Tuesday, May 08, 2007

K10 clock increase is 470 MHZ to be exact

INQ's reported 500MHZ hike was an approximation. The true number is 470MHZ (Of course, we know they are statistically equal.) Basically ,AMD's 65nm is doing 45nm speeds, the beauty of SOI+other exotic tricks.

This test show that AMD's 90nm X2 5600+ uses 20% less energy than Intel's E6400.

I heard Patty is polishing his resume and starts to call his old college friends. The problem for him is, there are only three x86 CPU makers and neither AMD nor VIA is interested.

100 Comments:

Blogger Aguia said...

Sharikou,

It seams that AMD is sending HP systems equiped with Athlon X2 and Intel Core 2 Duo to review.

Seams that they like the AMD system more.

Silent PC review

SharkyExtreme

8:42 AM, May 08, 2007  
Blogger jmack said...

Excellent read aguia.

After reading those articles, only a disillusioned fanboy would buy Intel.

8:56 AM, May 08, 2007  
Blogger Tommy said...

That's very true...

9:27 AM, May 08, 2007  
Blogger Only AMD said...

SWEET

10:44 AM, May 08, 2007  
Blogger Mahboobul said...

The war between AMD and intel was over the day C2D appeared in the market. It's all over for AMD. Perception is everything.

11:38 AM, May 08, 2007  
Blogger Scientia from AMDZone said...

This is the fudzilla link. The one in the article doesn't seem to be working.

mahboobul

"The war between AMD and intel was over the day C2D appeared in the market. It's all over for AMD. Perception is everything."

I thought it was over when Intel released SSE and then over again when Intel released Northwood. I guess this is the third time and presumably Nehalem will be the fourth.

12:55 PM, May 08, 2007  
Blogger Bubba said...

Oh please, quoting fudzilla.

Don't people care about their credibility?

1:25 PM, May 08, 2007  
Blogger pezal said...

Wow.. thanks for the links aguia..

4:45 PM, May 08, 2007  
Blogger R said...

Did AMD learn anything from this marketing exercise? No. Tomorrow goes back to business as usual and let Intel claim anything, and then walk all over AMD some more.

5:38 PM, May 08, 2007  
Blogger The Dude said...

Scientia from AMDZone said...

This is the fudzilla link. The one in the article doesn't seem to be working.

mahboobul

"The war between AMD and intel was over the day C2D appeared in the market. It's all over for AMD. Perception is everything."

I thought it was over when Intel released SSE and then over again when Intel released Northwood. I guess this is the third time and presumably Nehalem will be the fourth.


Actually, it was over the first time when Intel refused to give AMD the technical info for the i386, as previously agreed upon by the two companies. Intel agreed to this to fulfill a requirement by IBM that more than one source supplying its chips. AMD sought remedy in the courts, and in 1994 The Supreme Court of California sided with AMD.

Then it was over when AMD couldn't get K5 could to scale high enough in frequency to compete with Intel. Then AMD bought NexGen, and produced the K6.

Then it was over again when Intel moved to an exlusive socket for It's Pentium II processors, leaving AMD on socket 7. As I recall, socket 7 lived on for some time after that.

Soon, K6 was being left behind by the higher clocking, and generally better-performing Pentium III, and once again, rumors spread of AMD's demise.

Then AMD pulled out the Athlon...

6:40 PM, May 08, 2007  
Blogger Bubba said...

Then AMD pulled out the Athlon...

Which was then soundly beaten by the Northwood P4.

8:39 PM, May 08, 2007  
Blogger Poke said...

Funny how AMD fanboys keep on citing Tomshardware, Anand, Xbit, Hexus etc etc while insisting that they are Intel paid pumpers. You fanboys make me laugh. Bunch of hypocrits.

It doesn't matter, in the long run Intel is making billions in profit while AMD is sucking red and losing market share. Morgan Stanley will buy out AMD and sell its scrap heap FABs to make a nice profit.

9:17 PM, May 08, 2007  
Blogger The Dude said...

Then AMD pulled out the Athlon...

Which was then soundly beaten by the Northwood P4...

After over a year of regular Pentium ass-beatings at the hands of the Athlon.

9:41 PM, May 08, 2007  
Blogger core2dude said...

Reposting my own comment from the other thread.

Here is what will happen: Core 2 Duo will eat the dual cores (K8/K10) for breakfast. The advantage of K10 is not so much in the core. Afterall, it is more or less the same core. The real performance comes from the 4x4 crossbar, and increased bandwidth.

1333 MHz is plenty of bus for dual cores, as we all have seen (except for extremely memory intensive apps). With Penryn, Intel will up it a notch to 1600 MHz. That ought to keep K10 in check on any dual-core platform.

If K10 launches at 2.9 GHz, it may cause some problem for Intel in high-end DP and high-end 4P problems. You and other fanboys were plainly sounding ridiculous claiming that K10 will be a conroe killer at 2.5GHz.

And Scientia, when it comes to ridiculous, you yourself are not above others. Go back to your bog and read the posts where you claim Intel won't be able to launch Penryn at over 3 GHz. Guess what, Intel has already publicly demoed Penryn at 3.33 GHz.

The truth is, now Barcelona is looking a lot more competitive, and will have performance in the ballpark range of where Intel can take C2D. That will probably force Intel to release higher clocked C2Ds. C2D is not clocked higher because Intel does not need to. It is a chip that will easily go to 3.2+ GHz.

I have a C2D E6700. Guess what, I undervolt it to 1.22 V, and then overclock it to 3.2 GHz. And the system does not crash, ever! And if a CPU can give 20% overclock when undervolted by 5%, it is definitely underspecced.

Now I expect Intel to release Penryn at 3.5 to 3.6 GHz. That wills still be well within current TDP of 120 W for high-end QC. However, even at that frequency, K10 might be able to outperform it in specfp_rate.

9:45 PM, May 08, 2007  
Blogger rathor said...

Yeap, once again AMD is over Intel in almost every sector (gaming, office productivity, power consumption, and so on.). K10 it'll crush even "Penisryn" with all the tweaks inside; nor 45nm tech. or the future Nehalem it will not save Intel from certain death. (my opinion)

PS: I hate my C2D E6300, I'm so sorry that I have sold my X2 3800+...

10:30 PM, May 08, 2007  
Blogger Heat said...

PS: I hate my C2D E6300, I'm so sorry that I have sold my X2 3800+...

I will exchange my sempron for your e6300.....interested???

10:52 PM, May 08, 2007  
Blogger pezal said...

K10 it'll crush even "Penisryn" with all the tweaks inside; nor 45nm tech.

Penisryn?? wakakakaa.. you are so funny bud..

10:53 PM, May 08, 2007  
Blogger pezal said...

I will exchange my sempron for your e6300.....interested???

Im interested to exchange my celeron 1.70GZH with your sempron!

11:02 PM, May 08, 2007  
Blogger oneexpert said...

