Printable View
Yeah ,since you personally designed dozens of MPUs you already know that making those changes is like walk in the park. :rolleyes:Quote:
Originally Posted by Shintai
Heck,by the look of your post,it seems liek a 13 year old kid could have made a better job than AMD did,with such a small number of advancements.
Biggest and most rediculous understatement of the year. Congrats.Quote:
Originally Posted by Shintai
Also, increasing bus widths != larger buffers. I don't think you know what a buffer is.
whatever, star is more of a server oriented arch, surprise surprise. if it gets released in q2, it will give clovertown a very good run for its money. on the desktop, amd doesnt really care too much anymore. if it can satisfy dell, its set. the enthusiast market is what, probably less than 1% of the actual cpu market, whether we like it or not, we're not important. intel could continue to crush amd in the desktop, but again, only a small part of the market actually cares. nehalem for intel will make it very competitive in the server market, but amd is still ahead on the platform, as whatever intel's htx competitor is, its still one step behind since it depends on the high latency low bandwidth pci-e bus. then of course, nobody really knows how good cerebus will be. once amd gets its manufacturing facilities in line, it will be positioned to take a better run at intel on the desktop, but not yet, maybe in like 2009 or 2010.
Why do you say that? Intel did basicly the same thing with C2D from yonah. AMD basicly just copied what Intel did. It will be interesting to see if AMD's engineers are as good as Intels. What Intel did with C2D was impressive. I doubt that anyone will think amd's engineers are at 13 year old level. If K8l performs as well as C2D I would say AMD has done a wonderful job. But If anyone thinks K8L will be 20% better than C2D than I believe your in for a let down . Than as always I have to remind all again when you discuss K8L don't compare to C2D . As K8L will be up against. Wolf/York.Quote:
Originally Posted by informal
The real dog fight and what will determine AMD's future is Nehalem vs. AMD fusion. ATI has softened my position on AMD. SO I will watch this development with much interest from both sides. Who knows maybe ATI can save AMD.
I wrote that as a sarcastic comment to Shintai's post.You need to pay more attention.Quote:
Originally Posted by Turtle 1
And yes,i think AMD has great eng. department(probably more efficient as less teams are available for making different designs than intel has,and they are doing wonderful work and will continue to do that,don' you worry).
And i must ask you,why AMD needs saving?And from whom?
And ,who will save intel if Nehalem is less efficient than Cerebus?
Ah! those words! Well. First, bull******** is a casual word that's synonymous to "saying nonsense" in my culture. I know that it's hypocritically forbidden on syndicated TV stations accross the U.S. but that it's fairly common on rented DVDs (in the U.S. too).Quote:
Originally Posted by brentpresley
Second, it's not masterbate but masturbate and again, unless you're a moron (oops! typo. I meant mormon) you shouldn't be offended by the mere writting of the word let alone in a figure of speech.
amd definately doesnt need any saving. if it gets through this barrage of new intel architectures every two years and rapidly decreasing manufacturing processes, intel will be in a very tough spot. amd's goal is to increase its production capacities to intel levels at least for cpu production, and it will probably meet its expecations with three super fabs, fab 36, fab 38, and whatever theyll call the new york fab, and convert all of their lines to a modular architecture. thats why intel is trying so hard right now, as amd will very soon bring cpu upgradeability in terms of new features to gpu levels. of course, intel is not lagging behind, but it realizes that amd has a very effective long term plan.
With the price pressure intel is putting on AMD . AMD has to hit on all cylinders. No more misfires or they are in serious trouble.
Think of AMD position changes in the last 6 months. AMD went from a position of more cores are better to . More cores may not be so good . This is after AMD realized Intels full plan . 8 cores for nehalem 32cores for Geshner.
Geshner of course wouldn't be based on X86 plateform tho. But could run X86 code without a performance drop.
AMD has to have Ati to pull out of this one. I personnally hope fusion is a great success. Amd is going to rely on dual sockets and co processors to build on the future.
Intel is going multi 8+ cores and genesco to hold off AMD fusion in the short term.
Intel long term is aiming for 32+ cores all on 1 die for 1 socket. With memory on die. with differant cores to fulfill differant requirerments.
AMD needed ATI badly and that will reveal itself with the passage of time.
Not that we have that much time with global warming and the ice caps melting. You have heard of the Ice burg the size of texas in the atlantic have you not?
