Intel accused of inflating CPU benchmark results::SPEC says Intel’s Xeon processors were using a compiler that artificially inflated the results of its industrial benchmark by as much as 9%.
You must log in or # to comment.
Oh, I thought that was assumed, because Intel always does that. At least I noticed it very clearly already back in the 90’s.
What CPU did you notice that on in the 90’s and how did you notice it?
I remember specifically P4 was vastly overrated by Intel, where Athlon was actually generally faster at math, Intel used specially compiled code to show P4 was faster.
But there were several examples where Intel cheated, because Athlon was way better than P4.
I even went so far as to make my own test program, to see what the actual speeds were. I was an IT consultant, and was frustrated that customers were convinced Intel P4 was better than Athlon. And they also claimed Intel was more compatible, which was so annoying, since Athlon was actually more backwards compatible than P4.I as not expecting to read about socket A tonight but I still have a special place in my heart for my old athlon. I still have the CPU somewhere in my basement, I had that sucker over clocked for years and it played a lot of WoW on it.
To be fair google ultron alwways ran more efficiently on Pentium 4s
Yea, I suppose NASA optimizes for Intel. 😋
I’m not going to say you’re wrong but I’m surprised. It’s like we lived in alternate universes.
Not that I think Intel wouldn’t cheat, because they’ve showed what they’re “capable of” time after time, but what I remember about Athlon vs P4 was that there was something about heat and wattage specified by AMD that was criticized heavily. Athlon was also not something you’d choose for overclock because of this.
I just googled a little and there didn’t seem to be a trace of any controversy around P4 and tailormade compiler.
Oh boy the Tom’s Hardware “scandal”. That story was 100% planted by Intel. I think it was the K6, If the cooler dropped off, entirely the AMD thermal safety didn’t react quickly enough. In my 40 years in IT I have NEVER heard about a cooler falling off the socket even once, except for that paid for cesspool of shit article.
That story together with the 180° they did on RAMBUS to favor P4 have made me NEVER use TOM’s Hardware since. It was 100% dishonest paid for shilling. Either that or so idiotic it’s not worth reading either way. Even you mentioning that now about 30 years later, it still pisses me off. 🤬 🤬 🤬
Never heard about heat problems with Athlon, and P3 and P4 weren’t great overclockers either. Celeron was great, because you could up FSB 50%, Which made it actually faster than the top P3.EDIT PS:
No there were no journalists that covered/revealed ANY of Intels shenanigans at the time. The entire industry seemed to be in an Intel Vacuum.
But entusiasts all knew that Athlon was way better than P4.EDIT2:
There was also the issue that the Intel compiler had zero optimizations for any AMD CPU, but optimized heavily for P4. That was a general thing that Intel didn’t even hide.So the journalists are still covering for Intel to this day 15 years later, but the enthusiasts know the truth?
I’m still not saying you’re wrong but you have to admit it’s kinda strange a quick Google doesn’t reveal anything?
So considering Intel used illegal means to keep AMD out, and was fined a billion dollars for it. You think they didn’t touch gray areas in their marketing too? You seem to have no knowledge of what went on in IT in the 90’s, and be very naive, or maybe just a shill or a troll.
What in “Not that I think Intel wouldn’t cheat, because they’ve showed what they’re “capable of” time after time…” was unclear for you? Your conclusion from the above statement is that I’m probably a shill or just doesn’t know IT as good as you because I don’t agree with you on the spot? Really? Is that what you using your full mental capacity was able to conclude from my statement?
Let me just clear up it a little for you: The issue isn’t that Intel was (is) an asshole, it’s that you blurb out unsubstantiated claims and when called out on it you claim that there’s a conspiracy lead by Intel, all the journalists are on it but we all should take your word for it.
We both have all the information in the world literally in our hands and still you are unable to link to facts that support your statement and that is my fault somehow?
Nice talking to you. Have an awesome weekend. I will.
So they put out an article claiming that the thermal safety was defective, and the thermal safety was defective, and you see that as some grandiose conspiracy perpetrated by Intel? And you’re still upset about it?
