Intel admits that AMD's quad core is generations ahead
Intel dudes only know about double dies.
posted by Sharikou, Ph. D. @ 6/08/2007 03:59:00 PM
Journal of Pervasive 64 bit Computing
Main Blog Page
Site Feed:
sharikou.blogspot.com/atom.xml
Blog Roll
Hot pages
sharikou.blogspot.com/atom.xml
Blog Roll
Hot pages
- Price War Strategic Analysis
- AMD and Intel Performance
- Intel's false performance claims
- AMD's Architecture Lead Persists
- AMD and Intel Performance
- AMD and Intel Architecture
- Bensley/Woodcrest and Athlon MP
- Bensley/Woodcrest is 2P only
- 64bit CORE Performance!
- Conroe Busted
- Capacity: AMD vs Intel
- Conroe's Impact
- AMD and Intel in China
- A Beautiful Picture!
- Temperature and System Reliability
- Google and AMD
- AMD's Grand Masters
- Intel's Predicament
- AMD and Dell
- NGMA Core and Colwell 6th Gen Arch
- AMD and Intel History
- The Tale of Intel and Anand
Analysis on IT trends and competitive strategies, with emphasis on micro processors, computer systems and networks. Based on latest news, backed up with real data, this site intends to provide a true and realtime picture of the fast changing IT landscape. This journal strives to be accurate on facts and sharp on criticisms. You may email your opinion to sharikou@yahoo.com or post comments here, be cool and intelligent.
About Me
- Name: Sharikou, Ph. D.
Freelance journalist on IT matters. Some of my writings have been published on online IT journals. Any original content on this journal is Copyrighted, but it's free for non-commercial use. Any Trademarks used on this site belong to their respective owners. Some of the pictures are links. If there is any issue with the content of this site, please email sharikou@yahoo.com .
Previous Posts
- DELL Dimension E521 is a great machine
- Core 2 Duo user experience
- K10 Motherboard with HT 3.0 PCIE2.0 ready to frag ...
- Real world experience with Intel CPUs
- Athlon 64 x2 4400+ (65watt) for $69 free shipping
- AMD FAB30 is being converted to 65nm
- Intel should be ashamed of itself
- Intel's front bus architecture in plain English
- Why is Intel QX6700 quad core a $125 CPU
- Rahul needs a philoshopy lesson
100 Comments:
Q: What are Intel's thoughts about the delay of AMD's native quad-core processor Barcelona?
A: AMD's so-called native quad-core processor has a difficult challenge in technical and manufacturing terms and even Intel would have difficulty facing such challenges. Intel currently still adopts two dies on one chip for its Harpertown processor. The technology is much easier, the product has higher yields and performance is almost the same as the native quad-core processor. Therefore, Intel will not launch native quad-core processors in 2007.
I couldn't find any more interesting information from there but perhaps its just me
Oh Intel Fanboys, even Evil-Cancer Man director person is admitting that Intel is already total BK.
Do not believe his lie in article for simple fact: This interview was translated from Chinese.
Since Burninator make for great LAnguage Englilsh we now change words from Chinese to Proper English type:
Lie Speak: Intel's product launch is directed by the demand of clients, which has nothing to do with the competition in the market. Currently, product development is almost complete therefore it is natural for the company to launch the Harpertown processor ahead of schedule.
True speak: Intel already BK, and has no real product. We try to make-steal Barcelona, but AMD has too superior security that prevent us from stealing it! I soon have job repairing bikes. Harpertown is a stupid Name for CPU, but not matter since it total Vapor unlike super-real-quad Barcelona
Lie Speak: Intel used to have CPU as the unit for server migration in the past, but is now using platforms. The current platform is expected to last until Penryn processor launches in the fourth quarter.
True Speak: Intel used to have CPU, but these were all stolen from superior AMD technolologies. Current platform will lock up unless running benchmarks that we cheated on. Penryn is total VaporWare unlike real Barcelno that has already made us BK. If I am nice will super-success AMD give me job?
Lie Speak: AMD's so-called native quad-core processor has a difficult challenge in technical and manufacturing terms and even Intel would have difficulty facing such challenges. Intel currently still adopts two dies on one chip for its Harpertown processor.
True Speak: AMD's super-luck native quad-core is at least 10x generation in front of the stupid Intel cancer-chips. AMD has solved all production problems that InTEL types cannot make for solution due to being too busy trying to get AMD jobs (but not passing IQ test). Intel cannot even make 2 CPU dies to glue together, all benchmarks are paid for on pumper sites
Lie Speak: HPC (high-performance computer), blade server and workstation sectors are all expected to have good performance. In the past, middle and small enterprises had the highest server adoption rates, but now demand for HPCs, blade servers and workstations is increasing due to demand from multimedia and Internet search engines and the Internet communication markets. These fields are expected to perform better than others.
True Speak: Many Computer use markets will make for growth because of superior Barcelona, INTel can only make optimize CPU for games that no-one but Arstechnica pumper fanboys play (and even they the freeze and make cancer). AMD already has 90% marketshare and will have 100% next Fliday.
See Intel pumpers, AMD already won! Do not believe some Intel Pezal fanboy that say they BK 2Q '08 ALREADY BK!!!! INTEL SPOKEDIRECTOR ALREADY ADMIT IT!!!
"Intel admits that AMD's quad core is generations ahead"
True. Intel expects AMD's quad-core to come out just in time for Nehalem - Intel's next generation.
True. Intel expects AMD's quad-core to come out just in time for Nehalem - Intel's next generation.
