Sign in with
Sign up | Sign in
Your question

AMD has once again taken the technological lead ???

Tags:
Last response: in CPUs
Share
a b à CPUs
a b À AMD
May 30, 2008 5:49:25 PM

Hi guys,

I found this on Yahoo's finance site:
http://biz.yahoo.com/zacks/080527/12903.html?.v=1

Quote:
Zacks.com
Wait & See with AMD Shares
Tuesday May 27, 3:43 pm ET
By Ken Nagy, CFA

Advanced Micro Devices (NYSE: AMD - News) is the second largest producer of microprocessors and chipsets in the world. March quarter revenue missed the consensus by a sliver, although the EPS exceeded handsomely. Forward guidance is for revenue to be down in-line with seasonality (we estimate down around 4-5% sequentially). The long-awaited Phenom and Barcelona are now available, and should help market share gains in 2008.

Additionally, margin improvements shall continue as new products gain momentum and 45nm production ramps up. We think that despite the slow start, 2008 is likely to be a good year for AMD. We remain bullish about this stock, but prefer to take a wait-and-see approach for the time being.

We think we will see the advantages of this strategy unfold over the next two years. As AMD churns out faster and more efficient products at the low-to-mid-range, a large percentage of users with smaller functional needs are likely to switch to the lower-priced product. Further, with the launch of the latest processors, AMD has once again taken the technological lead.

However, the debt level remains very high, and execution remains an issue. We are, therefore, reiterating our Hold rating and $11.00 price target, which corresponds to a 1.1x multiple of our 2008 sales estimate.

Sejuti Banerjea contributed to this report.


The red highlight is my own, not the author's. Does anybody have any idea what he's talking about?
Thanks!

More about : amd technological lead

May 30, 2008 5:59:17 PM

Could be referring to the fact that the phenoms and barcelona are true quad core compared to the core 2 quads.
May 30, 2008 6:11:26 PM

Quote:
We remain bullish about this stock, but prefer to take a wait-and-see approach for the time being.
:heink:  Nothing like contradicting yourself in the same sentence.
Quote:
However, the debt level remains very high, and execution remains an issue. We are, therefore, reiterating our Hold rating
High debt level and questions about execution, and the author is bullish on the stock? :heink:  Glad he's not managing my portfolio.
aevm said:
Does anybody have any idea what he's talking about?
Thanks!
I don't think even he knows what he's talking about.
Related resources
May 30, 2008 6:15:18 PM

doomturkey said:
Could be referring to the fact that the phenoms and barcelona are true quad core compared to the core 2 quads.
...this translating into what real-world benefit?
May 30, 2008 6:16:10 PM

Look this topic has been discussed to death about a hundred times on these forums. When are you guys going to give it a rest? AMD doesn't suck and they don't own either, get that into your thick skulls. (Preemptive posting)
May 30, 2008 6:20:44 PM

modtech said:
Look this topic has been discussed to death about a hundred times on these forums. When are you guys going to give it a rest? AMD doesn't suck and they don't own either, get that into your thick skulls. (Preemptive posting)
Their product may sucketh not, but their execution sucketh mightily.
May 30, 2008 6:33:03 PM

^Agreed!
May 30, 2008 6:50:40 PM

I believe this is due to AMD being first to release quad cores using 65 watts, this will make a big difference to data centers where power efficiency and heat are of great concern. Remember also that server CPU's generate much more profit for Intel and AMD than desktop processors and this would be of interest to investors.
May 30, 2008 7:02:47 PM

Getting back to Aevm's original question, having been in stocks for over 40 years, I think I can venture an explanation, not withstanding that the writer of the article should have been a bit more clear in what he tried to say, and I may be getting his intention wrong regardless.

Yes, the writer may be bullish on the stock, but he's thinking long term and not short term when he says he wants to take a "wait and see approach". In other words, he thinks AMD will go up in price eventually, but don't buy quite yet. Part of the reason is that AMD does not pay a dividend, so any money invested right now is sitting there wasting time and not doing anything profitable. For an investor, this is a chancy thing. The time to buy is always at or about its lowest. The good side of chance exists that AMD might come out with something good and the stock will take off. This is especially true when considering that AMD produces more than CPUs, but also video cards (ATI) and chipsets. The bad side of chance is that either the current recession will hold back AMD not matter how good a product it produces, or that it will be a long time before AMD turns out another really good moneymaking product.

As for the technology lead, that might be referring to the fact that AMD has managed to produce a quad core chip that (finally) works, whereas Intel is still using two glued together chips in its quads. On a bright side, all AMD needs to do is to figure out how to raise the performance of the chip, not design one from scratch. As Strangestranger pointed out, this is a difference between technology and performance. AND needs to improve its performance of its quad core, while at the moment, Intel needs to design and sell a monolithic quad core chip, as well as making it perform. On the cloudy side, the average customer at Best Buy doesn't know the differences between monolithic quad core and glued quad core. He only knows that Intel's quad runs faster. As enthusiasts, we know the differences, but our purchases are a drop in the bucket compared to total sales.

Hopes this helps, Aevm.
a b à CPUs
a b À AMD
May 30, 2008 7:11:02 PM

Thanks Sailer. Yes, I think that's it.

Spongebob, I love your signature :) 
May 30, 2008 7:11:02 PM

sailer said:
the average customer at Best Buy doesn't know the differences between monolithic quad core and glued quad core.
And why should they care even if they knew? WHat real-world benefit does a monolithic quad buy the end user? For AMD's part, it seems to have been more of a liability up to this point.
May 30, 2008 7:25:06 PM

Any news article that ends with "OMG AMD4LIFE!!1!" is obviously garbage.

Technological lead, lawl. Just because something is more complicated doesn't mean it is better technology.

Results matter, how you get there does not.
May 30, 2008 7:32:19 PM

TechnologyCoordinator said:
Just because something is more complicated doesn't mean it is better technology.
Well said!
May 30, 2008 7:35:40 PM

borumas said:
I believe this is due to AMD being first to release quad cores using 65 watts, this will make a big difference to data centers where power efficiency and heat are of great concern. Remember also that server CPU's generate much more profit for Intel and AMD than desktop processors and this would be of interest to investors.



Yup. When I read that the 65W Quad Cores came to mind. I think that is exactly what he is talking about.
May 30, 2008 8:00:42 PM

spongebob said:
And why should they care even if they knew? WHat real-world benefit does a monolithic quad buy the end user? For AMD's part, it seems to have been more of a liability up to this point.


The answer was part of what I was writing. At this point, AMD's monolithic quad doesn't matter, as it doesn't have the performance. But, if the performance (speed) can be raised, it will matter. The monolithic quad in and of itself has not been a liability. A defect in the original chips occurred, but that has been solved. Now the next step is to raise the speed. Real world benefits may involve less power usage, better integration of the chip components, and eventually a faster chip. If and when that occurs, the Best Buy customer will then see two machines with similar speeds and will turn his attention back to that of which machine is cheapest.

By the way, I'm talking about technological theory and not meaning to say that AMD's present chips are better than Intel's chips, nor am I an AMD fanboy. My present gaming computer uses an Intel chip and I'm happy with it.

Edit: I just saw an article which further helps the point of AMD and its technology.

http://www.tomshardware.com/news/PC-supercomputer,5513....

The research group Vision Lab at the University of Antwerp came up with a different solution and constructed a PC that integrates four GeForce 9800 GX2 graphic cards (with a total of eight GPU cores) that runs CUDA-optimized tomography applications. The specifications include a MSI K9A2 Platinum motherboard, an AMD Phenom 9850 CPU, 4 x 2 GB Corsair TWINX DDR2 PC6400 memory, a Samsung Spinpoint F1 750 GB hard drive, a Thermaltake Toughpower 1500W Modular power supply unit as well as four MSI 9800GX2 cards. The researchers said that the cost of the system was less than 4000 Euro or about $5300.

Note the use of an AMD 9850 quad core chip. Not implying that an Intel chip couldn't have been used or might have preformed better. As it stands, people, especially those in business, will see the use of an AMD quad core and think that if its good enough to do that type of work, it will be good enough for their use. You might object and say that business usage and not enthusiast, but the vast bulk of AMD' income is from business purchases, as it is with Intel's income. Still, this shows that AMD's chips are technologically capable of doing very complicated work.
May 30, 2008 10:07:34 PM

AMD's quad is native...far more advanced full stop. Phenoms full performance has not been utilized properly yet....the better technology will pay off and the superiority over Core2Cheeseburger will be realized. Nehalem is a K10 clone...Intel know they will be left behind in the near future if they don't release a native Quad. Intel has less experience with on board memory controllers..there's a potential for bugs and that could be the reason Intel is limiting Overcloking because of this? watch this space.