LOL, nice one. I guess you dont like the facts? Seriously, is it so damaging and demoralizing? there was a huge difference on K8 and P4, leaving the P4 in the dust. Even new P4s didn´t help that. Yet, its not allowed the other way in some form? Why do you make it to that AMD engineers are like 13 year olds? Perhaps you missed what AMDs main design focus is on. Its servers, not desktops.Quote:
Originally Posted by informal
What are you talking about? Last time I checked I stated BOTH wider data paths and bigger buffers. :fact:Quote:
Originally Posted by Shadowmage
Do you understand English at all?I said :Quote:
Originally Posted by Shintai
So my point was that you think those changes are nothing at all,just minor tweaks to the core,piece of cake,nothing important...If you think that it's easy to add something as OOO loads to a cpu without much redesigning,you need to read a lot of stuff first.It's VERY hard thing and that's the reason why designs(be it ibm,amd ,intel etc.) take so much time to come out from fabs.From paper/software simul. to a working MPU,it's a very long road.Quote:
Heck,by the look of your post,it seems like a 13 year old kid could have made a better job than AMD did,with such a small number of advancements.
The advancements are significant,and by how much we'll find out pretty soon,come Q2 and server variants are out.And remember,on server front,AMD is not the one stuck with old FSB and off die mem. controller.They jumped to next gen IMC(supporting many different types of memory standards) and a new improved version of Hyper Transport(that will be totally revamped with introduction of DC2.0 in 08')
jees ... by the sheer number of rabid offtopic intel fanboys in here *i'm looking for the intel section sign on the door*
why can't you little agro boys just STFU if you don't have something on-topic to add?
i guess its just pure ignorance or stupidity...
I think they are still trying to get those years of repression from Prescott out of their SystemQuote:
Originally Posted by flytek
My thoughts exactly.Quote:
Originally Posted by nn_step
Now why do you think that. AMD never had a 20-40% advantage over the hot running pressies. I for one am happly Intel built Netburst arch. And the Itanic CPU. How much of what Intel has learned threw mistakes and misfires will help intel In future chips. I would bet a lot.
I was reading how Nehalem will use HT again. I really had to laugh! I would very much like to see were Intel has said future HT would allow 8 threads to be read from 4 cores. Than of course Nehalem will scale to 8+ cores on a die. So that would be 16+ threads at once. Come on were did Intel say the future HT would do that? Link anyone.
Most are aware of Intel's Mitosis. It makes much more sense that Intels new improved HT will allow 1 thread to be used by all available cores. 2 threads by 2 cores for each thread. In Nehalems 8 core die This will be a huge success in multi core apps. and in single core apps it will be untouchable.
I am betting this is what Intels new improved HT will be. As anything else is just plain foolishness.
Quote:
Originally Posted by nn_step
I second that. The Amd uATX systems rock. I'm building two now with Mobile Semprons, one to be used as firewall and the other as a fileserver. Dead cheap and consumes only a few watts when undervolted (the Asrock boards are nice for this) and with a Picu Psu. It's even more fun to put them together than a main rig. Because it's sooo darn cheap! You can even use the stock cooler on ultra low rpm. :)
It should be good when we can buy dual core 65nm Semprons. Then we are talking nice and cool running budget rigs!
I completely agree Intel has learned a great deal from their Mistakes, AMD learned from Intel's mistakes as well. Esp after they started picking up some of Intel's Top designers from their Itanium and Yohan Projects. Though the MOST important thing they have learned is that Doing processing is Easy, Keeping the processor busy but not wasting work is hard. And that is mostly what is going to help AMD, putting more work into idle hardware. There are only a few cases today where AMD isn't close to Conroe, and they specifically are when throughput is more important than latency.Quote:
Originally Posted by Turtle 1
Again nn you got me. Link to were Intels Top designers from Itanic and Yonah.Left to join AMD. I believe A few of India's top designers left Intel for AMD after their massive failure with CSI . I know a couple of flunkies left Itanic for AMD but I believe they got the boot. As for your Yonah Enginneers leaving Intel isreal for AMD germany I would very much want to see that link.Quote:
Originally Posted by nn_step
I agree that getting more out of less is effient. But I would say that Most of us are happy with C2D and many who are not. But they don't count as they won't buy c2d reguardless. All intel really needs to do with C2D is add ondie memory controller and CSI and its A whole new ballgame. But I believe Intel is going for the Kill with Nehalem .
