AMD Takes Award for Best Server Processor

H

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
AMD today announced that its AMD Opteron 6200 Series processor family, based on the new "Bulldozer" architecture, was chosen as the winner of The Linley Group's first annual Analysts' Choice Awards. The awards recognize the top processor products of 2011 in several major categories including embedded, mobile, PC, server, design IP and related technology. To choose each winner, the Linley Group's team of technology analysts focused on merits of the leading products that began shipping between January 1, 2011 and December 31, 2011.
 
stiltner said:
How?
Click to expand...

Probably won for Best long delayed AMD Server CPU introduced in the second half of 2011, either that of only fanboys voted. :D

It's a bit like Al Gore winning the Nobel Peace Prize in 2007 (apparently for having the largest personal carbon footprint of any American but telling the rest of us to behave).
 
I am a fanboy or "fangirl" I admit but it's also just because...well is there really any other competition? :p tehe
 
Because the interlogos processor is 90% of the performance of the Xeon, and it cost 40% less than the Xeons..
 
AMD just announced today that it won the prestigious FIRST ANNUAL Analyst's Choice Awards sponsored by AMD for it's award winning Bulldozer design according to an AMD spokesperson from the AMD corporate office. AMD then congratulated itself for such an outstanding achievement according to AMD officials.
 
The only reason I read these threads is because it's entertaining watching people argue about stupid bullshit.
 
Oh poor AMD. There is no such thing as "First Annual" in news articles, according to the AP. Maybe it's time for AMD to fire its PR department.
 
RealBeast said:
Probably won for Best long delayed AMD Server CPU introduced in the second half of 2011, either that of only fanboys voted. :D

It's a bit like Al Gore winning the Nobel Peace Prize in 2007 (apparently for having the largest personal carbon footprint of any American but telling the rest of us to behave).
Click to expand...

Or like Obama wining the Nobel Peace Prize before he actually did anything :)

(on second thought, if they had waited until he had a record to look at, even the lefties on the Nobel committee wouldn't have been able to stomach giving him the prize)
 
Benny Blanco said:
How you doin? You sound hot.
Click to expand...
Abit OT from me, but still, I'm not even wondering why technologies are still taboo to most women, when all the men in this field act like retards.

Back to topic: We are congratulating to AMD and sending flowers.;)
 
nutzo said:
Or like Obama wining the Nobel Peace Prize before he actually did anything :)

(on second thought, if they had waited until he had a record to look at, even the lefties on the Nobel committee wouldn't have been able to stomach giving him the prize)
Click to expand...

lol glad to see i wasn't the only one thinking this as well when i saw the al gore post further up.
 
walwalka said:
Because the interlogos processor is 90% of the performance of the Xeon, and it cost 40% less than the Xeons..
Click to expand...

If you look at SPEC benchmarks, which are traditionally the ones that are used to judge CPU performance, you would see that we are actually signifcantly faster than Intel:

SPEC Int 26% faster:
543 for AMD http://www.spec.org/cpu2006/results/res2011q4/cpu2006-20111024-18714.html
432 for Intel http://www.spec.org/cpu2006/results/res2012q1/cpu2006-20111219-19272.html

SPEC FP 44% faster:
403 for AMD http://www.spec.org/cpu2006/results/res2011q4/cpu2006-20111024-18715.html
278 for Intel http://www.spec.org/cpu2006/results/res2012q1/cpu2006-20111219-19195.html



TheGardenTool said:
Oh poor AMD. There is no such thing as "First Annual" in news articles, according to the AP. Maybe it's time for AMD to fire its PR department.
Click to expand...


an·nu·al [an-yoo-uhl]
adjective
1. of, for, or pertaining to a year; yearly: annual salary.
2. occurring or returning once a year: an annual celebration.
3. Botany . living only one growing season, as beans or corn.
4. performed or executed during a year: the annual course of the sun.
5. Entomology . living or lasting but one season or year, as certain insects or colonies of insects.

"Annual" implies that this will happen once a year, not that it has happened in the past. So if you are planning to do something annually, you ae doing it once a year, even if this is the first time. If you schedule a weekly conference call, it is a weekly call, even on the first day.

While you are correct that the AP style book prefers "innaugural" over "first annual", if you want to take on that issue you really need to point to Linley: http://www.linleygroup.com/press_de...ers-of-First-Annual-Analysts-Choice-Awards-50

We were going off of the headline that they used, part of that is that anything you do in the corporate world is tied to legal reviews and using the same phrasing helps cut down on the review cycles. If you think arguing "first annual" vs. "innaugural" is a pain on the internet, try it with a lawyer some time.


Also, this is the second "processor of the year" award that we received, the other was from InfoWorld: http://blogs.amd.com/work/2012/01/10/technology-of-the-year/
 
Rather interesting that you compare a 32-core AMD system to a 12-core Intel system... I also thought that lower power consumption is favorable for server chips and I highly doubt those two systems are comparable regarding their processor setups.