I have a e6400 and tested it for power consumption.
The 65 nm die is suppose to be power conserving but it is not in the intel chip and platform.
The e6400 platform and chip draw a whooping 85 watts of power when idle.
The e6400 chip and platform sucks 90 plus watts just to surf this blog.
The e6400 chip and platform sucks about 100 plus watts with both cores fully loaded.
My AMD 4800 65nm only draws 44 watts idle.
My AM2 4800 draws about 45 watts to surf this blog.
My AMD 4800 chip and platform draws 100 plus watts with both cores loaded but outperforms the e6400 by leaps and bounds.
If you want to save planet Earth and its resources buy AMD chips and platforms and save energy with out loosing performance.
If you wish to pollute and destroy the world, buy Intel chips and platforms, the real space heating energy wasters.
Anyway how can a super charged pentium 3 save intel?
No matter how fast you run a antique, its still just a antique.

11:36 PM, May 08, 2007  
Blogger rathor said...

I will exchange my sempron for your e6300.....interested???

Yeah, maybe 2 x semprons for 1 x e6300... That's fair.

2:21 AM, May 09, 2007  
Blogger Randy Allen said...

There's no need to believe people like oneexpert that are clearly attempting to mislead others.

We can see Core 2 Duo and Athlon 64 X2 power consumption at both idle and full load here:-

http://techreport.com/reviews/2007q2/core2-qx6800/index.x?pg=13

The performance can also been seen here:- http://www.xbitlabs.com/images/cpu/dualcore-roundup/charts/average.png

The E6400 has been replaced with the E6420, it has performance comparable to a 5600+. If you go for the higher performance you can see the 6000+ and E6600 are pretty equal.

You can also compare the E4400 (2Ghz/2MB Cache/800 FSB) to the 6320 (1.86Ghz/4MB Cache/1066 FSB). The extra cache and faster bus allow it to offer similar performance the E4400. Clearly, there is not a whole lot of difference between 2MB and 4MB of cache, and between 800 and 1066 FSB.

2:25 AM, May 09, 2007  
Blogger Randy Allen said...

No matter how fast you run a antique, its still just a antique.

So really, the new and upcoming K10 is just an upgrade of K8, which was upgrade of K7 which was just an upgrade of the K6 (which AMD didn't even design, they bought another company for that!).

Sure, each of those upgrades were quite substantial, but it's still based on that ancient K6 core when you look at it from your perspective. Comparing a Core 2 Duo to a Pentium 3 is just as crazy as comparing a K10 to a K6 CPU!

2:28 AM, May 09, 2007  
Blogger symbiansn said...

This comment has been removed by the author.

4:56 AM, May 09, 2007  
Blogger symbiansn said...

core2dude, with all that cognitive limits Intelers use to have, cannot recognize that AMD's spank new core is much superior than Intel's by design.

Multi-core and interconnect scalability is just another part of the show, where Intel will trail AMD apparently forever.

4:57 AM, May 09, 2007  
Blogger Rodney said...

AMD FTW! Huzzah!

5:21 AM, May 09, 2007  
Blogger Ycon said...

Intel offered a 533 MHz advantage with their 65nm compared to 90nm RIGHT OFF THE START and not 1 year later.

5:44 AM, May 09, 2007  
Blogger Aguia said...

Intel offered a 533 MHz advantage with their 65nm compared to 90nm RIGHT OFF THE START and not 1 year later.

Where did you see that?

The Pentium 4 90nm did 3.8GHz, I don’t think the 65nm clocked higher.

The Core 2 Duo didn’t exist in 90nm so it’s impossible to compare to the 65nm version.

However the mobile version at 90nm did 2266Mhz maximum, the current version does 2333Mhz, its an 67Mhz leap.

I’m very interested in where you got that data from.

6:11 AM, May 09, 2007  
Blogger pezal said...

Idiot fanbois said..

There's no need to believe people like oneexpert that are clearly attempting to mislead others. We can see Core 2 Duo and Athlon 64 X2 power consumption at both idle and full load here..



Push button said..

We looked at power consumption of our two testbeds, however AMD 4600+ is at a bit of a disadvantage here.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2903&p=7



The truth is..

The energy savings of the AMD 4600+ over the Core 2 Duo 6300 in real use will probably be at least 20%, and considerably more if the computer is powered on more than four hours a day, the conservative figure used for our calculations. For typical home or office use, as long as the AMD CPU performance remains within 10~20% of Intel's, it's the lower idle power that looks compelling to us. For a corporation running hundreds or thousands of PCs, >30% energy savings is nothing to scoff at in this day of rising energy and environmental costs.

7:14 AM, May 09, 2007  
Blogger Ho Ho said...

aguia
"Where did you see that?"

Compare P4D 840 vs P4D 955 and 965. First was 90nm 3.2GHz, second was 65nm 3.466GHz and third 3.733GHz. 955 was released simultaneously with single core 65nm CPUs, 965 was released two months later.

Good enough histlry lesson for you?

8:18 AM, May 09, 2007  
Blogger Aguia said...

This comment has been removed by the author.

8:44 AM, May 09, 2007  
Blogger Aguia said...

Compare P4D 840 vs P4D 955 and 965. First was 90nm 3.2GHz, second was 65nm 3.466GHz and third 3.733GHz. 955 was released simultaneously with single core 65nm CPUs, 965 was released two months later.

Good enough histlry lesson for you?


Not convinced.
The Pentium D 8xx is one die dual core processor, the Pentium D 9xx it’s two die dual core processor. Not really directly comparable.

I already have said the Pentium4 90nm clocked 3800Mhz, it’s not impressive that the 65nm part you are talking about hit 3733Mhz.

8:44 AM, May 09, 2007  
Blogger Heat said...

Rathor: Yeah, maybe 2 x semprons for 1 x e6300... That's fair.

Alrite you drive a hard bargain i will send you two semprons for your sucky crappy e6300 or i will give you a 3800+ for it whatever you prefer let me know if you are interested. Dont want a fellow tech enthusiast stuck with a crappy processor..........lets see you put your money where your mouth is...

9:19 AM, May 09, 2007  
Blogger Giant said...

R600 Review: http://it-review.net/index.php?option=com_content&task=view&id=1314&Itemid=1

No one wants this garbage. It gets flogged by the 8800 GTX in virtually every benchmark except for Company of Heroes. Yes, that's right! AMD's top of the line GPU thrashed by the six month old 8800 GTX! Nvidia didn't even need the Ultra to totally frag R600.

AMD BK Q2'08.

9:21 AM, May 09, 2007  
Blogger Heat said...

giant:No one wants this garbage. It gets flogged by the 8800 GTX in virtually every benchmark except for Company of Heroes. Yes, that's right! AMD's top of the line GPU thrashed by the six month old 8800 GTX! Nvidia didn't even need the Ultra to totally frag R600.

This is what happens when fanbois get giddy over slide shows and "expected, estimated" benchmarks and numbers. K10 is a tweaked K6 and will compete with the celeron but ultimately lose.

9:30 AM, May 09, 2007  
Blogger Mahboobul said...

AMD has obviously survived for 50 odd years. But it has never thrived. About a 18 months ago, it seemed that AMD had finally broken the trend. Alas! we were wrong.

9:32 AM, May 09, 2007  
Blogger Bruno Dieter Chan said...

http://www.theinquirer.net/default.aspx?article=39477

10:23 AM, May 09, 2007  
Blogger oneexpert said...