Phenom 9850 Unlocked multiplier! Phenom 9850 enthusiasts choice! Intel are jealous of AMD!

AMD4Life!
May 30, 2008 11:04:23 PM

I see that Thunderman has joined in to add some comments from the asylum.

I think its more like "there's a potential for bugs and that could be the reason AMD is limiting overclocking"
May 31, 2008 2:57:08 AM

thunderman said:
AMD's quad is native...far more advanced full stop. Phenoms full performance has not been utilized properly yet....the better technology will pay off and the superiority over Core2Cheeseburger will be realized. Nehalem is a K10 clone...Intel know they will be left behind in the near future if they don't release a native Quad. Intel has less experience with on board memory controllers..there's a potential for bugs and that could be the reason Intel is limiting Overcloking because of this? watch this space.

Phenom 9850 Unlocked multiplier! Phenom 9850 enthusiasts choice! Intel are jealous of AMD!

AMD4Life!



LMFAO. Dude, you sound just like "The Burninator" from Sharikou's blog comments.



Quote:

The Burninator said...
(7:40 PM, September 08, 2007)

EVILness of Intel Pumpers!! The AMD time of make-super-fastness in K10 is now upons!

Dr. Hector is good Because of his True Ameericanness. Sharikous hass shown that none of the peoples not of the Americanness can be making of any chips!!! Hectors is 100% always Americanss and has always spoken in most exquisiteness of perfections in the English! Therefore All AMD chips are of the superior form!! He has never used another non-English Language because of Sharikou showing that the non-Americans are all Paid Intel PUmpers who make evil cancer Pentium 3s!!!

Here are the true facts since K10 is the most widely of benchmarked chips EVER:
a. It is PROVEN to be 1000x faster in the only true benchmark which is the AMD SpecFP. Oh you Intel pumpers talk about lies with your "office" and "games" and "movies" and "databases".
THESE ARE LIES MADE BY THE INTEL!!
Only a slide from the pure AMD is of truthality! Only Randy Allen can be said to be of the truth... any liar "website" is only an evil racist Intel lie from not in Amerrica!!!

-4.3: It is shown 100% that any support of Intel is evil racism. Hector is known to be from the Mexico, and whenever any evil Intel pumper uses an Intel chip it means that evil Intel is being the racism against all of Mexico which is against the 34th amendment of all Consitution. Only evil wrong cancer-monopolies can use Intel and they will all be in the JAIL soon!! Any pro-Intel is always the racist!

r: K10 is super superior because of its super-cache that is NOT SHARED!! each super-core is having the own cache. Only evil Intel chips are using this primitive 1970's lie!

5: K10 is superior because of it's advanced shared Level of 3 cache! Only some stupid evil Inteller would ever think to not use shared cache! AMD invented 100% of all ideas ever used in CPUs with the K10. ALL INTEL CHIPS ARE ONLY RIP OFFING THE K10 SUPERIORNESS!!! THE P3 IS A LIE THAT WAS STOLEN FROM THE K10!!!

IIIV. The K10 Is so of the advancement that Intel must be made to be illegal. It is against the AMEEERICAN free-speach rights to allow any Intels to ever talk or use their evil chips because of the Intel Racism. Only AMD can ever be used for the thinking! Only AMD can use the free speech because of HEctor being a pure Ameeerican!! All Intellers must be put in to cages to keep good PURE AMD!!!

THe Time of Acting is HERE!!!! K10 is super-100% 100 YEARS EARLY THAN EXPECTED BECAUSE OF SUPER AMD!!!
a b à CPUs
a b À AMD
May 31, 2008 3:47:38 AM

Does the American Constitution even have a 34th Amendment ?

Thanks for the quote TC, that was a good laugh :) 
May 31, 2008 4:37:58 AM

lmfao hahahahahha Burninator was hillarious :p 
May 31, 2008 5:02:09 AM

omg....THUNDERMAN is back...! :D 
a c 96 à CPUs
a b À AMD
May 31, 2008 12:07:22 PM

I don't really know what a ""technological lead"" is supposed to mean but utilyzing split power planes/voltages over multiple synchronous data buses with varibly clocked multiple cpu/gpu cores seems pretty 'slick' to me.

