In addition to being a photographer, electrical engineer, general knowledge junkie and sports fan, I’m also a computer geek, and frequent computer enthusiast message boards in addition to many others.
Today, Intel released their Ivy Bridge processors, and a bunch of early reviews came out.
One of them noted that at the same voltage, the new processor was running at 82C, while the older chip could run there at 56C. The reviewer noted that their heat measurements were 46% higher with the new processor.
Anyone else see the problem? Would you have a problem with that statement?
Anyway, I posted a comments saying that the new chips were not 46% hotter, but rather 7.8% hotter, as when comparing different temperatures, you really need to do it in Kelvin, since Fahrenheit and Celsius both have arbitrary 0 degree marks. I even illustrated the point by showing that if the temps were done in F, you’d have a completely different % increase in temperature…so which one would be right. I also pointed out (after getting thoroughly jumped on in later comments) that it would seem absurd to publish in their review that a new chip was 2,200% hotter if the old one was 1C and the new one was 22C (the same percentage increase in actual temperature as the one in the review), wouldn’t it?
So, given this is a computer nerd technical forum I was posting on, and about 80% of the commenters completely blasted me for being overly pedantic and how of course they should be in Celsius and 46% is right because people understand Celsius but not Kelvin. (Nevermind that wasn’t the point…the point was the 46% is a completely meaningless ratio in this discussion and flat out wrong when comparing temperatures).
So, was I being overly pedantic, and would you have a problem with what I wrote? (note that I did not ‘flame’ the reviewer…I simply pointed out the inaccuracy and provided examples to illustrate why it was misleading in civil tone).
True… In the grand scheme of things, it’s very not important. But I’m more curious about how many dopers would have a similar response to a comment on a review like that like what I received, rather than ‘Yes, you’re right.’ I know I’m right in the comment, the question is more “would you have jumped on someone who posted that?”
As in, would you be upset that I was being pedantic, or would you agree (or simply let pass) the comment?
Heck, the article author came back and said “Well, the number in C is 46% higher than the previous one, so I am right too.” If I had written the article, and someone had pointed out my mistake, I’d say "Oh…wow…I forgot I couldn’t do that, thanks. " and either use the correct numbers, or, if too technical for your audience, rework the statement to not include a percentage. But I guess that’s just me.
As someone who is familiar with entropy and enthalpy and all that thermo-chemisty stuff: Kelvins are the only way to measure heat absolutely, so I’d say you’re right. But I still might be worried about a chip that ran at 87 degrees Celsius.
After I read the way you explained it, I understood what you meant. But I think you might have more people on your side if you made that argument on a physics board rather then a computer board.
7.8% is just as meaningless anyway, until the mechanics of this specific measurement are normed in terms of relevance to the audience. I don’t know what that actual mechanics are, but 46% just might reflect the relevance of the difference more than 7.8%
So my reaction, however, might be to say, “Yes, you’re correct in that sense, but I prefer to use 46% with C to indicate how important the difference is.”
It may help to know that thermal issues are a HUGE part of the discussion on these boards. We crazy computer people overclock our processors a good bit, and lowering temps is a big part of that. There are entire sub-forums dedicated to cooling, etc.
In overall discussion, the percentage difference isn’t a super important thing to take note of, but in a serious discussion of the heat generated by the new processor architecture, I was mildly irked by the inaccuracy, but it wasn’t a big deal…I was very surprised at the reaction to the comment though.
It’s actually well within operating temperature, but it is getting a bit toasty. For those of us overclocking the latest Intel chips, the goal is to stay somewhere around 70C under extreme load…but the previous generation was OK in theory to 99C and the new one is OK in theory to 112C, but no one wants to run them there.
It’s the reason we have such insane looking heatsinks like this:
IANAScienceGuy, but I’m sympathetic toward the OP. However, I’d need some cites to believe that temperature can be compared in percentages. I’d buy that something uses 46% more energy than something else, but right now “46% hotter” seems wrong.
It’s just the kind of thing Sheldon from Big Bang Theory would go on about…and make people want to punch him for doing it. Probably accurate, utterly useless, and more than a little annoying. Nothing personal…just how I’d react to the situation.
OP is 100% unambiguously correct. To say that a statement about temperature- in an article that specifically reviews thermal performance of the CPU - is some sort of nitpicking minutia is silly. Temperatures are a big part of what they’re comparing. It would be like a car review saying that a 400 HP car has twice the horsepower of a 300 HP car, because you start your scale at 200 hp.
Where the 0 in F and C are arbitrary, and so comparing temperatures as a percentage of their value over that arbitrary 0 point leads to meaningless - even misleading statements. Is 5 degree weather 5 times hotter than 1 degree weather? Would it rub you the wrong way if a metereologist said “it’ll be 5x hotter than yesterday!” because the projected temperatures were 1 and 5 over those two days?
Hmm… This may well be why my friends think of me as a more pedantic version of Sheldon.
I do make up for it by being considerably more attractive than the actor portraying Sheldon, and I also am quite normal in my response to fair members of the opposite sex, so it is not as if I am a complete nerd.
But given that editorial aside, Jman was attempting to question the accuracy of a science-based article and his objection is quite valid in terms of how the issue would be presented in ‘Physical Review’, for example. I suppose given the general slovenly attitude of most popularizers of science when it comes to accuracy, then such a mistake could well be expected and, in fact, essentially ignored.
None of the accents in the series really bother me because they don’t come out of the real world. Who’s to say what someone from the Riverlands sounds like? They take care in the show to talk about and show that different regions have different fashions and other differences - why not a wide range of accents?
In this instance, I don’t know whether it matters or not. But, enough people make important, significant errors with percentages by being careless with the base—what they’re percentages of—that I commend the OP for trying to fight the good fight against ignorance.
Yeah, is the relevant increase really how many degrees above absolute zero, or is it something else, like how many degrees above room temperature or how many degrees above ideal operating temperature or something like that?
That’ll teach you for posting on some other message board.
That kind of pedantic nit-pickery is one of (many) things I love about the Dope. Not the correcting per se, but the accompanying explanation that linked formerly separate concepts in my mind together and added to my admittedly small pile of knowledge.
I would have a problem with what you said because it looks like you put words in the mouth of the reviewer.
They said “the heat measurements were 46% higher.”
You said “the new chips were not 46% hotter.”
But they never said they were 46% hotter. Noting the percentage increase in that particular measurement is redundant (since they gave the temperatures themselves) but not meaningless. It isn’t meaningless because the readers are familiar with what various measurements, in those particular units, mean. They can translate the statement themselves.
I haven’t read the original review, of course, so they may have been implying what you thought they were, in which case I would have responded with a Futurama quote. You can guess which one.
But even ‘the measurements are 46% higher’ is a massively misleading statement. If you convert the units to Fahrenheit, it’s only a 35% increase. The purpose of the statement was to apply a metric to determine how much hotter the new chips were running. And if you’re going to put “Whoa…46% increase!” it should at least actually mean something. If the starting point was 1C and the exact same increase in heat was observed, it would have been “whoa…2,200% higher!” If it were 0C and it went up even a single degree. “Oh my GOSH…INFINITELY HIGHER TEMPERATURE!”