The data speaks for itself AMD is the winner in performance and low power consumption.
Never believe anyone who does not provide data from there own equipment tests.
Trust but verify.
The e6400 and e6600 chips and intel platforms suck a lot more power than the 65nm AMD 4800 brisbane.
Anybody who tries to convince you otherwise is simply a big fat liar.
The AMD 4800 chip and platform draws only 44 watts at idle while the e6400 and e6600 chips and platforms draw 80 watts at idle.
Watt meters are watt meters and I would bet the lives of my critics that my data is 100% accurate and correct and repeatable by any one of you reading this blog.
And any intel fanboy who is not aware that the c2d is nothing but a reworked pentium 3 is simply out to lunch.
Intel platforms and chips suck way too much power compared to AMD energy conserving designs.
When the c2d/pentium 3s were originally designed power was not the issue and never considered by intel.
Buy AMD and be a energy wise and save our planet.
AMD does more work for a lot less power and a lot less cost to buy.
Energy does matter and AMD chips and platforms save lots of energy.

12:08 PM, May 09, 2007  
Blogger Poke said...

No one wants this garbage. It gets flogged by the 8800 GTX in virtually every benchmark except for Company of Heroes. Yes, that's right! AMD's top of the line GPU thrashed by the six month old 8800 GTX! Nvidia didn't even need the Ultra to totally frag R600.

Well from the looks of it, Nvidia doesn't even need the 8800GTX, as the 8800GTS 640MB frags the HD2900 in almost every benchmark.

AMD BK Q2'08

12:50 PM, May 09, 2007  
Blogger oneexpert said...

How the data was gathered.
Parts std.hipro 400watt power supply,one samsung 250gb sata drive,one hp dvd unit, same memory stick.
AMD test one 4800 brisbane on a rs485 std. amd/ati chipset board.
Intel test 1 e6600 on a std. 945 chipset motherboard and 1 e6400 on a std. 945 chipset motherboard.
My setups are very similar to your home computer to give as accurate as possible data that you would experience.
Idle watts= e6600 80 watts.
e6400 was 80 watts.
Idle watts on AM2 4800 was 44 watts.
All were running linux 5.0 at idle.
AMD clearly the 44 watt winner.
To my critics why dont you buy a watt meter and give us your set up and data.
Anything but real data is just hog wash.
Buy AMD and help save our planet and resources.

1:04 PM, May 09, 2007  
Blogger pezal said...

No one wants this garbage. It gets flogged by the 8800 GTX in virtually every benchmark except for Company of Heroes.


Nvidia is AMD best partner.. So, it is does not a matter whether Nvidia or ATI win the contest it will be the same to us.

The true opponent of AMD is Intel.. I bet, there will be no more chances to intel to win the games when K10 is launched this Q3. Intel better get prepared. Muhahahaha..

5:45 PM, May 09, 2007  
Blogger Sharikou, Ph. D said...

Nvidia is AMD best partner.. So, it is does not a matter whether Nvidia or ATI win the contest it will be the same to us.


That's very true. AMD and Nvidia are on good terms today. Both want to get a piece of Intel's pie. The advances made by AMD/ATI just make Intel and its customers look like retards. K10 will open a perf gap that Intel will never be able to fill.

5:50 PM, May 09, 2007  
Blogger Giant said...

Well from the looks of it, Nvidia doesn't even need the 8800GTX, as the 8800GTS 640MB frags the HD2900 in almost every benchmark.

AMD BK Q2'08


The 8800 GTS is the "smarter choice" unless you game at really high resolutions. The HD2900 can compete better there because it has tons of memory bandwith. The GTX still frags R600 all over though, not to mention the Ultra, which is up to 15% faster.

8:05 PM, May 09, 2007  
Blogger Giant said...

AMD bought Ati, Nvidia's chief competitor. AMD thinks that they can just get away with that?

Not a chance. Nvidia and Intel are both larger than AMD, and will severely punish AMD for trying to take Nvidia and Intel's business.

AMD BK Q2'08.

8:09 PM, May 09, 2007  
Blogger Poke said...

This comment has been removed by the author.

8:13 PM, May 09, 2007  
Blogger Poke said...

Nvidia is AMD best partner.. So, it is does not a matter whether Nvidia or ATI win the contest it will be the same to us.

In your dreams fanboy,

http://xbitlabs.com/news/video/display/20070507140244.html

Intel and Nvidia Reportedly Prep Another Cross-Licensing Agreement.

Intel Seeking for Nvidia's GPU Technologies

"Given that in the past Intel bundled graphics cards from ATI and Nvidia with some of its mainboards and the fact that Intel is unlikely to bundle graphics cards from its arch-rival Advanced Micro Devices, close collaboration of Intel and Nvidia may also include promotion of Nvidia GeForce graphics cards, something, which is very dangerous for ATI/AMD."


http://www.dailytech.com/article.aspx?newsid=7036

"AMD scrambles to raise capital but risks long term company health"

In the long run AMD is losing market share and sucking red while Intel is making billions in profit.


http://www.dailytech.com/article.aspx?newsid=7051

"Intel Wipes Out AMD's 2006 Marketshare Gains in One Quarter."

Intel took all of AMD's 2006 market share in 1 quarter. Who is going to BK again?

8:20 PM, May 09, 2007  
Blogger Poke said...

I heard old Hector polishing his resume but I don't think Intel or Via are interested in having another janitor.

8:32 PM, May 09, 2007  
Blogger Giant said...

I heard old Hector polishing his resume but I don't think Intel or Via are interested in having another janitor.

No matter. Hector made $16m last year. He is fine even if AMD BKs in Q2'08.

8:38 PM, May 09, 2007  
Blogger mi7chy said...

Storm is Coming
AMD True Native Quadcore

http://www.hardocp.com/news.html?news=MjU2OTYsLCxoZW50aHVzaWFzdCwsLDE=

http://www.hardforum.com/showthread.php?t=1189400

8:58 PM, May 09, 2007  
Blogger Heat said...

Damn a picture of a desktop. i am soo excited.......

9:34 PM, May 09, 2007  
Blogger abinstein said...

"Damn a picture of a desktop. i am soo excited......."

It says in the article that the box encodes Spiderman3 to H.263 in almost real-time. It is something...

10:15 PM, May 09, 2007  
Blogger abinstein said...

Ooops... it's H.264, not H.263 :p

10:20 PM, May 09, 2007  
Blogger Giant said...

Wow. A picture. Intel showed off Penryn working back in January, and allowed sites like Anandtech access to Penryn systems at IDF in Beijing. They were restrictive, but that's to be expected, given that what they showed off is likely not the final working product.

I doubt people at this AMD event were allowed to go and take whatever benchmarks they wanted either.

10:46 PM, May 09, 2007  
Blogger pezal said...

Intel and Nvidia Reportedly Prep Another Cross-Licensing Agreement.

Intel Seeking for Nvidia's GPU Technologies

It is rumoured that Intel plans to enter the market of standalone graphics processors in 2009 or 2010. Moreover, Intel already confirmed that its Nehalem processors will have built-in graphics core, something, which may be considered as a threat for Nvidia and ATI, graphics product group of AMD.






