Cinebench and Passmark for starters. Again this is another place where the information is not readily available and I as the user am expected to visit every benchmark software homepage for documentation which may or may not be present when the professional review sites who should be aware of this information could disclose it for me.
When someone offers me some AMD vs Intel comparison based in biased synthetic benchmarks such as Sysmark, Cinebench, Passmark... I just say "No thanks", because I know the issue, but average Joe would benefit from your idea of a footnote disclaimer: "Sysmark and other benchmarks used in this review are biased pro Intel and don't represent real world performance".
I dont know and thats the key point. I dont have the technical expertise or hardware availability to determine all of these things. Which is why my criticism was not about ICC bias but the lack of technical reporting on the issue by mainstream sites and the laughable argument presented above that users should be able to determine this for themselves and excercise due diligence when purchasing. If you would re-read my original post for meaning perhaps you would understand.Ok, so windows synthetic benchmarks only? And how biased are those benchmarks? Because even Agner Fog recognized that not every ICC code is biased. And what about the games, mostly compiled with MSVC? What about the GCC linux benchmarks? What's the bias on these?
Still? Are you talking about the test with software anno 2005?
So what about the other compilers?
Intel compiler studio version 12.0 produces complete shit for Bulldozer....
Look at the individual Floating Point benchmarks. Nice try with the auto-
parallelization tricks....
Now Intel has to optimize their compiler for AMD's exotic architectures?
No, It's enough if they stop lying.
Hans
Now Intel has to optimize their compiler for AMD's exotic architectures?
No, It's enough if they stop lying.
Ok, so windows synthetic benchmarks only? And how biased are those benchmarks? Because even Agner Fog recognized that not every ICC code is biased. And what about the games, mostly compiled with MSVC? What about the GCC linux benchmarks? What's the bias on these?
But a score picked off of who knows where (a website you won't link?) that's apparently using Unixbench (BYTE magazine's benchmark) is as synthetic and ancient and more or less useless and problematic as they come. You should be saying "no thanks" to it too, but you don't because you're happy that it gave you the number you wanted.
I don't expect this to mean much to you, but this gives you idea of one current incarnation of the benchmark you linked: http://code.google.com/p/byte-unixbench/
The actual CPU parts of this are garbage (Dhrystone and Whetstone), while a lot of the rest of it has to do with performance of other parts of the system like HDD/SSD and GPU, making the overall score especially bad for comparing CPU performance. People like to compare scores made with different compilers too, sometimes they compare scores with current compilers against those with compilers that are 20 years old. Without knowing anything else we can't even say that ICC wasn't used, since ICC is available for Linux. Some of the tests are also very sensitive to different kernel versions. Without (much) more information we can't even begin to tell if the numbers are fair, although I can begin to tell that they're pretty irrelevant regardless.
Also not really sure Cinebench should be considered synthetic. While maybe not literally a benchmark of CINEMA 4D that's essentially what it is, and a benchmark of real software designed to do some useful task outside of test performance is not synthetic.
Now Intel has to optimize their compiler for AMD's exotic architectures?
Nobody worried about Intel having one advantage with ICC deliberately crippling AMD chips, but about Intel hiding this fact to programmers, OEMs, consumers... and denying it latter (when some scholars began to denounce ICC) until the FTC research and final settlement obligated Intel to put the disclaimer.
How exactly do you propose people exercise this due diligence when the information is not made available to them? The compiler issue was known for some time yet was rarely mentioned except much later when it was a legal issue. Then it was generally only covered as a settlement issue without technical discussion. Even today, years later there has been no real follow up technical reporting on the issue to be found on popular tech sites like Anand. How exactly are consumers supposed to reliably find this information when it is only covered on wordpress blog type sites at best and major sites ignore it.
This is not just a compiler issue or even a computing issue. It is unreasonable to expect manufacturers to disclose unflattering information about hardware and other limtiations of their products since it would negatively impact sales. When the supposed news organizations fail to cover said information for fear of losing advertising revenue or other relations, exactly what is the consumer supposed to do? As consumers we need news sites to actually report the information that the manufacturers are unwilling to publicize and when this clearly doesnt happen as in the case of Intel's compiler, it is hard to use due diligence as any argument at all.
Some of the biased benchmarks that I listed before can downgrade the real scores of non-Intel chips up to a 40% for favouring Intel chips.
GCC linux benchmarks don't cheat. Precisely AMD chips perform much better when openbenchmarks are used.
AMD knows how optimize their own architectures, 'exotic' or not. But AMD does just that: optimization. Intel cheats and lies.
No two people will perform due diligence to the same degree, but here are examples of what I do in the name of "due diligence":.......
Due diligence is not relying on others to do your due diligence for you.
