl+C
3. Ctrl+V in address bar
But who has time to game, with all these durned DVDS to illegalize;) A Core 2[or fast X2] should boost encoding and reduce CPU consumption. A GPU for kicks can accelerate basic 3D and HD, but I would never need an expensive one. If they bottleneck the system to a point where even most enthusiasts could not imagine buying, then it's not any more 'real world' then the tests they complain of.
I'm not sure what you're saying, but the weaker the GPU, the more reliant the game is on the CPU.
But why would anyone want to make their bottleneck the GPU? Oh yeah, graphic whores:p. Those monsters are too expensive and hot for me though. While [H] rambles on about 'real world situations', they should've used a mainstream GPU.
And back to the original ~20% discussion..That is 20% over X2.
Many people that upgrade from Pentium 4/D, single core Athlon 64[tempted by good PR] would notice a greater improvement.
Anyone running X2 3800+ would be foolish to go Core 2, especially if they just get E6300(unless they overclock), unless they're going for ultimate performance, and in that case, they're already filthy rich so performance per price does not matter.
Ok I confused myself, my point was that they used a top of the line GPU with top of the line settings:) All I would've asked for was for them to use a more mainstream GPU and mainstream resolution, since the difference in CPUs is more obvious at lower settings.
ashenman, are there any forums you frequent? I feel censored lately;)
Post a Comment