Muhaahahaa.. Dont u get the point fanboy? Intel hire Nvidia so that nvidia can teach Intel how to make graphic card... Wakkkaka.. Whats a funny..

Wow.. intel want to compete with Nvidia and AMD in Graphic contest.. I bet, intel will occupying the 3rd place in the graphic contest with a performance gap of 500% below NVIDIA and AMD scores.. Wakkkakaka funny..:-)

10:59 PM, May 09, 2007  
Blogger abinstein said...

"I doubt people at this AMD event were allowed to go and take whatever benchmarks they wanted either."

I simply don't get it. So you Intel fanboys would rather believe a few bush-button benchmarks than the rea-life, near real-time H.264 encoding result? Do you really buy computers just to show good benchmark scores?

11:27 PM, May 09, 2007  
Blogger Giant said...


I simply don't get it. So you Intel fanboys would rather believe a few bush-button benchmarks than the rea-life, near real-time H.264 encoding result? Do you really buy computers just to show good benchmark scores?


So we had, what, the encoding on a 4x4 8 core system with two quad core AMD CPUs? That's great. What I'd like to see is the same thing run on an 8 core Intel setup, a Mac Pro for instance.

How well can an 8 core 3Ghz Clovertown system do at this encoding?

12:18 AM, May 10, 2007  
Blogger Heat said...

abinstein:I simply don't get it. So you Intel fanboys would rather believe a few bush-button benchmarks than the rea-life, near real-time H.264 encoding result? Do you really buy computers just to show good benchmark scores?

I am not sure if you just act stupid or you really are that dumb. Barcelona is just around the corner and you want us to be happy with a picture of a freaking case!!!

Intel had penryn up and running almost a half a year plus before its release. Not to mention Intel had to implement an achievement like high K gate which was not thought to be possible in 45nm and let people run benchmarks on it!! You are telling me that what intel did is worse than a freakin picture of a case and their word on its running spider man 3!!

The only thing dumber i have heard from you is saying that famous post about 3/4 victory in benchmarks is not a unanimous win. Its almost like AMD fanbois are retarded at this point between you, pezal, sharikou and scientia's latest post on his blog.

pezal:Wow.. intel want to compete with Nvidia and AMD in Graphic contest.. I bet, intel will occupying the 3rd place in the graphic contest with a performance gap of 500% below NVIDIA and AMD scores.. Wakkkakaka funny..:-)

You keep laughing it up while intel holds the largest marketshare in graphics.... imagine it getting better what will happen of ATI/AMD's share then....wakakakakakakakakak....not to mention AMD's latest six month late garbage is competing with Nvidia's third best wakakakakaka. ....

RATHOR I AM STILL WAITING WITH MY 3800+ TO EXCHANGE WITH YOUR E6300 seems like another AMD fanboi talking out of his ass and shuts up when you call them on it........

12:33 AM, May 10, 2007  
Blogger pezal said...

How well can an 8 core 3Ghz Clovertown system do at this encoding?


It's actually 8/2=4 cores only.. ;-) Ekekeke.. Due to the bad scalability of the Intel 8/2=4 CPU, Im sure it will get frogged by about 50% behind the true 8 cores of AMD CPU.. Ekekekkee

12:44 AM, May 10, 2007  
Blogger pezal said...

You keep laughing it up while intel holds the largest marketshare in graphics.... imagine it getting better what will happen of ATI/AMD's share then....wakakakakakakakakak....not to mention AMD's latest six month late garbage is competing with Nvidia's third best wakakakakaka. ....


Yeah.. let’s laugh it up together.. Don’t worry AMD will grab back half of the market share this Q3 and grab all of the share by Q208… Wkakakakaka… ATI/AMD will be going much better when QCBarcelona and ATI2950XT being launched this Q3.. Wakkakakaka.. AMD’s 4th best graphic is match with Nvidia 3rd best graphic… Intel best graphic totally no match even with ATI and Nvidia low class graphic.. wakakakakaaka..

1:42 AM, May 10, 2007  
Blogger abinstein said...

"Barcelona is just around the corner and you want us to be happy with a picture of a freaking case!!! "

I don't want you to be anything. I just pity that Intel believers like you 1. couldn't read and acknowledge facts, 2. have bad reason to buy a computer (for benchmark, more precisely).

The fact is you can't do anything close to real-time H.264 encoding with 8-core clovertown... it's just too slow and too poorly (un)scalability

2:21 AM, May 10, 2007  
Blogger abinstein said...

"What I'd like to see is the same thing run on an 8 core Intel setup, a Mac Pro for instance."

Mac pro Clovertown encodes at about 1x to 1.3x speed of Conroe. Obviously, dual Clovertown will have even lower scalability over single Clovertown, probably at most 1.2x.

AMD is playing at its strength here. I can see the whole video processing industry being excited about Barcelona's released 2 months later. Mac users are of no luck because Apple decided to swallow Intel's FSB and MCM approach. Well, it's not too late for them to switch, though. Still 2 months away. :-)

2:29 AM, May 10, 2007  
Blogger Ho Ho said...

aguia
"Not convinced.
The Pentium D 8xx is one die dual core processor, the Pentium D 9xx it’s two die dual core processor. Not really directly comparable."


What exactly makes them not directly comparable? Do they magically have some kind of different heat dissipation methods due to not being sawn to two pieces? Only thing they did with 9xx was they saw the die to two pieces. They could have done texactly the same thing with 8xx seres but for some reason they did not.

"I already have said the Pentium4 90nm clocked 3800Mhz, it’s not impressive that the 65nm part you are talking about hit 3733Mhz."

65nm did that with dualcore, 90nm was 3.8GHz max with single core.


mi7chy
"Storm is Coming
AMD True Native Quadcore"


Wow, so many huge fans cooling that thing down

abinstein
"It says in the article that the box encodes Spiderman3 to H.263 in almost real-time. It is something..."

Anyone got the data how fast can Intel 45nm CPUs encode movies? IIRC they were twice as fast as their 65nm ones, at least with DivX. Also that AMD one had to use two quadcores to achieve real-time speed.

"So you Intel fanboys would rather believe a few bush-button benchmarks than the rea-life, near real-time H.264 encoding result?"

How fast is "near real-time"? How can we know it used all the optimizations that x264 has? I can encode a full HD movie into x264 at "near real-time" on my e4300 also but without all those extra optimizations burn CPU time like there is no tomorrow. To me that one sentence doesn't show absolutely anything as there is no good information there.


bezal
"Due to the bad scalability of the Intel 8/2=4 CPU, Im sure it will get frogged by about 50% behind the true 8 cores of AMD CPU.."

Without proper description of the benchmark we can't even try to make a valid comparison, can we?

" AMD’s 4th best graphic is match with Nvidia 3rd best graphic"

Do yo know that by the end of this year NV will most likely have G90 on the shelves, or at least 65nm shrink of its current 90nm GPUs. That is right, AMD 65nm high-end GPUs are fighting against NV 90nm ones and so far it seems as the 65nm is loosing.


aguia
"couldn't read and acknowledge facts"

So some random dude telling you something is a fact?

"The fact is you can't do anything close to real-time H.264 encoding with 8-core clovertown"

How do you know this is a fact? Can you provide links or we simply have to believe you since, well, you said that and you are always 100% correct?