Not really. Optimum code for Bulldozer isn't optimum code for Core, and we know that as GCC has no dispatcher on their code we cannot really be sure if code is being optimized more for Bulldozer or for Core. If I were to guess, and knowing the "open sauce" community, I'd guess that GCC is a little biased towards Bulldozer, but by no means as biased as ICC.
Lies? ICC is explicitly saying that it's not going to give the same performance with non-Intel processors. Where's the lie here?
Cheating? Why cheating, the compiler is performing as advertised. The mere fact that they don't treat AMD processors like Intel processors don't configure cheating. It isn't a vendor neutral compiler after all, it is an Intel compiler.
I think you don't realize how ridiculous this situation is. If Intel (or AMD) benchmarks were the most reliable way to measure CPU performance, sites like Anandtech or Techreport would have to find something else to do. It's exactly because company numbers are to be taken with a grain of salt that sites like Anand exist. Also it's because ICC isn't always the best alternative that Microsoft bothers to develop their own MSVC compiler, or that the open sauce community develops things like GCC or the Open64. To never trust a company benchmark should be a norm when evaluating a product.
The only company information you should trust is information that is covered by a SEC filling, and not because the company will lose credibility with investors, but because the officer can answer in person for whatever misleading information they give in the filling or in the given event. Remember John Fruehe. His lies started to be dismantled by Seifert himself, when he had to answer about Bulldozer performance predictions.
Nonsense.
This is the third time that you ignore the relevant information given to you. :whiste:
Cheating? Why cheating, the compiler is performing as advertised. The mere fact that they don't treat AMD processors like Intel processors don't configure cheating. It isn't a vendor neutral compiler after all, it is an Intel compiler.
You are mixing two different things and doing a mess: discussing if one given fair benchmark is adequate or not for measuring performance of any chip is very different than saying "no thanks" to benchmark biased towards Intel.
Moreover you omit that I did not take unixbench scores as final. At contrary I said that I was waiting for more benchmarks of jaguar.
Had Intel not been found guilty during the antitrust cases someone like mrmt & a few others would have just said that "since it has not been proven in a court of law Intel is not guilty" so AMD/fanbois stop your whining ! In essence what he's said that Intel needed a court order to divulge some basic info regarding the shady stuff with ICC, I mean how messed up is that, & Intel has still not admitted publicly on record about their OEM bribes after their settlement with AMD. So "screw you consumers because I can sink AMD anytime I want to & I'll continue ripping you off any way & every way known to mankind" because that's what Intel's attitude looks like to me !The big caveat is that Intel was forced to inform ICC customers about it's very strange way of deciding instruction set support when the FTC finally decided to look into the complaints from AMD, Via and Nvidia. Several of Intel's behaviors are at least up there with Microsoft ones, imo, Intel got off lightly (at least in the US) partly due to the investigation being done after the financial bubble popped and with the US trade concern heightened by the large fines the EU has levied against MS and Intel.
The general public still doesn't know about it, nor do most of'em care frankly but it's a PR disaster hence Intel never admitted the charges publicly & seemingly they still have a halo over their heads :hmm:I think that's a pretty safe assumption R0H1T, considering there were still denials of paying off OEMs even after information such as this: http://www.guardian.co.uk/business/2009/nov/04/intel-bribed-for-bribery-coercion was available.
"Cuomo published internal documents from Intel's executive suite as evidence for his accusations, including a 2005 email exchange between Intel's chief executive, Paul Otellini, and Dell's chief executive, Michael Dell.
In the emails, Dell complained of a loss of "performance leadership" through the use of Intel's microprocessors and appeared to be contemplating alternative suppliers. Otellini replied: "We are transferring over $1bn a year to Dell for meet comp efforts [sic]. This was judged by your team to be more than sufficient to compensate for the competitive issues.""
No that isn't just "acceptable business practices" unless you are of the opinion the RAM price fixing and LCD panel price fixing should have been allowed to continue. The sudden drop in monitor and TV prices wasn't due to industry benevolence.
Yes & this why I mentioned admission of guilt because there will inevitably be a consumer backlash, also it could be used as a point of reference in future cases, eventually leading to the erosion of their top & bottom lines upto a certain extent.That's why I mentioned RAM and LCD price fixing. The general public doesn't really know about it but I'm sure they'd be upset to see those products retail pricing back at the "fixed" level.
Due diligence is investigation before purchase. What you claim is your "due diligence" is actually self testing and analysis after purchase. Personal discovery and testing after purchase is not due diligence at all, due diligence relates to pre-purchase knowledge. Your due diligence would make tech news sites obsolete as you would be doing your own reviews and testing. A process which requires significant equipment and technical knowledge not available to most users. Then again maybe you are correct that as a user I cannot count any of the worlds top tech sites to actually cover certain information and I should expect to discover it for myself.