Sign in with
Sign up | Sign in
Your question

Quad-Core AMD Opteron deliver world record web performance

Tags:
  • CPUs
  • Performance
  • Quad Core
  • AMD
  • Opteron
Last response: in CPUs
Share
July 29, 2008 4:01:33 PM

Lets all congratulate AMD on this latest Achievement....Intel left for dust here

Article found here:
http://finance.google.com/finance?q=amd&hl=en

Article:
http://www.marketwatch.com/news/story/quad-core-amd-opt...{A91527AE-8E1E-41EC-8CCD-79B2284ED7F3}&dist=hppr

Article Quotes:

Quote:
SUNNYVALE, Calif., Jul 28, 2008 (BUSINESS WIRE) -- Demonstrating its superior ability to handle the strenuous and increasing demands placed on today's high-traffic Web-based businesses


Quote:
AMD Opteron(TM) processor Model 2356 and Quad-Core AMD Opteron processor Model 8356 now hold the top x86 web performance records for both 2P and 4P servers, as measured by the SPECweb(R)2005 benchmark. These records demonstrate that Quad-Core AMD Opteron processors deliver the ideal platform for dynamic web-based businesses to help scale with peak loads, improve server utilization, minimize data center power consumption and deliver next generation Web 2.0 infrastructure


Quote:
Web-based businesses, including hosted services for customers or internally facing business critical functions like CRM environments over a corporate intranet, often require more peak performance and scalability to handle heavy user traffic while balancing cost and power concerns


Quote:
These benchmark results are the latest evidence of the clear value that Quad-Core AMD Opteron processors offer an Internet business - or any data center that requires the ultimate in performance, reliability and power efficiency


Quote:
SPECweb(R) 2005 is an objective and representative benchmark for measuring a system's ability to act as a web server, simulating simultaneous user requests for standardized web transactions such as banking, e-commerce and downloading dynamic content


Quote:
With its Direct Connect Architecture, Web-based companies world-wide, including DAUM Communications, Rackspace and Strato, have benefited from Quad-Core AMD Opteron processor-based systems in their datacenter infrastructure.


AMD4Life!

More about : quad core amd opteron deliver world record web performance

July 29, 2008 4:06:02 PM

AMD is going to have a very short life, so it might be a good idea to change your signature. :) 
July 29, 2008 4:55:35 PM

antonmadcow amd like IBM is seeing the bigger picture.

who buys more cpu's at a PROFITABLE % for a manufacture????



how about the millions of chips sold to cooperations, government institutions " think hospitals, government building, data centers, university's and colleges " and the countless CPU's sold just to maintain the Infer structure known as the internet.

think about if in your data cluster you can do with 700 cpus vs 900 cpus from prior to the upgrade. think of the saving for a data center being space " square footage as good luck renting office space for less then 10$ a square foot per month " cooling , power and the cost to have IT techs to keep the system running. Less is MORE.

now think about if you could save a data center 1 $ per cpu per week in power. now expand your brain past that and think 1000's of cpu's

please take your troll posts elsewhere.
a c 127 à CPUs
a b À AMD
July 29, 2008 5:03:23 PM

macer1 hence why Intels Larrabee and Nehalem will do well in all those areas.

As for the OP, wow big suprise. AMD does better in multi core CPU systems like always. Wow. Amazing. Oh if they are so good how come AMD has yet to break even?

What does this have to do with the desktop market again? :heink: 
July 29, 2008 5:07:06 PM

General_Disturbance said:
Baron, is that you?

No way, Thunderman is WAY more entertaining than Baron :D 
a c 117 à CPUs
a b À AMD
July 29, 2008 5:14:33 PM

Does this mean better HD p0rn and fewer gambling losses?

Good for AMD. Hopefully better in the future above 2.3GHz with 'true' HT3. 45nm may shine on the Optys.

I'm still wunderin' when AMD chipsets will roll on the server side ....
July 29, 2008 5:14:40 PM

I would argue the exact opposite. Thurderman is more like 9-inch: he posts, then he runs for cover, while the rest of his threads goes down the drain.