It's already been shown that Clovertown frags Barcelona. Penryn and Nehalem are so advanced that AMD will never catch up.
AMD BK Q2'08.
AMD BK Q2'08.
Problem with wishing for something is that it might just come true. Is this what you really want? Think again...
If AMD goes BK, Samsung might pick up the scraps...
If AMD goes BK, Samsung might pick up the scraps...
Nah, I'm thinking it would be more like Applied Materials.
You're the most pathetic fanboy ever. You'll post anything to make Intel look worse.
Hey, so would you still buy an AMD chip at this current moment?
We only know that Intel is over 1 year ahead regarding quad-cores.
Barcelona and the whole AMD are simply DOA/BK. You like Chapter 11?
We only know that Intel is over 1 year ahead regarding quad-cores.
If Barcelona releases on November/December 2007, you maybe right.
The Burninator, funny once but you have managed to push a joke way beyond it's sell by date ......
I don't agree when he says native quad core is more diffcult, if that was so then Nehalem will not only be difficult because of that but also because of CSI and IMC and it is unlikely Intel will try and do 3 difficult things at once.
I do agree with him though that yeilds may be better with twin packaged dual cores.
Another note here is Intel's PR lies. Mr. Boyd says in this interview that 45nm wasn't rushed to market in reaction to competition. Then they tell The Inquirer that
"he felt there may not even have been a need to rush out Penryns so fast, but, since the 45 nm newbie is pretty much ready, may as well let it roll
Which is it? Am I the only one tired of corporate double talk? This is not an AMD vs. Intel thing. It's a corporation feeding the public a shit sandwich thing.
AMD cores are generations ahead?
Yup they may be generations ahead in thinking, but too bad there is this minor detail they have to be here NOW to make a difference.
Generations ahead, yes, but also generations to wait before they get here.
Core2 was generations ahead too but when it arrived it crushed the oppositions, return INTEL to the top, captured marketshare and INTEL made billions.
I will wager any sum with any sucker that when Barcelona comes it may win a benchmark or two. It will not make AMD much money, nor will it materially improve AMD's MS and return them to above the mid 20's and that is the fundamental reason AMD is finished.
Who cares about generations ahead or what. Bottom line is do you make money and do you have a sustainable roadmap that allows you to be in business the next generation.
AMD doesn't... they are finished
Yo Phony.. did you catch the
"Phenom(enal) fiasco "
Nebo at the Inqueirer has an interesting read about the wonderful outcome for AMD at the big show. Looks like AMD is up to its expected behavior. Lots of talk, lots of anticipation, lots of sizzle but no meat.
AMD is finished...
http://www.theinquirer.net/default.aspx?article=40221
Read and weap.
This is what you get when you are a 2nd rate technology company, too stupid of an executive to focus your limited resources on unique market segements. Instead you run around beating your limp dick and thumbing your nose at the big dick and expect he with billions in the bank, 10's of thousand of designers 4 65nm factories, and a full couple year lead in technology to take it laying down. Sorry technology, business and war's arent' one nor do companies survive with that kind of strategy.
You are right AMD is generations ahead. They are so far ahead that they are finished since they don't seem to think it is important to take care of business "THIS" generation.
Bye bye AMD
some dick said...
It's already been shown that Clovertown frags Barcelona. Penryn and Nehalem are so advanced that AMD will never catch up.
Penryn has half of the latency in SIMD divide ops comparing to Conroe - thats its only design advantage on core level.
In comparison Barcelona has x2 FP throughput, x2 SIMD throughput, 50% more in ALU throughput, x2 more bandwidth to L1 cache, 32byte prefetch (vs 16byte on K8) and much much more.
When running optimized code in some server that thing will be scary.
Ycon said...
We only know that Intel is over 1 year ahead regarding quad-cores.
Barcelona and the whole AMD are simply DOA/BK. You like Chapter 11?
I have a question for (more advanced) intel fans - do you guys get pissed of when some intel-degenerate like shown above, lays of his mind on some blog and makes all of you look, well, shall we say 'not bright'?
More bad news for AMD: http://www.theinquirer.net/default.aspx?article=40225
HD 2400 is slower than 8400GS
AMD's mid-range and low-end DX10 GPUs will be slower than Nvidia's gear.
This must be AMD's new strategy. Wait until the competition has an awesome product out, then more than half a year later release a product that barely competes. A sure way to BK in Q2'08 indeed.
Barcelona DOA. AMD had pathetic 1.6Ghz only for Barcelona. Intel had 3.2Ghz/1600Mhz FSB/12MB L2 Cache Harpertown systems running at Computex. Clearly, Intel's silicon is in far better shape than AMD's. By the time AMD gets Barcelona sorted out they'll be out of cash and Intel will be ready with Nehalem.
AMD BK Q2'08.
LOL Another one (look one post above)
^_^
This comment has been removed by the author.
Azmount Aryl said...
LOL Another one (look one post above)
^_^
actually i'm laughing at you :)
unless you laugh at the host comments, else there is no merit for you to laugh at Giant's comment; he has been posting parody with some valid points compared to the host:)
azmout_aryl
"In comparison Barcelona has x2 FP throughput, x2 SIMD throughput, 50% more in ALU throughput, x2 more bandwidth to L1 cache, 32byte prefetch (vs 16byte on K8) and much much more."
You forgot to mention this is compared against K8 and all that you listed is already there in Conroe.