It will be interesting to see what Chipzilla advances in the next 12-18 months ...
May 31, 2008 12:26:57 PM

Quote:
Further, with the launch of the latest processors, AMD has once again taken the technological lead.


AMD have a number of design items on their CPU that Intel are trying to replicate for their next gen.


1. IMC
2. Split power planes
3. Monolithic quad
4. L3 cache structure



While AMD's design lead in these areas has not translated into better performance due to problems in these (and other) areas of the CPU, they are more advanced than their competitor in architectural approaches.
May 31, 2008 1:06:44 PM

And Intel is more advanced than AMD in practical approaches that turn into products and results for consumers.


I say this as a cheesed off former AMD fanboy. It just doesn't make sense...
a b à CPUs
May 31, 2008 1:22:08 PM

Quote:
perhaps the author feels they have the technological lead (that took so much brain power it was astonishing)

regardless of if the author is right or wrong i believe you are confusing performance lead with technological lead. there is a difference.



Well in terms of multiple socket (past 2S) then Barcelona does scale much better than Intel's current offerings in X86 ... so it does have a technological lead it already held ... plus with the lower power envelope that is of interest to large server systems.

But yes ... the Author isn't tech literate .. he is tech lite.

There is a big difference between single socket systems like Joe Public has and higher end servers.

Albeit in a number of areas Intel's Zeon servers perform exceptionally well.

Phenom is what we got as a consolation from Barcelona being redesigned for server duty ... a poor consolation compared to the core2's performance in terms of gaming and most of the applications of interest to the home user.

May 31, 2008 4:21:11 PM

aznguy0028 said:
omg....THUNDERMAN is back...! :D 


Every time I see that name I think of lyrics of Dan Fogelburg:

I hear the thunder(man)
Three miles away
The island’s leaking
Into the bay
The poison is spreading
The demon is free
And people are running from
What they can’t even see.
May 31, 2008 4:48:01 PM

TechnologyCoordinator said:
And Intel is more advanced than AMD in practical approaches that turn into products and results for consumers.


I say this as a cheesed off former AMD fanboy. It just doesn't make sense...


What doesn't make sense?


That AMD tried a more complicated approach which they haven't got to work right yet?



Look at the Nehalem architecture - now compare the amount of goodies it is getting that K10 has, and K8 had.



Technological probably isn't the best term the author could have used, design lead might have been a better one.
May 31, 2008 10:56:25 PM

Amiga500 said:
What doesn't make sense?

That AMD tried a more complicated approach which they haven't got to work right yet?
Not just that, but that they practically bet the company on said misexecuted complicated approach.
June 1, 2008 12:04:04 AM

The best techonology is the one that performs best in the metrics used to measure it. There is no technological or design lead without that. Anything else is wasted complication, not superior tech or design.
June 1, 2008 1:07:22 AM

SMU_Pony said:
The best techonology is the one that performs best in the metrics used to measure it. There is no technological or design lead without that. Anything else is wasted complication, not superior tech or design.



Well put. RESULTS MATTER. You can't tout and promote AMD because it is monolithic.
a b à CPUs
June 1, 2008 1:07:59 AM

SMU_Pony said:
The best techonology is the one that performs best in the metrics used to measure it. There is no technological or design lead without that. Anything else is wasted complication, not superior tech or design.


Now that isn't quite correct.

1. Sometimes the metrics being used are not an accurate measure.

Take a number of benchmarks for instance.

If we used superpi as a metric then core2 is twice as fast as K8 / K10 ... when in real world gaming the benchmarks don't seem to measure the same ... do they? Perhaps 5 to 10% faster in actual fact ... why ... the benchmarks is only measuring a small part of the sytem ... well cpu capability.

The benchmark has poor construct validity ... or more precisely .. your applying the benchmark inappropriately.

2. The best technology may require new benchmarks to be able to best measure it ... as the technology has produced a paradigm shift. Software may need to be completely recompiled in order to make best use of the technology. There are plenty of examples of this as well.

Trying to isolate the webbing in a preconceived idea of a nomothetic net is surely difficult (if not highly improbable) when in fact much evidence suggests reality is in fact more likely socially constructed, than not.