Even if Intel did discover and publish the defect, what exactly did they do wrong? I would reasonably expect AMD and Intel to be testing each other’s hardware constantly. Would you have preferred that Intel didn’t publish their findings?
The thermal safety was not defective, only a few years prior thermal safety wasn’t even available. The article created an artificial situation that never occurs in reality, and claimed the CPU should be able to handle that.
The CPU handled a fan suddenly cutting off just fine, it handled being turned on without a cooler just fine. Only if the CPU was running full throttle, and the cooling block “fell off” suddenly and completely, the throttle wasn’t fast enough.
When did you ever hear about that actually happening?
So you’re saying that the CPU burning out when the cooler is removed, is the thermal safety working as intended? Sorry, I am not familiar with the situation, but the way you initially described the issue doesn’t sound like foul play.
Edit: y’all are simping for a $300 billion dollar company rn lol
There absolutely was. Intel got smacked on the wrist for doing their benchmarks using ICC… you know, the compiler that builds code that detects that it’s not running on an Intel CPU and disables all optimisations and extended instruction sets (like say MMX/SSE).
Let me repeat myself: I’m not defending Intel in any way. I’ve seen a lot of shit and murky business practices going down since the beginning of the 90’s when I started my career in IT, so I have no trouble believing that Intel did it again and again.
However, it’s not that hard to back claims you make with facts when on the internet. Normally ou can link whatever with two clicks maximum.
Still no one seems to want to help me read up on Pentium 4 and Intel cheating.
The ICC “optimization”, as far as I remember it, was related to the Xeon line of processors. If it was P4 related, please link so that I can read up on it.
[General Rant and not about you]I really don’t understand why it’s so hard for people to post links that back their claims. If you post a link you give people the opportunity to learn. If you don’t, it’s like you just want to be right and nothing else.
it’s not that hard to back claims you make with facts when on the internet.
False, when it’s 30 years or more back, it absolutely is, even things I remember from just 20 years ago can be hard to find. Sometimes it disappears other times it’s drowned out by similar stories that are newer.
You are completely delusional about the efficiency of search engines, and the memory of the Internet, not every thing stays up even for just a few years.
Also it may lack context of things that were common knowledge of the time.The ICC “optimization”, as far as I remember it, was related to the Xeon line of processors.
Maybe that too, but it’s still correct as @Kangie wrote, it was 100% also used against AMD on consumer products.
Intel used every trick in the book, even when they weren’t quite legal, because P4 was a shitty product, and they couldn’t compete on merits with it. They even tried to revive P3, but it failed above a certain clock speed, 1.13 Ghz if I remember correctly.
Oh AMD was definitely a cooking stove, but get a decent heat sink and paste and you are good to go.
Everyone does it, it’s basically expected at this point that any manufacturer will announce that their new chip will run an infinite loop in fewer microseconds than the neighbouring one would.
Ah yes, the Snake oil merchants tried to sell snake oil…again
Again?
But I suppose it’s not that different to when they rigged an industrial phase-change cooler onto one of their CPUs, then pretended it could perform like an AMD CPU lol
Well, it could*!
*with an industrial cooling system
Wait, people trust corporate benchmarks?
Most of humanity is pretty clueless about IT in general and are just used to trusting who they buy from with zero critical thought or research.
Again?!
i will only care if they ever see consequences for it.
If a chip has instructions to make certain tasks go faster, I’d say it’s fair to recompile the benchmark to use them. We didn’t stop adding new things after the 486.
Are amd, nvidia and Intel not accused of this every generation?
It’s dieselgate for processors.
Very mixed on this. If the compiler is for example using stuff like constexpr in C++ to do major calculations at compile time instead of at runtime then yeah its cheating.
But on the other hand if its doing some Microarchitecture specific optmizations like reordering instructions or replacing certain instructions with others - as long as these are available to the public its fair game i.m.o…
The article states that the compilation was tailored for this specific task, which means it doesn’t represent how the CPU would normally perform it. So it’s definitely cheating.
A benchmark is not supposed to be a compiler optimization competition. If they showed both, and revealed it was optimized, it would be another matter.Even if the compiler was available to the public most software doesn’t use it, so the benchmark is still not representative of real world performance.