No. Conroe can't match K8 on FP performance, yet twice of it. Conroe has 128bit pathways to L1 cache not 256bit like in Barcelona. It also doesn't have 32byte prefetch.
azmount aryl
"No. Conroe can't match K8 on FP performance, yet twice of it"
If you are talking about x87 FPU then yes, you are correct. I you are talking about SIMD then you are wrong. I is not difficult to see 2x (or more) higher SIMD throughput on Core2 when comparing against similarly clocked K8.
Another thing Penryn has is fast shyffling of SIMD register contents and real SSE4.
As we all know, Clovertown is more than enough to frag Barcelona. That is, if AMD ever actually releases Barcelona.
VIA is more of a threat to Intel than AMD is now. Intel should keep a close eye on VIA.
AMD is no threat at all. They will lose more market share every quarter this year until they go BK.
AMD BK Q2'08.
What you guys are expecting is that Barcelona will be slower than K8, is that it ?!?!
Or you guys are expecting Barcelona to be some fancy technology, much like the new Intel Turbo Memory (Santa Rosa), which in the end didn’t do anything to improve the performance ?
Ho Ho said...
If you are talking about x87 FPU then yes, you are correct. I you are talking about SIMD then you are wrong. I is not difficult to see 2x (or more) higher SIMD throughput on Core2 when comparing against similarly clocked K8
My original post said: "Barcelona has x2 FP throughput, x2 SIMD throughput,", so as you can see I've separated FP from SIMD in my statement. The most embarrassment you can earn is by arguing things that only exist in your mind DUE to your bad english skills. Do read better next time, that is my advice. Or you can just make all intelers look stupid (I actually rather see you doing that instead)
Hey Sharikou, what happened to those 4Ghz AMD CPUs you told us about?
You said last year:
AMD's Rev F socket AM2 will be available for system builders on May 15, 2006. At 65nm, using Stress Memorization Technology co-developed with IBM, AMD will be able to increase clockspeed to 4GHZ. AMD is also working on Z-RAM, a SOI based technology that may increase cache density by 500%.
If AMD goes BK, Samsung might pick up the scraps...
They will not be able to. AMD's BK is a plan between Morgan Stanley and Intel and Nvidia. Morgan Stanley loans AMD $2.5bn knowing that AMD is going to BK in Q2'08. When AMD BKs they cannot pay back MS and MS gets the whole company. Morgan Stanley then sell off AMDs fabs to whoever is interested, and sell all the IP to Intel and Nvidia. From all the IP + fabs they will make back more than double their original investment. Morgan Stanley are ingenious for devising such a clever plot.
AMD BK Q2'08.
aguia
What you guys are expecting is that Barcelona will be slower than K8, is that it ?!?!
woo... someone is getting awfully pissed.
Hey AMD fanboy, aren't you also speculating that Santa Rosa Turbo Memory would fail?
We're not expecting Barcelona to fail. We're just expecting AMD to not be able to deliver them on time to stop Intel's Penryn (which is very likely ATM), and possibly Nehalem.
I'm sure Barcelona will be an excellent server chip, but whether Agena and Kuma will perform up to Penryn will remain to be seen. After all, AMD's main rooting segment is in the server area, where Intel pretty much reigned desktop and mobile sector.
Think, and think again, before you post some POS like that.
abinstein
"Barcelona has x2 FP throughput"
Ok, and you know this how exactly? Perhaps you can tell me the page number I should start reading from the AMD optimization manuals.
Also those two benchmarks seen so far sure do not indicate what you said. As you analyzed yourself it probably doesn't have too good performance in all cases, e.g in crunching the code that PovRay is using. Kind of like P4 that was excellent performer for as long as you didn't have any conditionals anywhere.
They will not be able to. AMD's BK is a plan between Morgan Stanley and Intel and Nvidia. Morgan Stanley loans AMD $2.5bn knowing that AMD is going to BK in Q2'08. When AMD BKs they cannot pay back MS and MS gets the whole company. Morgan Stanley then sell off AMDs fabs to whoever is interested, and sell all the IP to Intel and Nvidia. From all the IP + fabs they will make back more than double their original investment. Morgan Stanley are ingenious for devising such a clever plot.
The plan is awesome, except for one small problem: Department of Justice. Nvidia will not be able to purchase the pieces that are related to graphics, while Intel won't be able to purchase any pieces related to x86 or graphics. Samsung, on the other hand, might be able to purchase the x86-related pieces for 4-5 billion dollars, giving it an instant license to manufacture x86 CPUs. That is a bargain. Samsung has money to sink into Opteron, and wage out prolonged price/design war with Intel. No one else in the world has that might (not even IBM). ATI could be sold off to SIS or VIA or whoever.
I'm sure Barcelona will be an excellent server chip, but whether Agena and Kuma will perform up to Penryn will remain to be seen. After all, AMD's main rooting segment is in the server area, where Intel pretty much reigned desktop and mobile sector.
At this point, I wouldn't even place any bets on DP server performance. Intel has shown that Penryn is a beast when it comes to DP servers. And Barcelona showing so far is not exactly spectacular. So it is very much possible that Penryn might eat Barcelona for breakfast on DP. With Tigerton, even 4P won't be the undisputed territory of AMD. So Barcelona will the new undisputed king of 8P. However that is as relevant as being the undisputed king of gay gazzles!
http://www.anandtech.com
/tradeshows
/showdoc.aspx?i=3006&p=1
"In the end, performance was absolutely terrible. We're beginning to understand why AMD didn't let us test Barcelona last month. It's not that AMD is waiting to surprise Intel; it's that the platform just isn't ready"
"None of the partners we talked to are really impressed with Barcelona"
"Continuing on the worst case scenario track, some partners don't expect to see 2.3 - 2.4GHz until Q2 next year"
"For all of us who have been crying for Barcelona benchmarks, you really don't want to see them"
This sums it up. Scaling is terrible, performance is terrible, heck, it's not even ready yet. AMD expects to launch this crap next month at 1.4-1.6GHz? Intel doesn't even need Penryn, K10 has already been pre-fragged by Clovertown. By the time AMD scales K10 to 2.6GHz, Intel will have Nehalem out.
woo... someone is getting awfully pissed.
Not really. It seems that in every discussion the K10 will be even worse than K8.
Can be a presscott or willatte processor, but who knows.
Hey AMD fanboy, aren't you also speculating that Santa Rosa Turbo Memory would fail?
Do you think what is already a fact is speculation? Didn’t HP and Sony already given up? They could loss sales with this since Turbo memory isn’t just technology, it’s marketing too.
And don’t confuse me with you yomamafor2, I want what’s best for me.
We're not expecting Barcelona to fail. We're just expecting AMD to not be able to deliver them on time to stop Intel's Penryn (which is very likely ATM), and possibly Nehalem.
Who cares AMD and Intel have been lowering their prices. That’s a win for us (me). If K10 cost 500€ the cheaper one, I even care less.
I have already cheaper Core 2 Duo and Athlons X2.
Think, and think again, before you post some POS like that.
You should also think too.
I think the good doctor is forgetting one tiny detail... Intel is actually selling quad-core CPUs.
Please note the use of a present tense verb in the previous sentence. Thank you, drive through.
I think what Sharikou means is that both Intel and AMD are developing "native" quad core CPUs, but only Intel is actually selling quad core CPUs now.
"Who cares AMD and Intel have been lowering their prices. That’s a win for us (me). If K10 cost 500€ the cheaper one, I even care less."
This is the kind of thinking that most of the fanbois lack. Whether AMD can deliver Barcelona or not, it is because of Barcelona that people will get timely improvements from Intel with regard to Penryn, Nehalem, or whatever new/fancy technologies and marketing terms in them.
abinstein said...
“This is the kind of thinking that most of the fanbois lack. Whether AMD can deliver Barcelona or not, it is because of Barcelona that people will get timely improvements from Intel with regard to Penryn, Nehalem, or whatever new/fancy technologies and marketing terms in them.”
I strongly agree with abinstein and its not just in CPU’s; you know it must piss-off Intel to have to buy ATI graphic solution from AMD, but it will only make Intel work harder to improve their own solutions. Watts per clock is another example.
you know it must piss-off Intel to have to buy ATI graphic solution from AMD
Why would we have to do that? If people aren't interested in playing games or doing other things that need strong 3D performance an IGP is fine.
If they need high end graphics they can purchase an Nvidia card for the best speed and image quality possible.
From the low-end sub $100 Geforce 8500 to the massive 8800 Ultra, Nvidia has all the markets covered. All DX10 compatible. All Ati has for DX10 is one product at $400. It runs hotter, uses more power and is slower than Nvidia's competing product that is $50 -> $70 cheaper.
AMD BK Q2'08.
Most generous confession of an engineer who understands the feats that AMD have accomplished and probably wished that he was involved in such a enterprising company
“Randy Allen
Why would we have to do that? If people aren't interested in playing games or doing other things that need strong 3D performance an IGP is fine.”
Randy, you are correct. Ask Intel why they keep buying ATI stuff.
http://www.pcworld.com/article/id,132751-c,graphicschips/article.html
The point is, Intel will find a way to compete and do better.
So what if Dell/HP/etc want some ATI GPUs to sell with Intel systems. What does that prove? Intel is not buying them. If they didn't supply the OEMs with GPUs for Intel systems they would just go to Nvidia, losing Ati lots of business.
AMD BK Q2'08.
“Giant said...
So what if Dell/HP/etc want some ATI GPUs to sell with Intel systems. What does that prove? Intel is not buying them. If they didn't supply the OEMs with GPUs for Intel systems they would just go to Nvidia, losing Ati lots of business. “
It still probably irritates Intel that anyone would use ATI graphic solutions for there CPU’s. The head line in PCWorlds article is “AMD Will Supply Graphics Chips for Intel CPUs” which as you stated is misleading. I stand corrected giant.
Back to my main point; Intel can do better and they will with competition from AMD. Do you think Intel would be cycling to 45nm so quickly if it weren’t for AMD? This is win win.
This comment has been removed by the author.
http://www.tomshardware.com
/2007/06/12/vigor_force_recon_qx4
/index.html
AMD's 4x4 with dual FX-74 space heaters gets fragged in and out by a Dell XPS720 eqipped with a QX6800.
Games,
http://tinyurl.com/2myrs2
http://tinyurl.com/3xaszn
Fragged
Audio,
http://tinyurl.com/2r88pc
Fragged
Video,
http://tinyurl.com/2jukud
Fragged
Their conclusion,
"Vigor Gaming's Force Recon QX4 won a single benchmark, PC Mark 2005's hard drive test."
Hardly a relevent benchmark and it was a very close call.
Under every scenario, AMD's space heater gets fragged. Maybe AMD should go into the space heating business but you know, it's already summer so they would still lose money.
Jeez! Look at the conclusion:
We can't fault Vigor for building a system using inferior processors when these are the best an AMD enthusiast can buy. Instead, we'll put the blame on AMD for not keeping pace with the competition, and hope for a quick turnaround as new products are released.
[sarcasm]Hahahhahah those tests are not fair. When those test are done on a Numa aware OS that 54% difference in performance will go away.[/sarcasm]
AMD and their fanbois......pathetic.....
HD2600 and HD2400 scores are in:-
http://it-review.net/index.php?option=com_content&task=view&id=1419&Itemid=1&limit=1&limitstart=0
Pathetic. Pre-Fragged by Nvidia's existing products.
AMD BK Q2'08.
Intel is scheduled to drop its price on the Core_2 Duo July 22nd by as much as 50%. The K10 must be pretty good. AMD did it to Intel at the core_2 launch. What comes around goes around.
"The K10 must be pretty good"
Hkepc wrote about the price drop some time in the beginning of the year. That had to be planned a lot sooner than that. Bottom line: K10 has nothing to do with it, it is all a simple business strategy.
Were not a couple of the AMD fanboi's claming that AMD would fix the 2900 with driver updates?
Well looks like not.
The first round of driver updates are out and every major site reporting stuff like:
"We hoped newer driver revisions would improve performance on the ATI Radeon HD 2900 XT. With the newer driver we used for this evaluation we did not see any “magic” happen when it comes to real world gaming experiences at resolutions at and above 1600x1200. The ATI Radeon HD 2900 XT is not even a match for even the much less expensive and much less power hungry 320 MB GeForce 8800 GTS. "
Whoops...
2900 fails, Barcelona not looking all that good, Intel's line up looking stellar and at 45um...
I for one REALLY hope AMD doesn't die, as competition is good for everyone... but they keep f'n up like this the shareholders are going to go on a warpath!
R said...
Intel is scheduled to drop its price on the Core_2 Duo July 22nd by as much as 50%. The K10 must be pretty good. AMD did it to Intel at the core_2 launch. What comes around goes around.
Hey fanboi, a couple of important bullet points for you:
1.) the K10 isn't expected till September/October.
2.) Intel's line up starts to include the new 45um chips in August.
3.) Intel can afford to cut chip prices with ease. AMD cannot.
I think you better get your facts right before you start sounding more like Penix and Ph(ake)d.
Reply to evil_merlin
If I were a fanboi, I would have mentioned specific details of the article in BUSINESS WEEK yesterday that implies Intel may be going on the defensive with the price cuts.
http://www.businessweek.com/technology/content/jun2007/tc20070612_414317.htm?chan=top+news_top+news+index_businessweek+exclusives
I’m no longer an AMD share holder, so the chip war isn’t an issue for me. I agree with your bullet points except #1). We may see a launch as soon as August; which makes the July 22nd Intel price cut very timely.
We hoped newer driver revisions would improve performance on the ATI Radeon HD 2900 XT. With the newer driver we used for this evaluation we did not see any “magic” happen when it comes to real world gaming experiences at resolutions at and above 1600x1200. The ATI Radeon HD 2900 XT is not even a match for even the much less expensive and much less power hungry 320 MB GeForce 8800 GTS.
Apparently the 2900 is good for low resolutions with no AA.
First & foremost there isn't an OS available that can use all the power in todays CPU'S. So why are we bickering about 45/65 nano or Intel vs AMD? It doesn't mean crap. Hardware is far ahead of software at this point in time.
"Only AMD wrote...
First & foremost there isn't an OS available that can use all the power in todays CPU'S. So why are we bickering about 45/65 nano or Intel vs AMD? It doesn't mean crap. Hardware is far ahead of software at this point in time."
so now your playing Scientias favorit card ? :p
Give me a CPU with 1000 cores at 10GHz and 5 times higher IPC together with enough memory bandwidth and I can use that horsepower quite easily.
As a matter of fact such a CPU wouldn't even reach top5 in top500 list.
OS uses only a fraction of your CPU power. In fact I could take a P3*, downclock it to 100MHz and pair it with GF2 MX and run Compiz on it without problems and it would be difficult to see the speed difference between a PC with several times more CPU power. In short OS doesn't care much about your CPU, it is the apps that do.
*) P3 because SSE is a minimum requirement for compiz to work. I wouldn't call Vista an especially efficient OS, Linux can do all that and much more with several times less recources that it can.
If anything then I'd say we are only scraping the surface of things yet to come. One obvious future direction would be a revolution in rendering when ray tracing enters the scene, perhaps even completely replaces rasterizing just by being much better scaling and recource efficient. In 2006 Intel promised to show first fruits of their ray tracign recearch in 3 years with first real products out in five. Until then research on ray tracign progresses faster than ever before with new discoveries done very often. It would be an understatement to say that ray tracing has become around 10-100x more efficient during last 5-6 years. In terms of building data structures the difference is even much bigger. What I'm talking about is algorithmic improvements, not HW enchanchments. Same kind of performance increase took place with rasterizing around 20 years ago and during the last 5-10 years only minor things have been discovered.
I’m no longer an AMD share holder, so the chip war isn’t an issue for me. I agree with your bullet points except #1). We may see a launch as soon as August; which makes the July 22nd Intel price cut very timely.
You've seen it all over the web, K10 isn't scaling well, has terrible performance and isn't even ready. From all the evidence shown by Anand, Tech Report etc @ Computex 2007, AMD is going to need a few more revisions to the chip before it's launch ready. Some motherboard partners even predicted that AMD won't be able to scale K10 to 2.4-2.6GHz until Q2'08(I want to add this is when Intel will go BK according to our good doctor and his horde of AMD fanpois). I highly doubt AMD will embarrass themselves by launching a half dead K10 at the projected clock speeds of 1.6-1.8GHz. AMD would be better to delay the launch but shipping a half dead CPU and delaying will both have a negative impact on AMD.
First & foremost there isn't an OS available that can use all the power in todays CPU'S. So why are we bickering about 45/65 nano or Intel vs AMD? It doesn't mean crap. Hardware is far ahead of software at this point in time.
That's a good one! Did you say that when AMD had the performance lead?
Intel's price cuts are intended to finish off AMD and that is what will happen. AMD's fastest CPU will be competing with a $160 Intel 2.33Ghz CPU. No one will buy ancient AMD 90nm space heater garbage.
AMD BK Q2'08.
I think we had better wait and see what speed the K10 actually shows up at before writing the epitaph already.
As a consumer I'd much rather have a closely fought competition than one manufacturer way ahead or not even there at all.
The ideal for me would be for K10 to be slightly ahead and Intel having to keep prices low to compete, then I can make a choice.
To Poke 5:31 PM, June 13, 2007 post
Yep, I’ve heard. I discount most of the nay-sayers. Some of motherboard partners are complaining, but not all. Scientia puts some positive spin on the launch. So, what is the truth? I base my humble opinion on Intel’s price cuts. Intel seems to be doing everything right for some time now. Intel knows how to make a profit even during impossible times, so why wouldn’t they wait on the price cuts a few months and make the extra margin on the quad until September.
If K10 is released half-baked, AMD would have a home made disaster as you pointed out. Emotionally I’m hoping AMD can pull a rabbit out the hat, realistically the odds seem to be in your scenarios favor.
The up-side? Inexpensive Core_2 Dual ($250) is going to be a fantastic deal.
The down-side? Sorry, I can’t think of a down side.
Ouch. AMD is really in a bind.
http://www.theinquirer.net/default.aspx?article=40336
In a nutshell, Opteron held over 13.3% of the dual socket workstations in Q2 of 2006. By Q1 of 2007 that number has fallen to 8%.
AMD cannot get a break as of late...
r
"I base my humble opinion on Intel’s price cuts"
Please explain why did Intel lower its prices in April.
"why wouldn’t they wait on the price cuts a few months and make the extra margin on the quad until September"
Just a random idea but perhaps it plans to introduce new CPUs to market? You know, 3GHz quads with decent TDP, FSB1333 and so on. Sure, their dualcores will be dirt cheap but who cares anyway? Or did you expect Intel wouldn't lower the prices of its CPUs when introducing new models?
Forgot to add the question of why does Intel drop its desktop CPU prices when desktop versions of K10 are not due to be released before Q4 or even later?
It's a brilliant move by Intel to price the Q6600 at $266. Now, at the current price of $530, Intel will make a lot of profit, but they aren't selling a lot of quad core CPUs to consumers this way.
They will make less profit on each processor with the lower price, but they will sell many times more CPUs at the reduced price, expanding the market for quad core CPUs and earning additional profit.
Yes brilliant indeed.
The key is FLOOD the market with your chips.
Intel is selling a lot more than just the cpu. With every cpu sold, Intel can also make a chipset sale, a motherboard sale.
All those big vendors will offer QUADCORE for less, it'll appeal to the market. They'll order the quads along with their mobos...
ho ho said...OS uses only a fraction of your CPU power. In fact I could take a P3*, downclock it to 100MHz and pair it with GF2 MX and run Compiz on it without problems and it would be difficult to see the speed difference between a PC with several times more CPU power. In short OS doesn't care much about your CPU, it is the apps that do.
Well excuse me!!! I will include applications/programs to say they cant use the power of todays CPU's..simple! They are not written for this type of use. Sorry! Prove to me they are.....
only amd
"I will include applications/programs to say they cant use the power of todays CPU's..simple!"
Now you lost me. Are you claiming CPUs have too much power that apps can't use or do you mean that underclocked P3 is not sufficient for running stuff?
Btw, remember you were only talking about OS, not applications before ;)
Well excuse me!!! I will include applications/programs to say they cant use the power of todays CPU's..simple! They are not written for this type of use. Sorry! Prove to me they are.....
I do a lot of video encoding and audio encoding. This takes 100% advantage of my Q6600 overclocked to 3Ghz.w
This comment has been removed by the author.
LOL, if you only knew what the panic is like in Dresden, Sunnyvale and East Fishkill.
We've seen the QX6800 frag the Quad FX. Yes, that's right: One Intel CPU frags two of AMD's fastest CPUs!
HD2900, 2600 and 2400 are all failures and were pre-fragged by Nvidia's Geforce 8.
Barcelona is too little too late, it has been pre-fragged by Clovertown. Pathetic. Intel has a one year lead on quad core CPUs.
Penryn coming this year, 40% faster than Clovertown. Fragged all over.
AMD's market share is crashing, losing 6% of market share last quarter. At this rate AMD will have zero market share by the end of the year.
AMD is finished.
AMD BK Q2'08.
oneexpert
"According to bloomberg news intel plans to cut there cpu prices to the bone because they fear the barcelona will cause there cpu sales to drop to zero"
Why does Intel cut desktop chip prices when desktop versions of Barcelona are still several months away?
"Intels platforms are antiques and so are there cpus."
Funny that all it takes is an antique to perform better than AMD's most advanced and fastest cretion.
"According to the bloomberg news intels profitability does not exsist since amd put the price cuts on them."
Any links to this? I don't expect Intels >1b$ profitability to drop this quarter nor the next. I do think though that things will be slightly worse for AMD in the next two quarters, that means continued >$0.5B in red.
"Intel has huge expenses while AMD is still fit and trim."
I guess you've missed the last two quarter revenue-profit numbers.
"AMD/ATI are releasing a huge volume of new products this year but Intel is not."
AMD itself has said that there won't be volume K10 this year and it will start picking up some time next year. It will probably be the same for Intel 45nm.
"Intels press releases are not products."
Neither are powerpoint slides and simulated CPUs.
"AMDs products are actual state of the art hardware."
There is not much use of that when it isn't enough to be faster than antiques.
"It should be a very interesting year for cpus, motherboard chipsets, and video solutions."
That I can agree with. Also I expect things to get even more interesting in the coming years.
AMD Barcelona yields are terrible:-
http://www.fudzilla.com/index.php?option=com_content&task=view&id=1484&Itemid=1
oneexpert, you need to stop living in this made up dreamworld Ph(ake)d tried so hard to make for you fanbois.
1.) Intel's "antiques" outperform AMD's latest and greatest, in some areas by huge margins.
2.) AMD fanboi's seem to forget that the next generation of Intel CPU's is coming up quite quickly.
3.) AMD fanboi's seem to forget that currently Barcelona isn't looking so good. Most people that have tested it and worked with it are complaining its not fast enough and its simply not ready yet.
4.) We won't even get into the 2900/2x00 debacle AMD is currently running thru.
5.) Intel can sell a C2D for 100 bucks and still be making money on the processor. It's estimated that it only costs Intel about $35-50 to make a middle bin C2D.
6.) Yes Intel is large. Yes they have expenses, but they are still making a PROFIT. Something AMD isn't... AMD can be lean and clean all they want, but they are still losing money.
You really need to get in touch with reality...
i think someone just hanged himself. It's either life has become unbearable due to AMD's blunders or it could be the shame from the awful Intel BK mis-predictions.
It gets even worse for AMD. The Intel V8 platform benches are out and it is simply crushing 2 QuadFX chips. Not to mention it shows under a vast majority of benchmarks, QX6800 crushes 2 quadFX's as well...
Of course the AMD fanbois will just say that this site is being paid off by Intel...
http://www.hothardware.com/Articles/Intel_V8_Media_Creation_Platform__Dual_Sockets__Dual_Xeons/
Things look real bad for AMD now. Intel is bringing Core 2 performance for as low as $84:-
http://www.techspot.com/review/53-pentium-e2140-e2160/page6.html
The new Pentium E2100 processors offer exceptional performance at an amazingly low price point. Furthermore, these new processors have inherited many of the characteristics that helped make the Core 2 so impressive in everybody's eyes (everybody but AMD's, we assume anyway).
But cutting the smart-cache down to 1MB these things are incredibly cheap to produce. Cutting the cache to 1MB from 2MB in the E4xx series did not drop a lot of performance. The biggest increases from the cache are in video encoding. You can gain up to 10% there with 4MB of smart-cache.
AMD is going to burdened with a lot more inventory of it's cheap CPUs. AMD will have to halve it's processor prices if it wants to compete.
AMD BK Q2'08.
This comment has been removed by the author.
Where as AMD/ATI are releasing one product after another and on schedule.
R600? In excess of seven months late. K10? It's "mid 2007" right now. Delayed until Q4'07. Pathetic.
AMD is losing upwards of $600m a quarter and is losing market share 6% a quarter. By the end of Q4'07 AMD will have 1% market share left.
Intel's stock is at $24.24,a new 52-week high for Intel with a value in excess of $139bn!
AMD's stock is at $13.63, only one dollar away from their 52week low of $12.60. Market cap? Just $7.5bn.
In other words, the investors know Intel is in a strong position gaining market share every day while killing off AMD in the process.
In other words, Intel is worth more than 18.5X more than AMD.
Intel reports strong profits exceeding $1.5bn each quarter. That's right! Intel makes more in profit than AMD posts in revenue!
C2D owns all the benchmarks and the performance benefits are very real. My E6600 system is over 40% faster than my old AMD 4200+ (slightly overclocked to 2.4Ghz for even Mhz comparison) system in video encoding. With the Q6600 that's another 70 -> 75% faster. Well over twice as fast as the AMD 4200+ at the same frequency of 2.4Ghz.
A c2d cannot run windows me.
Tell me why anyone would want to run a dual core or quad core CPU on an OS with no support for SMP??? Idiot. For that matter you couldn't run Windows ME on an Athlon 64 X2 system either.
All R600 based cards pre-fragged by Nvidia's Geforce 8 series. Barcelona pre-fragged by Clovertown and Penryn. Pathetic.
This must be AMD's new strategy: Wait at least seven months for the competition to have an excellent part in the market then AMD will release it's slower performing part!
AMD BK Q2'08
Intel is an "A" company, AMD is a "B" company.
Nvidia is an "A company", ATI was a "B" compnay.
You know what happens when a "B" company buys a "B" compnay? You get an "F" compnay.
While both Intel and Nvidia hit 52-week highs, AMD is steadily sliding back to 52-week low; only a dollar to go.
@onexpert
"At least AMD portables did not burst into flames."
You do know that ALL AMD cpu's blow up ALL kinds motherboards?
/sarcasm off...
When are you going to post K10 benchmarks? Or did you ditch this blog after you saw the benchmarks?
Things are not looking good for AMD. Even an old Pentium D frags Athlon 64 X2 in budget CPUs:
http://www.xbitlabs.com/images/cpu/pentium-e2160/divx.png
AMD BK Q2'08.
This comment has been removed by the author.
oneexpert said...