Sorry ... I went for a coffee before posting ... the dog wrote that last sentence ... I'll have him castrated.

:) 
June 1, 2008 4:19:37 AM

Reynod said:
The best technology may require new benchmarks to be able to best measure it ... as the technology has produced a paradigm shift.
Hehe... create new tests to best highlight the benefits and advantages of new technology... are you talking bechmarks or marketing?
a b à CPUs
June 1, 2008 4:25:24 AM

Aaahh ... your trying to dumb this down aren't you?

*prepares wet trout just in case*
a c 96 à CPUs
June 1, 2008 4:37:41 AM

sailer said:

Edit: I just saw an article which further helps the point of AMD and its technology.

http://www.tomshardware.com/news/PC-supercomputer,5513....

The research group Vision Lab at the University of Antwerp came up with a different solution and constructed a PC that integrates four GeForce 9800 GX2 graphic cards (with a total of eight GPU cores) that runs CUDA-optimized tomography applications. The specifications include a MSI K9A2 Platinum motherboard, an AMD Phenom 9850 CPU, 4 x 2 GB Corsair TWINX DDR2 PC6400 memory, a Samsung Spinpoint F1 750 GB hard drive, a Thermaltake Toughpower 1500W Modular power supply unit as well as four MSI 9800GX2 cards. The researchers said that the cost of the system was less than 4000 Euro or about $5300.

Note the use of an AMD 9850 quad core chip. Not implying that an Intel chip couldn't have been used or might have preformed better. As it stands, people, especially those in business, will see the use of an AMD quad core and think that if its good enough to do that type of work, it will be good enough for their use. You might object and say that business usage and not enthusiast, but the vast bulk of AMD' income is from business purchases, as it is with Intel's income. Still, this shows that AMD's chips are technologically capable of doing very complicated work.


The group wanted to have four PCIe x16 lanes to run the four 9800GX2s. This meant they needed a board with the AMD 790FX chipset as only those and the expensive and older ASUS L1N64 have four PCIe x16 lanes. They very well may have gone to an Intel solution if Intel had anything that can accommodate 4 PCIe x16 lanes for anywhere near what the K9A2 Platinum costs. The group used the Phenom X4 9850 as it was probably the fastest chip (to them) that they could put in the board and it is reasonably-priced at $235. The selection of the CPU + board was not made on technical prowess but simply what would best accommodate their intended setup.
June 1, 2008 5:19:33 AM

amd is so far back if intel told eveyone to take a year off
when intel got back amd would be still about in the same place
a c 96 à CPUs
June 1, 2008 5:28:47 AM

TechnologyCoordinator said:
Any news article that ends with "OMG AMD4LIFE!!1!" is obviously garbage.

Technological lead, lawl. Just because something is more complicated doesn't mean it is better technology.

Results matter, how you get there does not.


AMD does have a lead in technology in certain areas, though:

- Chipsets. The 700-series chipset line has the best IGP chipset (780G) and arguably the best performance desktop chipset as well (790FX.) Their chipsets are also very power-efficient although NVIDIA has a slight edge in IGP desktop chipset power consumption with the GF 8000 series.

- Low-power desktop CPUs. The Phenom X4 9100e is the only quad-core dissipating less than 65 watts. The 45-watt Athlon X2 4000e series have no official competition from Intel although many of the lower-clocked Core 2 Duos could be rated at 45 watts if Intel chose to do so.

- Servers, particularly 4S and 8S units. The 1600 MHz FSB Xeon Harpertowns (particularly the 3.40 GHz model) and the Xeon 5400 chipset have made big dents in AMD's lead in the 2S space but it will take Nehalem for Intel to really compete in the 4S and 8S arena.

- GPU efficiency. The Radeon 3000 series is quite efficient. The 4000 series doesn't look like it will suck down too much more power but the NVIDIA GT200 series sure will- the >600 mm^2 die on 65 nm on the GTX280 will ensure that. I wonder why NVIDIA went with 65 nm rather than 55 nm from TSMC as AMD was already using 55 nm to make their GPUs.