The name Nehalem is a Town I believe . But whats really interesting is the other spelling Neplilim (Children of the gods)
interesting use of wordsQuote:
Originally Posted by Turtle 1
http://news.com.com/2100-1006-6055595.html?tag=tbQuote:
director of Itanium circuits and technology
I wonder if that means he knows alot about Itanium's design :rolleyes:
If your trying to get the best bang for your buck its AMD. If you want performance and money is no object then its Intel. AMD's mobos are priced nearly half the cost of a decent Intel board. AMD's high end boards start at about $80 for most. Intels start at $150 with the same features as the $80 or $100 AMD's you got to put $50-70$ in more just for a decent OCer when most AMD boards will top out a cpu for $80 to $100 vs $150-$170. So a person with cost in mind is going to go with AMD. But if your all about performance and know your stuff you'll go Intel.
Some people still think Higher GHZ = Better so they will still go with Intel if they see a P4 or they might go with AMD over a C2D because AMD's GHZ is higher. To a normal person there thoughts are. More GHZ = faster again flashes in there minds. Some people don't know about C2D. And just the K8 and go or stay with that. Most users never come here or see any kind of forums or reviews for that matter. Some people don't even care.
Amen to that. I think it's safe to say that the people loyal to AMD will remain loyal even if it means keeping their s939 for a while. It's also safe to say that the (bad?) "salesmen" flooding "our" threads 1) won't sell us anything and 2) will remain Intel fanboys.Quote:
Originally Posted by Serge84
For some, there is no justification in paying a higher price for bragging rights and, for others, those same bragging rights are of the utmost importance.
exactly, upgrading the CPU isn't on anyone's top list. More Ram, Better GPU, and more storage usually come first and are often the most helpful.Quote:
Originally Posted by SkunK
So The Top HP designer left Intel after less than a year with Intel . Didn't take long for Intel to get rid of the loser. When he didn't deliver on time he got the boot . Big deal. Still waiting On those Israel engineers leaving Intel for AMD. Link please. While your looking check and see how many ATI designers left ATI/AMD for Intel.Quote:
Originally Posted by nn_step
Or he left a ship before it sank,and joined a new one,with better future...Quote:
Originally Posted by Turtle 1
Rest assured dear Turtle1,AMD is not going anywhere anytime soon.And they are ready for the battle,just as you think intel is (i don't doubt that either)
Shunk you got it all wrong. This thread was about K8L es chips In DEC . Never happened. Most OF the hype has been proven to be just that hype.Quote:
Originally Posted by SkunK
AMD says K8L is faster than AMD X2 by 40%. From the die shots this is a pure BS % unless their saying in certain apps. If thats the case C2D is faster by 40% in certain apps.
I could careless if you buy AMD or Intel. That would apply to every member of this forum . Except those with hidden agenda's.
But you guys are over hyping K8L . Its also strange thats its the same people who said Intel fixed the Spring IDF 06 results. I just find it amusing from an enthusist point of view that these people can be so broadminded when it comes to AMD and yet so narrow minded when it comes to Intel.
I have never every said AMD64 was not a good chip for its time. I like ondie memory controller. I believe C2D should have had it. But Intel wanted to have CSI inplace as well. So its wait and see. Just as K8L is wait and see.
But this hype of the K8L when we basicly know the changes that are being made. Is a joke. Than the AMD demo of the K8L running Task manager. And no one allowed to see inside the box. If AMD was going to show K8L . It should have been 1 core running real apps. and everyone allowed to see inside the box.
Sorry if it looks like a fish and it swimms like a fish and it smells like a fish its fishy.
How many knew what C2D was 1 year before it was released. or 6 months befor it was Shown . Very few. there was know hype except a few who had Dothan rigs that new it was going to be great.
I doubt you would see any Intel fanboys in these threads if you guys didn't over hype . Or say things like K8L will destroy C2D. . Its really pure BS. when you say K8L will destroy C2D when you all know K8L will be up against wolf/york. But you deny that. Just as you denied the spring idf 06 results.
Instead you say Intel won't have 45nm out in 3rd qt. But than you say AMD a much smaller weaker company will be out ahead of the roadmaps. But I guess if as a company you put out a bunch of road maps you can point to the right one and say right on schedule. I guess AMD has learned from the Inquirer. Put out enough fud some is about to be true. Example 4x4 was orginally hyped as an enthusist gamer setup. When AMD failed they changed it to multi thread Ultra. You guys bit right into it hook line and stinker.