I would agree with you fanboys that intel c2ds own all the benchmarks.
Unfortunately nobody runs benchmarks on there computers, they run open office, maybe windows(not windows me on a c2d), linux, games and movies.
And as I have pointed out many times, c2ds have problems running typical software packages, they lock up or simply cant run standard software commands without errors.
So Oblivion, Company of Heroes, FEAR, Stalker, Word, Excel, Photoshop etc etc count as bechmarks and not typical software.
I can also see the internet straining under the massive amounts of complaints from people experiencing the same problems as you running 'standard' software on C2D.
Other than that I just like to say thanks for such a quality post :)
Why is it that only the AMD fanboi clownshoes seem to have problems with "C2D locking up" and "have had so much trouble with a e6600 and e6400 misrunning basic programs".
What a fucktard.
I wonder what happened to the good doctor?
This comment has been removed by the author.
My friend has an AMD 4400+ system. He was trying to encode some video and audio, the system just locked up. It was obvious that AMD's ancient 64bit SSE and only the pathetic 3 issues wide core were to blame. The system just couldn't handle it.
So we did it in on my E6600 system and it just zipped right through. 128bit SSE and the ability to process 4 insturctions per cycle make short work of this workload.
Everyone knows that C2D is for real work. AMD CPUs are a joke, toy CPUs. Pathetic.
AMD BK Q2'08.
AMD cpus do random events very, very, very, well.
This is understandable, considering that the most commonly used source of randomness is thermal noise. Core 2 runs cooler, and hence probably gives lower entropy to the random number generator (RNG). AMD space heaters have no shortage of entropy.
That is the only explanation I can come up with...
This is understandable, considering that the most commonly used source of randomness is thermal noise. Core 2 runs cooler, and hence probably gives lower entropy to the random number generator (RNG). AMD space heaters have no shortage of entropy.
In not-so-rare case, if some AMD fanboy misunderstands this comment, please note that this comment is purely sarcastic. There is absolutely no reason (aside from bugs) for one CPU to handle randomness better than the other. And looking at the errata for both the CPUs, AMD CPUs seem to be more prone to errors than Intel CPUs.
AMD CPU's give you a true gaming experience?
What dope are you smoking kid?
Most gaming frame rates are consistently better on C2D than any AMD CPU.
Take AMD's best CPU and best video card against Intels best CPU with Nvidia's best video card and AMD looks like that poor retarded kid at the prom.
oneexpert
"Randomness is very important to computers and computer programs especially for games."
Yes, it is but there is no x86 function that gives you random numbers. Generating random numbers is just a simple math. Of course on some OSes there are a bit different solutions available also. For example in Linux you have /dev/random that uses data gathered from all sorts of sources and generates true random bits.
In short there are two ways of getting random numbers:
1) use a simple math algorithm that generates a certain sequence of pseudorandom numbers from given seed
2) use things like /dev/random that generates real random bits
In games nobody uses 2 as it is very slow and doesn't have the predicrability of method 1. In games you need the data to be random enough but repeatable. To get that you'll need to know the algorithm and seed value.
So basically if you claim that Intel CPUs are bad at random nubmer generation then you claim that it has some huge problems with simple integer math. Of cource you are wrong with that, no such problems are known. If you don't believe me just do some tests with rand() and tell me about the results you saw on Intel and AMD CPU.
I stand corrected on my post about the July K10 launch possibilities. Intel fans were correct in the rumors of a Sept – Oct volume launch. Its all over the net, no links are needed. I was writing checks with my heart that reality can’t cash. Sorry.
poke
I wonder what happened to the good doctor?
He is probably currently hiding under the bed, tears dripping down his cheek, murmuring.. "my stock.. my stock..."
This comment has been removed by the author.
oneexpert
Do the Doctors know you escaped?
oneexpert said...
...Best gaming and power saving chips AMD.
Highest power sucking chips INTEL...
yeah, those were the days...
oneexpert
"I have done extensive tests on random number generation with cpus."
Please describe those tests a bit more. Example code would be excellent but for starters I'd be happy just to know what did you use to generate random numbers, some OS facilities like /dev/random, simple libc rand() or hand coded merseinne twister? You know without any kind of explanation your text suits nicely together with all the other FUD being spread.
"I was reading toms hardware report on the new x2 2350 and have reviewed the test results. I see the $1400 ex6800 c2d was fragged by the $200 am2-6000 and the $80 x2-2350."
Care to elaborate on what benchmark that was? I bet it was some weird thing that even didn't use quadcore features. Comparing quadcore vs dualcore in such a benchmark would be the same to compare how fast can you drive 1/4 mile by motorbike and 20 ton trailer.
"Like I said before the c2d does not run regular software well"
And that software is ... ?
"c2d pdf creation time scores are very poor."
Funny that you said that, I wrote an application that used iText to generate PDFs and it ran extremely fast on Core2. What libraries or programs did you use?
enumae
"Do the Doctors know you escaped?"
Perhaps Sharikou, our beloved doctor, is taking care of him?
I just found the benchmark you were referring to here. Some comments on it.
1) That weird EX6800 seems to be regular dualcore X6800, not quadcore as I assumed. THG seems to be making up CPU names these days.
2) PDF generation was not CPU bound as differences between 2.83 and 1.8GHz Core2 and 2.1GHz and 3GHz K8 were negible. Bottleneck had to be somewhere else.
Another interesting thing was that 1M L2 cache E2160 was beating AMD CPUs in most of the benchmarks. Only 6000+ was staying ahead of it.
Post a Comment
<< Home