I am NOT one of the people that constantly bash AMD's new architecture, but I don't see how that is a fair comparison.
 
It's a performance comparison. I chose my highest performance and their highest performance.

From a power perspective we are about the same in power.

From a cost perspective, thiers are ~$1600, ours are ~$1000.

If you wanted to do a dollar for dollar comparison, we did that at launch. We took their top selling processor, the E5640 which is ~$800, and compared our $800 processor, which was 89% faster.

For the comparisons I laid out here I was just looking at raw performance, but if anyone wants to go down the price/performance or performance/watt path, that is even more of a lead for AMD; I actually chose the most conservative comparison for CPUs.
 
Just because there are no results for Intel chips with more cores does not mean they do not exist. Intel is shipping 8 and 10 core chips. I would think a dual-CPU system with Xeons with more than 6 cores would be a fair fight. Otherwise, I just see that the AMD hardware received its rating first. I don't want to get into a big argument over this so I will just leave it at that.
 
The cost-to-performance-to-core is actually pretty good when compared with similar Intel setups so long as it's optimized it gets better, but it's the comparison to the older 61xx server chips where there isn't a big stride forward outside of instruction sets and encryption. AMD did well with 61xx, they just didn't improve much on it with Interlagos unless you recompile, but telling people they need to do more work to see the benefits isn't taken lightly. In fact it's often ignored completely.

http://www.anandtech.com/show/5058/amds-opteron-interlagos-6200/14

That is the good news. We cannot help but to feel a bit disappointed too. AMD promised us (in 2009/2010) that the Opteron 6200 would be significantly faster than the 6100: "unprecedented server performance gains". That is somewhat the case if you recompile your software with the latest and greatest optimized compiler as AMD's own SPEC CINT (+19%), CFP 2006 (+11%) and Linpack benchmarks (+32%) show.

One of the real advantages of a new processor architecture (prime examples where the K7 and K8) is if it performs well in older software too, without requiring a recompile. For some people of the HPC world, recompiling is acceptable and common, but for everybody else (that is probably >95% of the market!), it's best if existing binaries run faster. Administrators generally are not going to upgrade and recompile their software just to make better use of a new server CPU. Hopefully AMD's engineers have been looking into improving the legacy software performance of their latest chip the last few months, because it could use some help.
Click to expand...

There's a big difference between real world environments and company run setups and demonstrations meant to showcase products. Generally, any and every advantage will be used to detract from the competitor and show their own product in a positive light, so don't take an AMD presentation and consider it bearing the merit of an Anandtech review by Johan. Or JF's words as truth. I think we've all learned that by now
 
charlie-sheen-winning-resized-600.jpg


:D
 
This is business. Cores per rack space, and compute power per dollar are key items. Power consumption is not nearly the issue you might want to believe. I believe that Bulldozer will not give as much to the server market as previous processors, but don't ever think that AMD has not absolutely trounced Intel for quite a while in that space, from a specifications standpoint. Brand loyalty would be the only reasoning behind most intel purchases in the server space for a few years now. Possibly anti-competitive back-room dealings, but those don't count because there's no proof.

AMD has nothing to gain in terms of value in the server space. They stand to lose that lead if Intel comes out with a compelling processor. The low power game is already lost by AMD, is my guess. Now it's down to mainaining the compute per rack value.

If Intel rolls out an 8-way 8 core server, and AMD only has a 6 way 8 core server, you can bank that Intel will have the best core count crown.
Comparing the best offerings in a price range is completely legit. I am sure that if you look around, Intel has a few awards geared towards their offerings.
 
pelo said:
The cost-to-performance-to-core is actually pretty good when compared with similar Intel setups so long as it's optimized it gets better, but it's the comparison to the older 61xx server chips where there isn't a big stride forward outside of instruction sets and encryption. AMD did well with 61xx, they just didn't improve much on it with Interlagos unless you recompile, but telling people they need to do more work to see the benefits isn't taken lightly. In fact it's often ignored completely.

http://www.anandtech.com/show/5058/amds-opteron-interlagos-6200/14



There's a big difference between real world environments and company run setups and demonstrations meant to showcase products. Generally, any and every advantage will be used to detract from the competitor and show their own product in a positive light, so don't take an AMD presentation and consider it bearing the merit of an Anandtech review by Johan. Or JF's words as truth. I think we've all learned that by now
Click to expand...

Telling people to recompile is exactly what happens every time Intel adds new instructions. "Recompile or get a little less", more specifically.

Sounds like AMD is just following industry standards.
 
Solomutt said:
This is business. Cores per rack space, and compute power per dollar are key items. Power consumption is not nearly the issue you might want to believe. I believe that Bulldozer will not give as much to the server market as previous processors, but don't ever think that AMD has not absolutely trounced Intel for quite a while in that space, from a specifications standpoint. Brand loyalty would be the only reasoning behind most intel purchases in the server space for a few years now. Possibly anti-competitive back-room dealings, but those don't count because there's no proof.

AMD has nothing to gain in terms of value in the server space. They stand to lose that lead if Intel comes out with a compelling processor. The low power game is already lost by AMD, is my guess. Now it's down to mainaining the compute per rack value.

If Intel rolls out an 8-way 8 core server, and AMD only has a 6 way 8 core server, you can bank that Intel will have the best core count crown.
Comparing the best offerings in a price range is completely legit. I am sure that if you look around, Intel has a few awards geared towards their offerings.
Click to expand...

If that power brings along extra heat with it (and it usually does) then yes that is a big issue. Doesn't matter where the costs come from hardware, software, hvac, power, etc. if you have to stay within a budget.
 
Last edited:
Good for AMD!

And here we go with the nerd rage. I bet if a story came out that AMD discovered a cure for cancer, it would take about 3 posts before you'd get how much AMD sucks. I don't get it. Yeah Bulldozer was a disappointment. So what? Should AMD just close it doors and all their employees hang themselves? Jesus! I'm going to buy a FX-8120 just out of spite.
 
Solomutt said:
Telling people to recompile is exactly what happens every time Intel adds new instructions. "Recompile or get a little less", more specifically.

Sounds like AMD is just following industry standards.
Click to expand...

Recompiling doesn't happen unless they're forced to and it's not something that's done on a whim. It doesn't take 5 minutes and it generally equates to more work being done for an obvious return in performance. It's not that they're asking people to recompile to see the best out of that server chip, that's the case with every server chip ever released, but that it doesn't provide a significant performance boost UNLESS the software is recompiled, and that's where it fails.

They are cheaper, though. And although it's nowhere near the +50% throughput that was originally promised it's still offers a viable alternative if starting fresh. The prices on those things are fantastic.

I wouldn't pin that same OEM-fiasco on Intel this time around despite holding 90% server market share. Back then they trailed in performance, now that's not the case anymore. You don't get many platform upgrades in the server space, certainly not like the desktop, so chipping away at Intel's stranglehold will take a long time
 
MacLeod said:
Jesus! I'm going to buy a FX-8120 just out of spite.
Click to expand...

Come now, come now, let's not do anything drastic or do something you'll soon regret.

Nobody wants all AMD employees to hang themselves, thats never been put on the table. Just the CPU design wing and marketing people and managers. The graphics department and other branches are fine with most people. Oh but talking of the graphics department, that person who makes the "ruby" woman who looks like a dude...think they better go. :p

I don't think you can choose a "best" server processor. Theres a million applications. If you wanted something with higher CPU power or IPC, or something lower wattage, you'd go for a xeon. But there are going to be some things where the bulldozer CPUs are a better fit.
 
P4Power said:
Just because there are no results for Intel chips with more cores does not mean they do not exist. Intel is shipping 8 and 10 core chips. I would think a dual-CPU system with Xeons with more than 6 cores would be a fair fight. Otherwise, I just see that the AMD hardware received its rating first. I don't want to get into a big argument over this so I will just leave it at that.
Click to expand...

You are absolutely right.

Intel has an E7 2870 which is a 2-way 10 core, 20 thread processor.

The performance is almost identical to AMD's 16-core 16 thread Opteron 6282.

However AMD is priced at ~$1000. Intel is priced at ~$4200.

Do you really believe that is a fair comparison? That is why I chose the 5690. Remember that the 2P market is ~75% of the total server market.
 
P4Power said:
Just because there are no results for Intel chips with more cores does not mean they do not exist. Intel is shipping 8 and 10 core chips.
Click to expand...
I think when comparing similar server prices to AMD's star performer, JF's comparison is fair, at least right now. That value proposition will significantly change later this quarter when 8c Sandy Bridge-E Xeons are released.

And it's a little sad that 12 Westmere cores (Westmere-EP is nearly 2 years old) aren't that far behind 32 Bulldozer cores.

Anyways, to compare the best Intel DP SPEC scores against AMD's best DP SPEC scores, it probably should be against this system (I know the MP Westmere-EX based E7 is a more expensive CPU):

Bull SAS NovaScale R480 F2 (Intel Xeon E7-4870, 2.40 GHz) 20c 2s 10c/processor
SPEC CPU2006 rate 553 (AMD 32c BD system is slightly slower)
SPEC FPU2006 rate 378 (AMD 32c BD system is 7% faster)

Maybe not too impressive, but that E7 system runs at significantly slower clock speed than the X5690, a full GHz slower. 8c SNB-E should easily erase that difference.
 
But you are ignoring the prices. At $4200, they are more than 4X the cost of the AMD processor. Would you pay $8,400 for 2 CPUs when for much less you can buy the whole AMD server?
 
JF-AMD said:
But you are ignoring the prices. At $4200, they are more than 4X the cost of the AMD processor. Would you pay $8,400 for 2 CPUs when for much less you can buy the whole AMD server?
Click to expand...

Because he is an Intel fanboy. I mean, look at his username.
 
You must log in or register to reply here.
Back
Top