BM on the other hand, has the balls to stand for his perspectives, no matter how wrong they are (CPU storage in styrofoam anyone?). Not to mention that BM is hilarious to argue with, as he often trips.
July 29, 2008 5:22:01 PM

jimmysmitty im just stating that at the time when the life cycle of your "data" center or servers are up most management "who sign off on the expense" want to know EVERYTHING.


now if the IT department can with current technology save this company X $$$$ every year with faster equipment then to the management its a no brainer.

and lets be honest most management couldn't tell you the difference between a 3.6 ghz P4 and a e4400. All they see is the big numbers and they are sold on the product.

hell go down to a local best buy and watch people get sold on a bigger LCD over better components in the box...........
a c 127 à CPUs
a b À AMD
July 29, 2008 5:25:19 PM

^This happens all the time yes. My own opinion is that in the next few months the whole PC market will be reshaken. Nehalem will be out for servers and desktops and that will cause AMD a bit of dismay.

But what happens in the server market rarely affects the desktop market.
a b à CPUs
July 29, 2008 5:48:59 PM

jimmysmitty said:
macer1 hence why Intels Larrabee and Nehalem will do well in all those areas.

As for the OP, wow big suprise. AMD does better in multi core CPU systems like always. Wow. Amazing. Oh if they are so good how come AMD has yet to break even?

What does this have to do with the desktop market again?


Wow! I could feel the sarcasm oozing thru my monitor as I read this... :ouch: 

Regardless, you gotta AMD props for pulling it together with the B3 revisions, especially after the whole TLB debacle. Better late than never in this case.

Albeit, Nehalem and/or Larabee may do well in the same areas, they are essentially unproven in the server market. AMD has already hurdled any issues with implementing an IMC processor and platform, whereas Intel will have some growing pains with Nehalem. And while AMD may have lost a little ground to Xeons in recent Quarters, it's news like this that will further embed and reinforce Opteron in the market making it that much more difficult for IT and Data Centers Managers to switch to a Nehalem platform. The new B3 opterons are a drop in solution for the Skt1207 whereas Nehalem will require an entirely new platform. That's makes it that much more difficult to justify the increased capital expense of a new platform as opposed to a drop in solution.
July 29, 2008 6:21:44 PM

Oh... yes... I guess that I have to back up my perspective somehow... let's see... how could I do that... hmm... how about...

$4.11! DUH!

Would the precious few people here who know how to spell "business" raise their hands? Reading the posts of the blind leading the blind AMD fans you'd think that AMD had an exemption from market economics! :) 

These Opteron results will mean next to nothing to the clients. They are not going to invest in a chip manufacturer who may not be around next year. IT Managers got to their positions by not being myopic.
July 29, 2008 9:15:09 PM

antonmadcow said:
Oh... yes... I guess that I have to back up my perspective somehow... let's see... how could I do that... hmm... how about...

$4.11! DUH!

Would the precious few people here who know how to spell "business" raise their hands? Reading the posts of the blind leading the blind AMD fans you'd think that AMD had an exemption from market economics! :) 

These Opteron results will mean next to nothing to the clients. They are not going to invest in a chip manufacturer who may not be around next year. IT Managers got to their positions by not being myopic.

Yes, there is something weird going on with this thread. You are being incessantly downrated for stating what seems to me to be obvious truths. Even Thunderman cannot deny, AMD's well is is running dry :p 
July 29, 2008 9:25:52 PM

AMD's finances are crippled...they have the moral victory. I can see the company hitting big soon. ATI 4000 series is already destroying Nvidia and the Spider platform is unique with up to four cards in Crossfire. Four 4870's combined with the Phenom 9950 will absolutely destroy anything Nvidia and Intel can offer in games. AMD is leading the gaming market clearly with a strong platform.
What we can gather by early evidence is Nehalem is probably going to be a failure.
a b à CPUs
July 29, 2008 9:36:42 PM

What evidence do you have that Nehalem will fail? If anything, the evidence shows it will be out earlier than expected, for cheaper than expected, and be a significant boost over Core.


EDIT: I love how a post simply asking for evidence gets rated down so much.
July 29, 2008 9:38:34 PM

thunderman said:
... What we can gather by early evidence is Nehalem is probably going to be a failure...


link please!?
July 29, 2008 9:50:35 PM

thunderman, your blind fanboyism is truly disgusting. Carry on.
July 29, 2008 9:59:26 PM

LOL. How many of these post are you ganna make Thunderman?

In a few months you will be on here asking what Intel procs are better because your like the other fan boys. When the boat sinks you will jump on the ship thats there to rescue you eventually.

What about a intel quadcore @4.0 with quad xfire. Your little AMD wont even come close.
a c 127 à CPUs
a b À AMD
July 29, 2008 10:01:13 PM

chunkymonster said:
Wow! I could feel the sarcasm oozing thru my monitor as I read this... :ouch: 

Regardless, you gotta AMD props for pulling it together with the B3 revisions, especially after the whole TLB debacle. Better late than never in this case.

Albeit, Nehalem and/or Larabee may do well in the same areas, they are essentially unproven in the server market. AMD has already hurdled any issues with implementing an IMC processor and platform, whereas Intel will have some growing pains with Nehalem. And while AMD may have lost a little ground to Xeons in recent Quarters, it's news like this that will further embed and reinforce Opteron in the market making it that much more difficult for IT and Data Centers Managers to switch to a Nehalem platform. The new B3 opterons are a drop in solution for the Skt1207 whereas Nehalem will require an entirely new platform. That's makes it that much more difficult to justify the increased capital expense of a new platform as opposed to a drop in solution.


I was being quite sarcastic yes. But here is the interesting part. I do give AMD props for finally getting more Opterons of the B3 stepping out. I just don't see it as relevant to anything if AMDs desktop market is not doing as well.

As for Nehalem, don't judge too early. What happens when people judge a new technology before there are benchmarks? Makes them look bad if they are wrong. Everyone hyped Phenom and it wasn't anything near its hype. Nehalem may be a newbie to the IMC but Intel is not and has tried on several occasions (before AMD) to create them and I am sure they are able to make it work. We will have to wait and see.

thunderman said:
AMD's finances are crippled...they have the moral victory. I can see the company hitting big soon. ATI 4000 series is already destroying Nvidia and the Spider platform is unique with up to four cards in Crossfire. Four 4870's combined with the Phenom 9950 will absolutely destroy anything Nvidia and Intel can offer in games. AMD is leading the gaming market clearly with a strong platform.
What we can gather by early evidence is Nehalem is probably going to be a failure.


LMAO. You forget that Intel has decently priced X38 and X48 mobos that also offer CFX at the same speed (4 8x lanes) as AMD. So lets take a 9950 with 4 4870s (or better yet 2 4870X2s in 2 x16 mode) and a Q6600 OCed to 2.6GHz (to match the 9950) and same either 4 4870s or 2 4870X2s in CFX and see which one will win.

Oh one thing, well two really, if you think AMD is a angel company you are blind. They are just as corrupt as every other company out there and they only want your money like Intel, IBM and all the others. And the second thing, I don't think that winning a moral race will pay the bills. If they fail financially thats it. Nothing short of a miracle can bring them back.

God fanboyism is fun to rip apart.
July 29, 2008 10:02:03 PM

cjl said:
What evidence do you have that Nehalem will fail? If anything, the evidence shows it will be out earlier than expected, for cheaper than expected, and be a significant boost over Core.


skittle said:
link please!?


It will fail because Nehalem is a longer word than Phenom. And it's hard to market the Nehalem name. I mean, Phenom is easy. Phenom-eee-nal. w00t! That is trademarked, by the way. Try with Nehalem. Hmm.... Ne-hay....uh...yes what I mean.

You don't need links. It fails simply because it's a much longer word and hard to pronounce.

:pt1cable: 
a c 127 à CPUs
a b À AMD
July 29, 2008 10:04:29 PM

^LMAO!!!!

It will probably be Core 3 or something to that effect.
July 29, 2008 10:05:15 PM

LMAO. Look at thundermans profile and look at the threads he has posted. Every single on is AMD fanboyism. AMD should be paying this guy.
a c 127 à CPUs
a b À AMD
July 29, 2008 10:07:42 PM

^No he should be paying them due to all the damage he is doing....all the posts are just....that stupid....
July 29, 2008 10:15:28 PM

evilshuriken said:
thunderman, your blind fanboyism is truly disgusting. Carry on.

:lol:  How can you not see the sarcasm dripping from every one of his posts?! No -real- poster ever says things like "AMD4Life!" and "they have the moral victory" ;) 
July 29, 2008 10:19:46 PM

Well. I used to think he was just a good troll. But now I'm not so sure, I mean, who puts that much effort into being a troll?
So I concluded that he must not be a troll at all, but in fact, a true "believer".
a b à CPUs
July 31, 2008 2:22:38 PM

I just love how the ratings are going ...