"Mac pro Clovertown encodes at about 1x to 1.3x speed of Conroe. Obviously, dual Clovertown will have even lower scalability over single Clovertown, probably at most 1.2x."

A lot of uncertanity there. I take you actually have no idea of the MacPro performance.

2:37 AM, May 10, 2007  
Blogger pezal said...

Without proper description of the benchmark we can't even try to make a valid comparison, can we?

With the hints given of course we can do the estimation.. That is according to my OPINION fanboy.. Ekekeke.. Stupid fanboy cant even estimate how long his ..... are. ekekekee



That is right, AMD 65nm high-end GPUs are fighting against NV 90nm ones and so far it seems as the 65nm is loosing.


Yeah.. it's similar to intel 65nm CPUs that is so far seems like its still loosing to AMD old 90nm ones.. Ekekeke..

3:04 AM, May 10, 2007  
Blogger Giant said...

I don't think so. AMD's fastest CPU is the 6000+. Intel's sixth fastest desktop CPU, the E6600, competes with the 6000+ and is actually a tiny bit faster.


http://www.xbitlabs.com/images/cpu/dualcore-roundup/charts/average.png

Nice try though.

(Intel has E6700, X6800, Q6600, QX6700 and QX6800 all faster than E6600)

3:30 AM, May 10, 2007  
Blogger Poke said...

Yeah.. it's similar to intel 65nm CPUs that is so far seems like its still loosing to AMD old 90nm ones.. Ekekeke..

Then why does every tech websites I visit conclude that C2Ds frag Athlon X2s? Oh, thats right, they are Intel paid pumpers. No wonder why AMD fanboys cite the same Intel paid pumper websites to support their flawed arguments.

It doesn't matter in the long run because AMD is losing market share and sucking red while Intel is making billions.

AMD BK Q2'08

3:32 AM, May 10, 2007  
Blogger pezal said...

AMD's fastest CPU is the 6000+. Intel's sixth fastest desktop CPU, the E6600, competes with the 6000+ and is actually a tiny bit faster.

AMD fastest CPU is QXPhenom 2.9GZH.. AMD 6000+ is the 9th faster of AMD CPUs..

So, thats mesn Intel 6th faster Vs AMD old 9th faster.. After 60-90 minutes due to the overheating+power hungry, intel E6600 will be totally frogged by AMD 9th faster CPU.. ekekeke

3:40 AM, May 10, 2007  
Blogger pezal said...

Then why does every tech websites I visit conclude that C2Ds frag Athlon X2s? Oh, thats right, they are Intel paid pumpers. No wonder why AMD fanboys cite the same Intel paid pumper websites to support their flawed arguments.

Nope.. Is actually because of you are idiot fanboy.. Idiot like you of course cannot differentiate which one is old which one is latest.. Although the old one still perform better, but Idiot person like you will be choosing idiot perception. Ekekekekekeke...

3:54 AM, May 10, 2007  
Blogger Tanrack said...

can I buy a QXPhenom 2.9GZH? I need it by tomorrow

4:22 AM, May 10, 2007  
Blogger pezal said...

can I buy a QXPhenom 2.9GZH? I need it by tomorrow

Isn't you want to buy Penisryn?

4:40 AM, May 10, 2007  
Blogger Giant said...


AMD fastest CPU is QXPhenom 2.9GZH.. AMD 6000+ is the 9th faster of AMD CPUs..


Really?! Please kindly direct me to a website where I can buy anything faster than a 6000+ for socket AM2.

4:53 AM, May 10, 2007  
Blogger pezal said...

Really?! Please kindly direct me to a website where I can buy anything faster than a 6000+ for socket AM2.

You are kidding not buying.. ;-) I know. Die hard Intel fanboy like you better waiting for the penisryn. It will be more interesting to see AMD phenom frogging the pensryn soon.

5:07 AM, May 10, 2007  
Blogger mi7chy said...

Intel Hack Jobs

- hyperthread phony multiprocessing
- netburst low instructions per cycle
- front side bus bottleneck
- bloated cache to make up for architecture deficiences
- phony quad core
- scaling issue with >2 core

6:36 AM, May 10, 2007  
Blogger Evil said...

But pezal, you silly little AMD fanboi, brainwashed by Ph(ake)d... last week Toms Hardware was pumped by Intel according the fanboi's around here...

But yet suddenly its OK to use them as they posted a pro AMD review?

How does that work?

If its pro Intel, Intel MUST HAS TO BE! They are pumping the review sites!!!! OMG!!!

But when they post positive AMD, everything is good?


Holy double standards Batman.


Get your head outta your ass pezal, you are looking quite the stupid

7:08 AM, May 10, 2007  
Blogger Evil said...

From AMD's Q-10 filing: BIG OUCH

Computing Solutions net revenue of $918 million in the first quarter of 2007 decreased 31 percent compared to net revenue of $1,337 million in the first quarter of 2006. The first quarter of 2006 did not include revenue from the sale of ATI chipsets. Net revenue decreased primarily as a result of a 32 percent decrease in average selling prices and a 15 percent decrease in unit shipments for our microprocessors as compared to the first quarter of 2006. Average selling prices for our microprocessors decreased due to competitive market conditions and aggressive pricing by our principal competitor. Moreover, our competitor offered quad-core multi-chip module processors during the first quarter of 2007, and since we did not offer quad-core products during this period, we discounted the selling price of certain of our competing products which adversely impacted our average selling prices, margins and profitability. Unit shipments for our microprocessors decreased for a number of reasons: we experienced lower demand for our microprocessors, in part because our competitor offered quad-core microprocessors during this period while we did not; we experienced a decrease in sales through our distributor channel; and we continued to experience challenges in the ability of our supply chain to deliver products in the right mix and on a timely basis. Sales through our distributors were adversely impacted in the first quarter of 2007 because we were not able to adequately meet their demand in the second half of 2006.

Computing Solutions net revenue of $918 million in the first quarter of 2007 decreased 38 percent compared to net revenue of $1,486 million in the fourth quarter of 2006 primarily as a result of a 37 percent decrease in unit shipments and a 10 percent decrease in average selling prices for our microprocessors as compared to the fourth quarter of 2006. Unit shipments for our microprocessors decreased for the reasons set forth above. Average selling prices for our microprocessors decreased due to continued aggressive pricing by our competitor and the competitive product environment discussed above.

Computing Solutions operating loss was $321 million in the first quarter of 2007 compared to operating income of $312 million in the first quarter of 2006. The decrease in operating results was primarily due to a 31 percent decrease in net revenues, an increase of $63 million in research and development expenses and an increase of $39 million in marketing, general and administrative expenses. Net revenue decreased for the reasons set forth above. Research and development expenses and marketing, general and administrative expenses increased for the reasons set forth under “Expenses” below.

Computing Solutions operating loss was $321 million in the first quarter of 2007 compared to operating income of $65 million in the fourth quarter of 2006. The decrease in operating results was primarily due to a 38 percent decrease in net revenues, a $15 million increase in research and development expenses and a $25 million increase in marketing, general and administrative expenses. Net revenue decreased for the reasons set forth above. Research and development expenses and marketing, general and administrative expenses increased for the reasons set forth under “Expenses” below.

7:13 AM, May 10, 2007  
Blogger Randy Allen said...

See this:

http://www.sec.gov/Archives/edgar/data/2488/000119312507108224/d10q.htm#tx64403_7

Scroll to page 26 and read for yourself how Intel is killing AMD.

Average selling prices for our microprocessors decreased due to competitive market conditions and aggressive pricing by our principal competitor. Moreover, our competitor offered quad-core multi-chip module processors during the first quarter of 2007, and since we did not offer quad-core products during this period, we discounted the selling price of certain of our competing products which adversely impacted our average selling prices, margins and profitability.

The price war is taking a huge toll on AMD. That's good. And while AMD talks about quad core CPUs Intel actually ships them.

Computing Solutions net revenue of $918 million in the first quarter of 2007 decreased 38 percent compared to net revenue of $1,486 million in the fourth quarter of 2006 primarily as a result of a 37 percent decrease in unit shipments and a 10 percent decrease in average selling prices for our microprocessors as compared to the fourth quarter of 2006. Unit shipments for our microprocessors decreased for the reasons set forth above. Average selling prices for our microprocessors decreased due to continued aggressive pricing by our competitor and the competitive product environment discussed above.

In addition to lower ASPs AMD is losing tons of market shares. No one wants AMD Turdion or Crapteron processors. AMD is totally fragged in servers, mobile and desktop CPUs.

Nvidia and Intel have created a strategic alliance to utterly destroy AMD and force them to go BK so that Morgan Stanley gets the company and sells off the scraps to Nvidia and Intel.

You are kidding not buying.. ;-) I know. Die hard Intel fanboy like you better waiting for the penisryn.

So the 6000+ is AMD's fastest CPU after all?! The previous statement stands:- AMD's fastest CPU competes with Intel's sixth fastest CPU. AMD's fastest GPU (when it finally get launched - seven months late) will compete with Nvidia's third fastest GPU.

It will be more interesting to see AMD phenom frogging the pensryn soon.

AMD fans said wait R600! It will frag Geforce 8800 GTX, worth the wait! Turns out the opposite was true. It's the same here, Barcelona is all talk. Clovertown is so advanced with it's four issues wide core that it has totally prefragged Barcelona.

AMD BK Q2'08.

7:18 AM, May 10, 2007  
Blogger pezal said...

last week Toms Hardware was pumped by Intel according the fanboi's around here...


Thats according to them not me. I never used to said Toms is intel pumped.

I think you are the one that is looking quite stupid here, especially when you are wearing that batman costume. Where the hell you get it? You aren’t look like a superhero but most probably like Mr bean.. Ekekekeke…

7:22 AM, May 10, 2007  
Blogger pezal said...

This comment has been removed by the author.

7:32 AM, May 10, 2007  
Blogger pezal said...

So the 6000+ is AMD's fastest CPU after all?! The previous statement stands:- AMD's fastest CPU competes with Intel's sixth fastest CPU.

Nope.. Why acting like stupid donkey fanboy.. AMD already declare about their new CPUs line up.. Aren't you know about that? It is remaining the same 6th Intel fastest Vs 9th AMD

7:33 AM, May 10, 2007  
Blogger Randy Allen said...

Nope.. Why acting like stupid donkey fanboy.. AMD already declare about their new CPUs line up.. Aren't you know about that? It is remaining the same 6th Intel fastest Vs 9th AMD

Ok. So if AMD has faster CPUs, where can I buy one???

7:53 AM, May 10, 2007  
Blogger Randy Allen said...

Here is a link to the fastest AMD CPU one can buy:-

http://www.newegg.com/Product/Product.aspx?Item=N82E16819103773

Can you provide a link to a faster CPU from AMD?

7:56 AM, May 10, 2007  
Blogger Tanrack said...

pezal said Isn't you want to buy Penisryn?

Is it faster than than the QXPhenom 2.9GZH?

8:11 AM, May 10, 2007  
Blogger enumae said...

Sharikou

AMD Cutting 430 Jobs

Any comment?

8:12 AM, May 10, 2007  
Blogger pezal said...

Ok. So if AMD has faster CPUs, where can I buy one???

Whats wrong with you all guys keep pushing asking the same question. Why dont you go directly to AMD factory to buy it? Ekekkekee..

8:14 AM, May 10, 2007  
Blogger Lou Ceifer said...

I have FINALLY learned the secret and reasons behind the fanatical Intel mouthbreather's zeal on this blog today. This is an old post, but a very informational one and very insightful to why Intel fears AMD like they do and in turn, thier fanboys as well.

Taken from ZDNet's article, "AMD scores huge 7-round TKO over Intel in dual-core benchmarks" on 12/07/05 by COMALite J:

http://talkback.zdnet.com/5208-10532-0.html?forumID=1&threadID=15629&messageID=313261&start=-9896

Intel is TERRIFIED of AMD. Why? Because Intel remembers Zilog!

Back in the days not long after the invention of the wheel, there was something called 8-bit CPUs. Yes, only 8 bits. Their clock speeds were measured in single-digit MHz. That was not a typo. That was an "M", not a "G".

Intel had the first of these, the Intel 8008, shortly followed by the 8080, an improved version of the 8008. The 8080 typically ran at 4MHz, but was generally considered inferior to the MOS Technologies 6502 (used in the Apple I, ][, ][+, ///, and the first //e computers, as well as in Commodore PETs and Atari 400/800/1200s) that only ran at 1MHz, and yet could do at least as well in benchmarks as a 4MHz 8080, even though the 8080 had more instructions (in many ways, the 6502 was a precursor of the RISC concept).

Digital Research came out with a popular OS for the 8080, called Control Program for Microcomputers, or CP/M. Many companies made CP/M computers, and the first generation ones all used Intel 8080s.

Then some tiny upstart company named Zilog makes a CPU that is compatible with the 8080, has many more instructions (using opcodes unused by the 8080 and so remaining fully binary compatible with it), can run at faster clock speeds. This CPU is called the Z80. Later versions, the Z80A and Z80B, run at faster clock speeds than any 8080, up to a whopping 8MHz! And yet they still cost LESS than the Intel 8080!

Suddenly, almost EVERY CP/M computer maker has switched over to the Zilog Z80! All newcomers to the CP/M market also use the Z80 family. The market share of the 8080 drops to the single digits. The INSTALLED BASE share becomes majority Z80, so much so that developers actually feel comfortable taking advantage of the extra Z80 instructions for even better performance, even though this means that their programs won't run on an 8080-based CP/M machine!

Even NON-CP/M machines come out that use the Z80, the most famous being the Tandy Radio Shack Z80, or TRS-80, Models I, II, III, IV, and 12. TRS-80s used their own proprietary OS, TRSDOS, which took full advantage of the Z80 instructions and so could not run at all on an 8080. Even Apple Computer, when they made a coprocessor card so that Apple ][ users could run CP/M, used the Z80, not the 8080.

Intel tried to counter with the 8085, which had its own set of a handful of extra instructions that were incompatible with the Z80, but nobody used it. It was way too little, way too late.