AMD certainly has the more complex platform and processor macro-architecture than Intel does, but that doesn't really matter as much as there isn't much of a tangible benefit from some of it at the present. The Phenom is extremely complicated compared to the relatively simple Core 2 Quad but is generally a touch slower clock for clock in consumer-type applications. The Phenom has very complex power-management features with differing core clocks and voltage planes while Intel has about the same level of speed/voltage management in the Core 2 desktop chips as they had ten years ago in PIII-Ms. Yet the 45 nm Core 2s generally consume less power than the Phenoms, particularly at idle where the Phenoms' power management should result in the opposite outcome. To make a car analogy, Intel is like a Viper while AMD is like a BMW M5. The Viper has a simple, huge, old-school pushrod V10. The M5 has an extremely complicated naturally-aspirated DOHC V10 with an 8250-rpm redline. But the M5 will lose to a Viper in pretty much all races even though the M5 has a lot more technology under its hood than the Viper.
June 1, 2008 6:44:36 AM

I don't really need to write anything. All this is from fangirl's emails

Quote:
Curious Rant: Well I depends on which type of applications your talking about. I mean like Barcelona (K10) was meant to be a server chip in the first place and a 2GHZ one can easily tie with some of the best 2.8GHZ Xeons. But if we're talking about games, which mind you is firstly not optimized for AMD and is very sensitive to cache and cache speed the Bacelonas kinda fails due to its lack of fast L2 cache. In fact most of the L2 work gets done in the L3, thus decreasing performance. However in the server space connection speed is the parallel of L2. Since AMD's HT is miles ahead of Xeons, it tends to be faster in those application.

I've seen many people insult AMD over the past, I dunno... AMD made a server processor, not a general purpose one. It's not at the top of the server hill (gets bested by clockspeed). Now before you mention anything about the Pentuim 4 being a parallel, its not. AMD's chips of the time were running way hotter than the Pentuim 4s. It was the Pentuim Ds that kinda screwed up as it dispersed way too much heat against similar speed Athlons.

AMD's chips are hot because they are monolithic and have less surface area to dispense heat, combined with an inferior 65nm process and thus resulting in huge low yield and etc. Now I'm not insulting Barcelonas, just explaining why they don't do well.

I don't know how blind the world is but most should be able to see that Barcelona is actually two Athlon X2 cores with one HT, improved stuff and a die shrink. Its not that much better than a Core 2 Quad. A completely new archeture would have been more suitable, but considering what pressure AMD was under...

In a way AMD has a technological lead, but it tends to benefit server applications whilst Intel's super-fast bounty of L2 cache which replicates the problem AMD had when Intel had quad pumped FSB and AMD only had dual, slower FSB, Its not fast enough, AMD's cache and it shows in real-life applications.

So you must realize that, in the end, AMD's Barcelona is a server CPU, not a gaming one. This way true also of the Athlon 64, but at the time it was closer in cache speed and quantity and Intel was being hampered by the relatively slow FSB bus.
It's like how ice must be consumed with water! So do you understand?
Any questions?
-fangirl =P


Anyone understand anything she said? I just asked her why is the Core 2 Quad better than a Phenom...
June 1, 2008 8:15:49 AM

macgirlfriend said:
I don't really need to write anything. All this is from fangirl's emails

Quote:
Curious Rant: Well I depends on which type of applications your talking about. I mean like Barcelona (K10) was meant to be a server chip in the first place and a 2GHZ one can easily tie with some of the best 2.8GHZ Xeons. But if we're talking about games, which mind you is firstly not optimized for AMD and is very sensitive to cache and cache speed the Bacelonas kinda fails due to its lack of fast L2 cache. In fact most of the L2 work gets done in the L3, thus decreasing performance. However in the server space connection speed is the parallel of L2. Since AMD's HT is miles ahead of Xeons, it tends to be faster in those application.

I've seen many people insult AMD over the past, I dunno... AMD made a server processor, not a general purpose one. It's not at the top of the server hill (gets bested by clockspeed). Now before you mention anything about the Pentuim 4 being a parallel, its not. AMD's chips of the time were running way hotter than the Pentuim 4s. It was the Pentuim Ds that kinda screwed up as it dispersed way too much heat against similar speed Athlons.

AMD's chips are hot because they are monolithic and have less surface area to dispense heat, combined with an inferior 65nm process and thus resulting in huge low yield and etc. Now I'm not insulting Barcelonas, just explaining why they don't do well.