July 31, 2008 6:50:37 PM

homerdog said:
:lol:  How can you not see the sarcasm dripping from every one of his posts?! No -real- poster ever says things like "AMD4Life!" and "they have the moral victory" ;) 

If you want to see a true blind fanboy look up BaronMatrix :whistle:  (either that or he owns a lot of AMD stock..)



Posts like this may make me start rating some of this drivel. What does this post have to do with me?
July 31, 2008 7:29:02 PM

^Sorry Baron, that was pretty inappropriate. Edited :sol: 

BTW, I didn't know you were posting here again. Should I take this as a sign that AMD has something big up its sleeve?
July 31, 2008 7:59:40 PM

The funniest thing about that is that it's the 8356 NOT the 8360SE. I can imagine what the Shanghai @ 2.7GHz will do. From some of the numbers in HPC, Nehalem WON'T close the gap. Intel's own Penryn release numbers show the 2.3Barcelona outrunning a 2.93Penryn and close to a 3.2GHz. I believe Anand posted those numbers a few months ago. LS-DYNA and Fluent were the apps in question.
a c 127 à CPUs
a b À AMD
July 31, 2008 10:04:54 PM

^Um a die shrink? Maybe add 10%. Thats all really. And don't talk about what you don;t know BM.

Oh and I know its you doing the down rating. Real mature.
July 31, 2008 10:30:17 PM

BaronMatrix said:
The funniest thing about that is that it's the 8356 NOT the 8360SE. I can imagine what the Shanghai @ 2.7GHz will do. From some of the numbers in HPC, Nehalem WON'T close the gap. Intel's own Penryn release numbers show the 2.3Barcelona outrunning a 2.93Penryn and close to a 3.2GHz. I believe Anand posted those numbers a few months ago. LS-DYNA and Fluent were the apps in question.

Got any links? I'd like to know more about Shangai seeing as how AMD seems to be keeping everyone in the dark. If it can beat Nehalem in HPC like you say it can that would be something... Nehalem is going to be a HPC beast.
a c 102 à CPUs
July 31, 2008 11:55:20 PM

macer1 said:

and lets be honest most management couldn't tell you the difference between a 3.6 ghz P4 and a e4400. All they see is the big numbers and they are sold on the product.


Most people here couldn't tell the difference between a 3.6 GHz P4 and an E4400 without looking as they are roughly the same performance in single-threaded apps and both show up as two CPUs in Task Manager. Almost all office apps are single-threaded, so the two processors feel very similar.

Quote:
hell go down to a local best buy and watch people get sold on a bigger LCD over better components in the box...........


Even the cheapest hardware today has more than enough power to do typical home-user stuff (Web browsing, e-mail, consumer digital photography, playing music), so a larger, higher-resolution monitor is more useful for those purposes than faster CPUs that would spend 99% of their time idling.
July 31, 2008 11:57:02 PM

homerdog said:
Got any links? I'd like to know more about Shangai seeing as how AMD seems to be keeping everyone in the dark. If it can beat Nehalem in HPC like you say it can that would be something... Nehalem is going to be a HPC beast.



Find Intel's benchmarks for Penryn server. LS-DYN and Fluent. I'd do it but...
August 1, 2008 12:49:49 AM

^Huh? Remember, it's Shanghai vs Nehalem; Penryn will be phasing out by the time AMD get to 45nm ;) 
a c 127 à CPUs
a b À AMD
August 1, 2008 1:01:10 AM

I was going to say what does Penryn have to do with Nehalem? Nehalem is a whole new beast and has a IMC with links as fast as (or faster Im not sure which as I have seen both sayings) AMDs HTT 3.0.
August 1, 2008 11:35:56 AM

^Common System Interface is faster than HyperTransport, which isn't to say that HT is slow but CSI is very fast. The interface doesn't matter as much in the desktop market, but in HPC it can make all the difference along with an IMC.

I still don't see how Shanghai could be faster than Nehalem at anything, but as is always that case with AMD I would love to be wrong. A world without AMD is not a world I want to see :cry: 
a c 127 à CPUs
a b À AMD
August 1, 2008 2:13:12 PM

^Agree. I have been lead to believe (from some AMD fans here) that Shanghai is the desktop version of Deneb. Or is it the other way around?

I don't see a die shrink making Barcy get a 25% performance jump. If it was a die shrink and a rework of the arch maybe. But not when its just a die shrink.

I still cannot believe that Baron wants to use Penryns results (considering Penryn is still on the FSB) as proof that Nehalem will not be faster.