Zilog even came out with a version that had 16-bit extensions, but would still run 8-bit software as-is and at full speed, with those speeds being much faster than before (double-digit MHz!): the Z800.

Intel was on death’s door, hemmorhaging red ink. They'd lost the 8-bit CPU market that they had invented. They made a last-ditch effort to enter a new market, one that already had some other CPUs: the then-new 16-bit CPU market.

Intel created the 8086 CPU, with full 16-bit internal and external busses. While not directly binary compatible with the 8080, software written for the 8080 could be easily re-compiled, and the segmented memory model made it easy for such programs to run as they were used to.

About the same time, Zilog was working on their own fully 16-bit CPU, the Z8000 family.

But a 16-bit computer would be much more expensive than an 8-bit one, because it would need external busses and thus motherboard traces that were twice as wide. It would take 16 memory chips, not just 8, to fill a row, for instance.

Both saw that, for the initial transition, there would need to be a line of CPUs that were fully software-compatible with the full 16-bit versions, but had only 8-bit external data busses. Such CPUs would be slower than the 16-bit versions of the same family would be at the same clock speed, because whenever the CPU needed to fetch a 16-bit number, it would need to use two fetch cycles instead of only one because the number would have to be split into two 8-bit bytes. They also could not address as much memory, being limited to the 64kB memory space of the 16-bit address bus that 8-bit CPUs used. But systems built using them could be made much less expensively than full 16-bit versions.

Intel called theirs the 8088 (8086 with an "8" on the end to denote the 8-bit external data bus, instead of the "6" that denoted 1"6"-bit), and Zilog called theirs the Z8002 (they named the fully 16-bit version the Z8001).

Getting these hybrid 8/16-bit versions to work properly was tricky, though. One fateful day, IBM decided that they wanted to enter the personal computer market then dominated by the Apple ][+, TRS-80 [Z80-based] Model III, and various CP/M machines [nearly all based on the Z80]. They wanted the machine to be 16-bit, as the next step up from the 8-bit CPUs then dominating the market. But they didn't want it to really be all that fast: to save costs, they wanted an 8-bit external bus version of a 16-bit internal CPU. And they had no intention of clocking it to its full speed.

IBM asked both Intel and Zilog if they had an 8-bit external, 16-bit internal CPU ready. Neither company did, but both were close, merely having a few remaining glitches to work out.

Zilog told the truth to IBM, that their Z8002 was almost ready but that there were still some glitches to be resolved, but that they should be before IBM was ready to launch the Personal Computer (code-named "Peanut"). Intel LIED to IBM, claiming that their 8088 was ready to go (it wasn't -- it still had about as many glitches as the Z8002).

THIS LIE IS THE **ONLY** REASON THAT INTEL STILL EXISTS AT ALL, let alone has their enormous domination of the desktop CPU market! Intel owes that domination entirely to that lie. Had Zilog been dishonest and Intel honest, Intel would be a footnote in the history books now, just as Zilog is in reality. If IBM had not gone with the 8088, there's no way that Intel would've survived even another half a decade.

(A similar thing happened with the OS for the new computer: IBM wanted an OS for the 8086 [remember that the 8088 was to be fully software {and thus OS} compatible with the 8086!], and Digital Research had one: CP/M-86, the official 16-bit version of the venerable 8-bit OS that had been Intel's only reason for existence for so long. Microsoft had no OS, but said that they did. Gary Kildall, the owner of Digital Research, had been out flying his private plane when IBM came knocking, and his wife refused to sign a non-disclosure agreement, so IBM went to Microsoft instead. Bill Gates got the contract, and immediately went out and BOUGHT an 8086 OS called "QDOS" ["Quick-and-Dirty OS"], which was a rip-off of CP/M-86 by a company called Seattle Computer Systems. Microsoft paid only $50 thousand dollars for this, and it became the basis of THEIR domination of the desktop OS market.)

Anyway, now you know why Intel takes the threat of tiny AMD so seriously. They've seen their total domination of the CPU market taken away from them before by a superior upstart, and they barely survived it then. They see AMD as as much a threat now as Zilog was then. They have no intentions of letting it happen again. THIS time, there's no big savior corporation creating a whole new (and vast) market trying to decide which CPU to use, who could save them (IBM is a pale shadow of what they were back then).


Most enlightening.

8:41 AM, May 10, 2007  
Blogger Randy Allen said...

Yes. AMD will be crushed like an ant.

AMD BK Q2'08.

8:58 AM, May 10, 2007  
Blogger Lou Ceifer said...

This comment has been removed by the author.

9:51 AM, May 10, 2007  
Blogger Lou Ceifer said...

Randy Allen spewed...

"Yes. AMD will be crushed like an ant.

AMD BK Q2'08."


Thus sayeth the Intel fanboyz chorus. *yawn*

Here's another footnote for you beanheads referring the previous article I just copied here, ZiLOG is still around by the way.

http://www.zilog.com/

Surely if Zilog survived Intel's rise to a CPU monopoly, I don't think AMD will have a problem surviving thier current situation either. Namely because of K10, the video encoding demo results referred to by Kyle at HardOCP is quite amazing, I know movie studios will look forward to that for sure. AMD's diversity is also a great thing, I expect their other markets geared towards home HD digital TV and video solutions to increase as well.

9:57 AM, May 10, 2007  
Blogger Evil said...

Why can't the Intel fanboi's say AMD BK by Q208?

After all, it seems fine for the AMD fanboi's to say Intel BK Q208....

11:18 AM, May 10, 2007  
Blogger abinstein said...

"Ok. So if AMD has faster CPUs, where can I buy one???"

You will be able to buy a Barcelona before you could buy a Penryn.

11:21 AM, May 10, 2007  
Blogger abinstein said...

"So the 6000+ is AMD's fastest CPU after all?! The previous statement stands:- AMD's fastest CPU competes with Intel's sixth fastest CPU."

It really depends on what application you use to benchmark.

For video/3D/synthetic fanboy website benches, what you said might be true.

For SPEC, which is a much more balanced and standard measure of CPU performance, the lead of Core2 is very limited, probably 10% performance or 15% clock difference.

11:24 AM, May 10, 2007  
Blogger abinstein said...

"AMD's fastest CPU is the 6000+. Intel's sixth fastest desktop CPU, the E6600, competes with the 6000+ and is actually a tiny bit faster."

This is pure ignorance. You can't determine CPU performance ranks just by a few selected programs. I own an E6600 and while I'd always use it for divx encoding, for most other workloads it's often slower than my X2 5600 (which is cheaper, btw).

"AMD fastest CPU is QXPhenom 2.9GZH.. AMD 6000+ is the 9th faster of AMD CPUs.."

Nop... AMD's fastest CPU on desktop is 6000+, which isn't competitive to E6700 or X6800. However, for 4 cores or higher, multi-socket Opteron is still very competitive to most if not all Core2-based servers due to Opteron's good scalability.

Once Barcelona is released, Core2 stands no chance in the server space.

11:40 AM, May 10, 2007  
Blogger Poke said...

No matter. Hector made $16m last year. He is fine even if AMD BKs in Q2'08.

Hector should be taken to court for misleading investors. They should sue him for all his money. After he gets out of jail, he will need that job as head janitor at Intel.

