backtop


Print 24 comment(s) - last by Fritzr.. on Aug 2 at 8:43 PM


AMD "Barcelona" 2.6 GHz "simulated" performance projections. (Source: AMD, Alexei Nechuyatov)

Dual-socket AMD "Barcelona" 2.3 GHz projections. (Source: AMD, Alexei Nechuyatov)

Single-socket AMD "Barcelona" 2.3 GHz projections. (Source: AMD, Alexei Nechuyatov)
In this day and age we don't even need silicon to benchmark our processors -- some of us anyway

Last April, benchmarks of AMD’s Barcelona running at 2.6 GHz made rounds across Internet tech sites. These benchmarks, according to AMD, provided a 21% performance gain over quad-core Xeon 5355 in SPECint_rate2006 and a 50% gain in performance over Xeon 5355 in SPECfp_rate2006.

This looks pretty phenomenal until you read the footnote courtesy of AMD: "Estimated performance @ 2.6GHz based on internal AMD simulations."

Much to my surprise, these benchmarks surfaced again over the last few days.  This time, however, the author did not tell you the benchmarks were simulated and at least four months old. He did not tell you that the benchmarks ran were presented to him on a PowerPoint document, and he did not tell you the numbers featured in his "review" were of a 2.6 GHz simulated chip instead of a 2.3 GHz simulated chip.

In fact, AMD publically denied the Barcelona would top out at anything over 2.0 GHz at launch when it does launch on August 27

So what's going on here?  Shoddy journalism and even the author wouldn't deny that. Any author willing to pad his byline with such omissions of fact would almost certainly have an alternate agenda brewing.

I didn't give these benchmarks much thought when I first saw them in February, and given the confirmed top-out frequency from AMD, I certainly don't give these benchmarks traction now either.  I suggest those with interests in Barcelona wait until someone tests actual DVT or Retail silicon.

The funny thing is I cannot recall the term "simulated benchmarks" in my pre-Barcelona lexicon.  Since when did benchmarking a 1.6 GHz processor give the "simulated" performance of a 2.3 and 2.6 GHz chip?  Does anyone find it odd that Googling for "simulated benchmarks" yields only 574 entries -- all from AMD, Microsoft and Sanda.  AMD holds the number one spot.

Maybe simulated benchmarks will go down with some of the other great marketing terms of the last decade: FUD, paper launches and ship dates.



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

By ScythedBlade on 7/4/2007 2:43:51 PM , Rating: 2
Yea, I think theo's relatively new ... but he does say they are PAPER benchmarks ... and theoretically calculated what COULD possibly happen.




By KristopherKubicki (blog) on 7/4/2007 2:51:29 PM , Rating: 3
Stating it, in passing, in the last sentence of the article doesn't really discount the other errors mentioned.


By Dianoda on 7/4/2007 3:20:35 PM , Rating: 2
Well, the INQ is hardly known for the quality of the information posted there. If they think they have a story, they post it, regardless of the source/reliability of said information.

Remember the story during Computex about Intel licensing SLI from NVidia for the upcoming X38 chipsets? There is no way that story turns out to be true (while it is almost certainly technically possible for the X38, officially it's not gonna happen).


By ScythedBlade on 7/4/2007 5:36:52 PM , Rating: 3
Dianoda ... wrong ... Intel's Chipset is going to have SLI.

Anyway, here's the thing about the Inq. It's actually RIGHT most of the time. However, you have look for keywords ... they usually say its speculation if its "paper" ... hence, we can't discredit Inq here ... and you have to look for the authors. Remember the author who was whamming Conroe and all those false information? That guy was Fudo ... he was fired and now runs his own site: fudzilla.com ... the same person who got fake barcelona benchmarks.

The Inq posts everything ... stories and opinions. But they have a 97% of being right ...


By GaryJohnson on 7/4/2007 6:37:10 PM , Rating: 2
quote:
It's actually RIGHT most of the time.


It doesn't really matter if they're right most of the time if they're wrong some of the time. It takes countless good stories to counter the credibility lost because of one bad one.


By ScythedBlade on 7/4/2007 7:49:56 PM , Rating: 2
Or firing. In this case, the fired Fudo no longer is on TheInquirer ...

The reason I read theinquirer is usually because its the fastest site. Heck, it's why there are always tons of people looking at spoiler pics and summaries on mangahelpers.com ...


By danrien on 7/5/2007 9:34:34 AM , Rating: 3
i read the inq daily just for a good laugh. their writers have some of the greatest dry humor out there, unlike most tech sites, which take a pretty serious take on the whole deal.


By emboss on 7/6/2007 12:24:31 AM , Rating: 3
The benchmarks are probably pretty accurate as to how a 2.6 GHz (current silicon) Barcelona would perform. For example, replacing the right crystal(s) in a system with ones operating at half the frequency, then you'd get the same benchmark results (as long as the benchmarks didn't depend significantly on HDD performance). However, you could use double the multiplier in the CPU, and so get accurate results for a chip clocked at twice the speed. Not saying his is how AMD has done it (they may have just made them up) but it's certainly possible to get accurate simulated numbers.

Of course, the deceptive thing here is (as noted) that Barcelona isn't going to be hitting 2.6 GHz any time soon, and they weren't comparing against the fastest Xeons.


This isn't the worst of it....
By JumpingJack on 7/4/2007 11:33:25 PM , Rating: 5
Xbitlabs picked up on these benchmarks:

http://www.xbitlabs.com/news/cpu/display/200707022...
Title of the piece: "AMD Shows Off More Quad-Core Server Processor Benchmark Results." and a tag line: "AMD Demos More Barcelona Test Results" ....
Xbit contradicts itself, and in fact, in text states projected performance.... but that is not what a casual reader would take away from the byline.