I did read that the CSI (QPI now) on the next Itanium, Tulkila?, is as fast as a 4P AMD server. But I am not that sure. And we will have to wait and see the results on both Shanghai and Nehalem, especially Nehalem to see if Intel is able to impliment the IMC and links good.
August 1, 2008 2:50:18 PM

People you are getting really crazy with the grading of posts.The thing about it is that points are being devalued with out though and with that makes the thread invalid.I am an AMD fan,but fair is fair.Please read the post ,twice if you have to, but don't just mark them down.Antonmadcow do make very little sense most times though.
a c 127 à CPUs
a b À AMD
August 1, 2008 2:54:28 PM

^Yup. And you notice its only to those who are not AMD fans that are getting graded down. And funny things is most of the ones I get graded down on are either infomation or non AMD bashing just talking.

Its a sad sad world we live in.
August 1, 2008 3:04:49 PM

THG needs to fix it or given us the ability who is grading
August 1, 2008 3:10:03 PM

Ah, it's just a little toy for those that just think that anything negative about AMD is AMD bashing.
I'm sure this will get voted down, only because of what I typed, even though there is nothing negative towards AMD.
It's just childish behavior.
I frankly don't care, but if they want to make this place into an AMD fansite, there are many already in existence.
August 1, 2008 3:45:36 PM

jimmysmitty said:
I still cannot believe that Baron...

Well I hate to go back to the Baron bashing, but surely you can believe he would use irrelevant benchmarks to prove his "point." That's kinda what he does ;) 

Back on topic, Shangai is more than 'just' a die shrink. It triples the L3 cache and has some other enhancements which should improve IPC quite a bit in certain applications. Whether it will be enough to take on Nehalem is questionable, I lean towards "no way in hell," but we'll see :) 
a c 127 à CPUs
a b À AMD
August 1, 2008 3:46:14 PM

Its for anything that seems like AMD bashing or those who have talked negative about AMD beofre or those like me who disagree with BM.

Where I disagree is the crap he spreads and thats all. Fun times.

Soon this site will be crap. It was so nice in the begining.
August 1, 2008 3:53:47 PM

homerdog said:
^Huh? Remember, it's Shanghai vs Nehalem; Penryn will be phasing out by the time AMD get to 45nm ;) 




From t
August 1, 2008 3:54:41 PM

speedbird said:
TC was the person who promoted the rating system. I mean it's always been there, but it was pretty much ignored before. I'm not laying all the blame at TC, but it certainly started it off I believe.

http://www.tomshardware.co.uk/forum/251608-10-declare-r...

Unfortunately it looks like TC's plan is backfiring...
jimmysmitty said:
Soon this site will be crap. It was so nice in the begining.

Here's hoping I don't get banned for agreeing with you :whistle: 
a c 127 à CPUs
a b À AMD
August 1, 2008 3:59:36 PM

^The difference was that he only wanted to use it to block out the troll posts like those by kassler at the time.

Of course you see it never went through until recently.
a c 127 à CPUs
a b À AMD
August 1, 2008 4:02:47 PM

homerdog said:
Unfortunately it looks like TC's plan is backfiring...

Here's hoping I don't get banned for agreeing with you :whistle: 


Yes TCs plan is backfiring. The trolls are down grading the ones with the links to proof.

I think the last thing I like about THG is the way they stick up for PC gamers. But the forums is on a very slippery sloap with only one ski left and heading for a cliff. I just don't want to be there when they go plunging.
August 1, 2008 4:11:11 PM

homerdog said:
^Huh? Remember, it's Shanghai vs Nehalem; Penryn will be phasing out by the time AMD get to 45nm ;) 



My keyboard is possessed. I was using Penryn as a baseline. It's(Nehalem) supposedly 20% faster and not due to HyperThreading. See Anand. Things are totally different with K10 (K10.5) in server. We've already seen the proof with this very post. Intel is 500MHz ahead in clockspeed. Well, really 700MHz based on 2356(2.3GHz).

I'm not saying Intel doesn't have a chance but remember there will be no MP Nehalem this year. I heard Q209. Shanghai also upgrades to HT3 links which will really count in heavy multithreaded loads. AMD is being hush-hush because they are emphasizing Deneb improvements. Fudo thinks it will beat Yorkfield.

A 15-20% improvement (reported) over K10 just makes the wall for Nehalem higher, especially in server loads.
      • 1 / 2
      • 2
      • Newest
!