I don't know how blind the world is but most should be able to see that Barcelona is actually two Athlon X2 cores with one HT, improved stuff and a die shrink. Its not that much better than a Core 2 Quad. A completely new archeture would have been more suitable, but considering what pressure AMD was under...

In a way AMD has a technological lead, but it tends to benefit server applications whilst Intel's super-fast bounty of L2 cache which replicates the problem AMD had when Intel had quad pumped FSB and AMD only had dual, slower FSB, Its not fast enough, AMD's cache and it shows in real-life applications.

So you must realize that, in the end, AMD's Barcelona is a server CPU, not a gaming one. This way true also of the Athlon 64, but at the time it was closer in cache speed and quantity and Intel was being hampered by the relatively slow FSB bus.
It's like how ice must be consumed with water! So do you understand?
Any questions?
-fangirl =P


Anyone understand anything she said? I just asked her why is the Core 2 Quad better than a Phenom...

It was a valiant effort, but there are a few flaws.
Quote:
AMD and is very sensitive to cache and cache speed the Bacelonas kinda fails due to its lack of fast L2 cache.
Not exactly. The core 2 suceeds because it's L1 data cache is well optimized for. See, with cache/memory, the smaller and closer to the chip, the better, but it still relies on the software to tell it what to load.
Quote:
Since AMD's HT is miles ahead of Xeons, it tends to be faster in those application.
Where HT helps most in servers, is where it is used as a dedicated link between chips, and for memory access.
Quote:

AMD made a server processor, not a general purpose one.
It's a general purpose processor, but so are all cpu chips. The thing it does best is multitasking. Where it falls down in desktop environment is lack of software support.
Quote:
AMD's chips of the time were running way hotter than the Pentuim 4s.
Some of the athlon xp chips did run hot, though by today's standards, with one of today's hsf they would be cool. The A64s which started shortly after the prescotts, were a very cool running chip.
Quote:
AMD's chips are hot because they are monolithic and have less surface area to dispense heat
Because the barcelonas are on 65nm, they have a much larger surface area. The AMD chips run hotter because they use a higher voltage, and have more internal resistance. AMD has a great process, but Intel's is much better.
Quote:
Barcelona is actually two Athlon X2 cores with one HT, improved stuff and a die shrink.
No die shrink, bothe the K10s and K8s are being made on 65nm. AMD has been using 65nm for about a year.
The K10s are sort of like the K8s if you forget about SSE4a, the doubling of the SSE bit width per core, the improved prefetch the out of order execution, the fact that each K10 core is only 3/4ths the size of a K8 core, and the rest of the total redisign.
Quote:
whilst Intel's super-fast bounty of L2 cache
Again, it is how well programs use the cache, (esp D1) that gives Intel that edge.
I doubt that will help you understand any better, sorry.
June 1, 2008 12:37:42 PM

MU_Engineer said:
AMD does have a lead in technology in certain areas, though:

- Chipsets. The 700-series chipset line has the best IGP chipset (780G) and arguably the best performance desktop chipset as well (790FX.) Their chipsets are also very power-efficient although NVIDIA has a slight edge in IGP desktop chipset power consumption with the GF 8000 series.


Agreed. The 700 series is awesome for what it does.
a c 96 à CPUs
a b À AMD
June 1, 2008 6:29:16 PM

TechnologyCoordinator said:
Agreed. The 700 series is awesome for what it does.


sb750 (this month if we can believe the buzz) should be a solid improvement to AMD 7xx chipsets and may well fix the PLL loopiness and improve OCs ...

A 5-10% improvement wouldn't be earth-shattering to Chipzilla but would give Team Green a little hope ...

The 780gx IGP with the frame buffer looks sweet - especially if it drives the 780g's down in price $10-$15
a b à CPUs
June 1, 2008 7:59:00 PM

When will people realize that once upon a time AMD made Intel chips? And that Intel had an IMC? :lol:  Oh.... the irony...

Quote:
In February 1982, AMD signed a contract with Intel, becoming a licensed second-source manufacturer of 8086 and 8088 processors. IBM wanted to use the Intel 8088 in its IBM PC, but IBM's policy at the time was to require at least two sources for its chips. AMD later produced the Am286 under the same arrangement, but Intel canceled the agreement in 1986 and refused to convey technical details of the i386 part. AMD challenged Intel's decision to cancel the agreement and won in arbitration, but Intel disputed this decision. A long legal dispute followed, ending in 1994 when the Supreme Court of California sided with AMD. Subsequent legal disputes centered on whether AMD had legal rights to use derivatives of Intel's microcode. In the face of uncertainty, AMD was forced to develop "clean room" versions of Intel code.


Quote:
Timna was the codename of a proposed Processor Family by Intel. The project was announced in 1999 and was designed in Haifa, Israel. "Timna" is also the name of a city in Israel. It was supposed to be the first Processor with an Integrated GPU and Memory controller which was designed to work with RDRAM. The price of RDRAM did not drop as expected by Intel. So, It was decided to use the Memory Translator Hub(MTH) that is also used by the Intel 820 chipset to link Timna with SDRAM. Later, a serious defect was discovered in the design of the MTH and so the Intel 820 based motherboards using it had to be recalled. The MTH was rebuilt again but problems still remained. Timna was therefore cancelled on 29 September 2000.


Links:
http://en.wikipedia.org/wiki/AMD
http://en.wikipedia.org/wiki/Intel_Timna



June 1, 2008 9:09:22 PM

While I agree AMD's northbridges are fantastic their current southbridges leave a lot to be desired.
June 1, 2008 10:49:24 PM

Shadow703793 said:
When will people realize that once upon a time AMD made Intel chips? And that Intel had an IMC? :lol:  Oh.... the irony...

Quote:
In February 1982, AMD signed a contract with Intel, becoming a licensed second-source manufacturer of 8086 and 8088 processors. IBM wanted to use the Intel 8088 in its IBM PC, but IBM's policy at the time was to require at least two sources for its chips. AMD later produced the Am286 under the same arrangement, but Intel canceled the agreement in 1986 and refused to convey technical details of the i386 part. AMD challenged Intel's decision to cancel the agreement and won in arbitration, but Intel disputed this decision. A long legal dispute followed, ending in 1994 when the Supreme Court of California sided with AMD. Subsequent legal disputes centered on whether AMD had legal rights to use derivatives of Intel's microcode. In the face of uncertainty, AMD was forced to develop "clean room" versions of Intel code.


Quote:
Timna was the codename of a proposed Processor Family by Intel. The project was announced in 1999 and was designed in Haifa, Israel. "Timna" is also the name of a city in Israel. It was supposed to be the first Processor with an Integrated GPU and Memory controller which was designed to work with RDRAM. The price of RDRAM did not drop as expected by Intel. So, It was decided to use the Memory Translator Hub(MTH) that is also used by the Intel 820 chipset to link Timna with SDRAM. Later, a serious defect was discovered in the design of the MTH and so the Intel 820 based motherboards using it had to be recalled. The MTH was rebuilt again but problems still remained. Timna was therefore cancelled on 29 September 2000.


Links:
http://en.wikipedia.org/wiki/AMD
http://en.wikipedia.org/wiki/Intel_Timna

Oh no, not the timna IMC again. Too bad you didn't read the wiki about rdram
Quote:
RDRAM includes a memory controller on each memory chip, significantly increasing manufacturing complexity compared to SDRAM, which used a single memory controller located on the northbridge chipset.

Not to say that timna didn't have a mem controller on the chip, it just wasn't a "real mem controller"
a b à CPUs
June 2, 2008 12:04:38 AM

endyen said:
Oh no, not the timna IMC again. Too bad you didn't read the wiki about rdram
Quote:
RDRAM includes a memory controller on each memory chip, significantly increasing manufacturing complexity compared to SDRAM, which used a single memory controller located on the northbridge chipset.

Not to say that timna didn't have a mem controller on the chip, it just wasn't a "real mem controller"

True, but it's still the forerunner to the IMC.
June 2, 2008 1:13:17 AM

Are you are trying to say that a lot of people tried to get an ODMC to work, but AMD was the first to get it right? Then I agree.
June 2, 2008 2:56:32 AM

endyen said:
Are you are trying to say that a lot of people tried to get an ODMC to work, but AMD was the first to get it right? Then I agree.


Are you seriously claiming that AMD were the first company to put an on-die memory controller on a CPU?

Hint: Intel had one on the 4004 in 1971.
June 2, 2008 5:22:29 AM

Yes, a lot of special purpose ig chips include a cache controller. They are good for a K-bite or two. As far as x86 class chips, AMD is the one.
!