12:12 PM, May 10, 2007  
Blogger Ho Ho said...

pezal
"With the hints given of course we can do the estimation.. "

Yes, we can form some opinon based on the hints. Mine currently is that AMD is not showing real world numbers since it wouldn't make it look so good.


"Yeah.. it's similar to intel 65nm CPUs that is so far seems like its still loosing to AMD old 90nm ones.. Ekekeke.."

Certainly not in maximum performance segment. As you don't use GPUs in servers then scalability is not needed and even still SLI scales/works better than CF.


"AMD fastest CPU is QXPhenom 2.9GZH.. AMD 6000+ is the 9th faster of AMD CPUs.. "

1) sell me one now
2) show me some real world benchmark numbers

If you fail to do either then at least give me dates when I can buy it or when can I see the benchmark numbers from real world applications.



mi7chy
"- hyperthread phony multiprocessing"

See the other thread and think of Nehalem.


"- netburst low instructions per cycle"

Northwood was quite good against K7. Soon after Intel hit the thermal wall and got surprassed by K8.


"- front side bus bottleneck"

Even today this is a bottleneck in few selected workloads


"- bloated cache to make up for architecture deficiences"

It is a design decision. AMD requires low latency RAM, Intel can do just fine with considerably cheaper higher latency one.


"- phony quad core"

Works fine for the vast majority of clients and is considerably cheaper than monolithic core.


"- scaling issue with >2 core"

Highly application dependant. When your programs are badly written then using different architecture will only make the issue a bit less but won't remove it. Just take Cinebench. Any schoolkid with basic threading knowledge can code a raytracer with perfect scaling on tens of PC's connected over 100mbit LAN with insane latencies and nonexistent throughput. For some reason their can't work well even on directly connected AMD box.



abinstein
"You will be able to buy a Barcelona before you could buy a Penryn. "

I don't want 4x4, I don't want server CPU. I want single CPU desktop PC.

"which is a much more balanced"

You are joking, right?


"for most other workloads ..."

Care to make a list?


"Once Barcelona is released, Core2 stands no chance in the server space."

Why? Because it will also have four cores? Because of direct connect? If so then why has Intel taken so much server share from AMD with its old 2P solutions?

12:17 PM, May 10, 2007  
Blogger Poke said...

Barcelona has no chance as it already has been pre-fragged by Clovertown. A K8 with a patched up Hyper Transport won't help AMD's financial situation. It doesn't matter what you AMD fanboys think because the FACT is AMD just lost all of its 2006 market share gains in 1 quarter. They are sucking red and losing millions.

You AMD fanboys are pathetic, you can cry all you want but the end result won't change. Intel making billions, AMD losing millions.

12:17 PM, May 10, 2007  
Blogger R said...

Hey Lou Ceifer, thanks for nice read. Be Well.

12:51 PM, May 10, 2007  
Blogger R said...

Because of the Osborne effect AMD doesn’t release much data; especially on the near future release of Barcelona.

1:00 PM, May 10, 2007  
Blogger abinstein said...

"You will be able to buy a Barcelona before you could buy a Penryn. "

I don't want 4x4, I don't want server CPU. I want single CPU desktop PC.


What are you, reading impaired? Did I mention 4x4 at all? Barcelona is a great server CPU as well as desktop.


"which is a much more balanced"

You are joking, right?


Tell that to the SPEC people. At least spec is what the processor research community use to determine microarchitecture progress. Much more balanced than your ray tracing crap.


"for most other workloads ..."

Care to make a list?


I have made the list, but either you're so dumb-wit or lack of lasting memory you simply can't get hold of them. Such intentionally ignorant response is truly pathetic.


"Once Barcelona is released, Core2 stands no chance in the server space."

Why? Because it will also have four cores? Because of direct connect?


Your mind is blind. Opteron at the moment already has better SPECint_rate and SPECfp_rate than Core2 at 8 cores. Barcelona is only going to extent that lead.

Core2's MCM & FSB is a joke for scalability, which already suck around 2-socket 8 cores. In short, you get only 6.5 cores with an 8-core Clovertown.


If so then why has Intel taken so much server share from AMD with its old 2P solutions?

Intel is a great marketing company, and most people (like yourself) follow marketing blue crystals, not technology innovations.

2:13 PM, May 10, 2007  
Blogger Ho Ho said...

abinsein
"Barcelona is a great server CPU as well as desktop."

When can I but a relatively high-end desktop version of Barcelona? Wasn't it some time in late Q3/early Q4? How much would a Q6600 competitor cost, less than $266?


"At least spec is what the processor research community use to determine microarchitecture progress"

Those benchmarks (specFP/int) measure only a few specific areas and you definitely can't say they are balanced or resemble real-world workloads, at least not the ones majority uses.


"Opteron at the moment already has better SPECint_rate and SPECfp_rate than Core2 at 8 cores."

Ok but how many people will buy 8+ cores?


"Core2's MCM & FSB is a joke for scalability, which already suck around 2-socket 8 cores."

Interesting that this doesn't disturb the people who are buying those machines.


"In short, you get only 6.5 cores with an 8-core Clovertown."

As I've said N+1 times, this is highly dependant on your specific application that 6.5 is most definitely not the overall average.


"Intel is a great marketing company, and most people (like yourself) follow marketing blue crystals, not technology innovations."

Marketing can't make something out of nothing but they can make competitive product to sell better. Intel was loosing marketshare fast with its P4D but once Core2 arrived in masses it retook a lot of it very fast.

I myself use the thing that gives me the biggest bang for my money. So far this has mostly been Intel, had my CPU upgrades been shifted by one year I would have used more AMD CPUs than Intels (K8 instead of Northwood, x2 instead of P4D, K10 instead of Core2).

2:38 PM, May 10, 2007  
Blogger Evil said...

AMD fanboi's seem to forget that Intel isn't sitting still...

1.) Intel's new high-k dielectric 45nm process

2.) Frontside bus speeds of up to 1600MHz

3.) 50W to 120W power utilization on Quad Core depending on clock speed

4.) Dynamic acceleration tech

5.) Much improved VT

Now thats just with Penryn. Lets not forget that Nehalem is not far behind it either.

So yeah, its POSSIBLE that Barcelona takes the speed crown from C2D, but once Penryn hits, its a whole different story.

People who underestimate Intel typically do what AMD is doing now, panic....

2:50 PM, May 10, 2007  
Blogger enumae said...

Sharikou

Any comments on this.

"Quad-core, codenamed Barcelona will launch later this summer, in the July, August kind of time frame -- followed by [Agena FX] on the desktop," stated Robert Rivet, AMD executive vice president and CFO.

6:18 PM, May 10, 2007  
Blogger Giant said...

Hector should be taken to court for misleading investors. They should sue him for all his money. After he gets out of jail, he will need that job as head janitor at Intel.

After Morgan Stanley takes charge of AMD once it goes BK they will fire the entire management team. Then Intel will hire them all:- Hector can be the head janitor, and the rest of the AMD management team can be serving under him as janitors.

7:36 PM, May 10, 2007  

Post a Comment

Links to this post:

Create a Link

<< Home