TechReport also parrots Theo:
http://techreport.com/onearticle.x/12799
Title: "Barcelona synthetic benchmark results bode well"

Definition of Result:
"to spring, arise, or proceed as a consequence of actions, circumstances, premises, etc.; be the outcome. "
http://dictionary.reference.com/browse/Result

Xbit, Theo, and now the trusted TechReport are culpable of spreading inaccurate and false information, the credibility of these sites just took another step down in my book.

This is truly getting out of hand... regergitation of old Feb. 2007 slides as new actually benchmarks... it is becoming almost impossible to know what 'journalist' to trust.




RE: This isn't the worst of it....
By Master Kenobi (blog) on 7/5/2007 10:32:25 AM , Rating: 2
Techreport issued a link to this article. But I would also like to point out that DailyTech tends to be better on the journalism. Although us bloggers here like to rant about certain things :)


RE: This isn't the worst of it....
By Fritzr on 8/2/2007 8:43:37 PM , Rating: 2
TechReport started out fine ... the Synthetic in the title correctly says these are simulated. But the article is written as if the results are taken from real world CPUs running at these freqs.


By SandmanWN on 7/5/2007 2:36:51 PM , Rating: 2
quote:
it is becoming almost impossible to know what 'journalist' to trust.

Don't trust people. Trust No One.

Every company spins its products, all of them. News and Journalism are no different. They spin to the tune of the company backing them.

Figured you would know that by now. :)


By Performance Fanboi on 7/7/2007 11:58:30 PM , Rating: 3
Exactly how is Xbit being missleading? Are you telling me people read the title but can't hang in past the first 30 or so words?

From the 4th paragraph:Unfortunately, Barcelona’s test results are “estimated performance” of the chip at 2.60GHz “based on internal AMD emulations”, whereas the first quad-core processors from AMD due to be available in September will only hit frequencies of up to 2.0GHz. Meanwhile, Intel Corp. plans to release its quad-core Intel Xeon X5365 (3.0GHz) chip towards September as well, whereas no quad-core AMD Opteron products with higher than 2.0GHz clock-speeds are projected until Q4 2007.

I was not mislead (and doubt anyone who read the article was)and you took your quotes out of context.


Theo "Reverse Hyperthreading" Valich
By jhtrico1850 on 7/4/2007 10:22:33 PM , Rating: 4
By ScythedBlade on 7/4/2007 10:28:03 PM , Rating: 2
Thanks dude ... GUYS, WE HAVE OUR SECOND FUDO!!! (aka, false author)

It's Theo .. so don't trust the AMD fanboy called Theo ....


By AvidDailyTechie on 7/5/2007 8:04:34 PM , Rating: 2
WOW

This is the reason why I have to read so many damned news sites or review sites about the exact same product because so many people favor bias over what’s actually the superior product...

Aren’t they aware that they're consumers too?


Simulate by deinition is............
By gudodayn on 7/4/2007 9:47:53 PM , Rating: 1
Simulate by deinition is "to give or assume the appearance or effect of often with the intent to deceive"!
Now, I have nothing against AMD. I have a 4200x2 OCed to 2.9GHz running at home and a E6300 as well. They are both more than capable with the things I do!

But this simulated benchmark thing with AMD......what is it's point?

1. Is it saying that "it is" faster than Intel's CPU or AMD will refund out money 10 fold?
AMD's ans.: No, not deinitely because it is simulated.

2. Is it saying that the CPU exists and simulations were ran according to specs.?
AMD's ans.: We are still making final tweaks to it so no comment at this stage!

May I ask AMD to give us a benchmark of an actual K10 CPU with it's detail specifications? Why be so secretive about it's detail specifications? Are you (AMD) not planning to sell the CPU in the future? You (AMD) do know once it's sold on the market, the specs. gotta go with the product to the public, right?




RE: Simulate by deinition is............
By Armorize on 7/4/07, Rating: 0
RE: Simulate by deinition is............
By SandmanWN on 7/6/2007 9:40:21 AM , Rating: 3
tisk tisk on you. The tech industry is run by the numbers, raw data, and analysis. It doesn't need to be less conservative. Thats the problem with the Inquirer. They are so liberal that it doesn't matter if its false or not. They will run with it anyway and deal with the repercussions later.

Bad reporting by lazy people (THG, INQ, etc) is confusing even the most knowledgeable person.


By wordsworm on 7/19/2007 8:26:56 AM , Rating: 2
I think that they're trying to show that they're still a valid competitor. "Don't worry, AMD investors, we've got some great hardware on its way." I think the age of AMDs revival is near. I don't think they're at any risk of going under. With the ATI acquisition getting settled, I think the world is going to start noticing some hard core competition for both the CPU and GPU market. If I had some money to put into stocks, I'd go buy AMD right now, not because of these simulated benchmarks either. I think their latest releases, like the new BE chip (of which I'll have at home in about 48 hours), are some great performers.

A lot of people have been slamming AMD products unfairly. Although they don't stack up to Intel high end at the moment, they are still extremely attractive at low~lower middle range. I bought the BE because of it's low consumption and it's power as a capable dual core 64 bit CPU.


It's even worse!!!!!!!!!!!!!!!
By casket on 7/6/2007 1:53:06 AM , Rating: 3
Look at the slides carefully. They say "2p". These are not only simulated, but 2 processors (8 cores).

There is no mention of how a single processor stacks up.




By SandmanWN on 7/6/2007 9:34:15 AM , Rating: 2
Given that its Barcelona and intended for the server market it makes sense that its two sockets. The two socket server is the most popular server setup in the industry.

Probably see "simulated" single socket setups when the desktop line gets closer to release.


"The whole principle [of censorship] is wrong. It's like demanding that grown men live on skim milk because the baby can't have steak." -- Robert Heinlein

















botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki