4,580 Posts served

11,094 Conversations started

I don't know why I let this stuff bug me

By Clay Breshears (Intel) (196 posts) on February 9, 2009 at 12:54 pm

This rant is in no way intended to reflect on the professors, researchers or other staff at Rice University.  Well, maybe the person who wrote the online news story about the introduction of new PCMOS (probabilistic CMOS) technology.  The first sentence of the story ends with:

"...scientists have created a microchip that uses 30 times less electricity while running seven times faster than today's best technology."
 

I'm excited by the claim of being able to run 7 times faster.  That's great!  However, I am a bit confused about the claim of power savings by the technology.  Specifically, how does one get something that requires "30 times less" of some other thing.

When I was growing up and doing arithmetic, "times" was a multiplier that signified the adding of things together.  If I take 30 times 10, I don't get a number less than 10 as the (correct) answer.  Could the PCMOS actually be giving back electricity?  Put one watt into the chip and 30 watts are returned to the power source?  If so, we've solved the energy crisis for the next thousand years.  Hallelujah!

All right, I guess the writer of the story and the headline "Revolutionary microchip uses 30 times less power" probably meant that the chip used 1/30th of the power of current technology.  If so, why not say that?  Or, if fractions would confuse anyone that would be interested in such an announcement, say  "a microchip that uses 3.3% of the electricity of current chips".  To me, 3.3% is much more dramatic than 1/30th.

And it's much more cogent and believable than "30 times less." 

Either way, though, it sounds like cool stuff.

Categories: Uncategorized

For more complete information about compiler optimizations, see our Optimization Notice.

Comments (9)

February 9, 2009 12:47 PM PST rreis Total Points: 2,910 Brown Belt

Sorry but I think it's clear what he wanted to say. 30x less, so 1/30th... don't quite follow your rant.

February 9, 2009 2:15 PM PST Clay Breshears (Intel) Total Points: 30,566 Black Belt

My rant was because it wasn't clear. At least it wasn't to me. If I say to my enemy, "I hate you 30 times less today", has my spite increased thirty-fold, increased by 1/30th, reduced to 1/30th of what it was yesterday, or do I love him now 29X more than I used to hate him? (Or do I think about my burning hate only 70 times today when it was 100 times yesterday?) Take parallel speedup. If I say my parallel app has a speedup of 200%, how much time does it take relative to the serial execution? Does it take half the time (with 100% speedup being the same as serial time) or does it take 1/3rd the time (with 0% speedup being the same as serial time)? If it's the former, I should say 'I have a 2X speedup' (not two times less execution time). And don't get me started on whether 12pm is noon or midnight. (No matter which way you choose, I can find someone that disagrees with you.) My rant is for clarity in scientific exposition. Juxtaposing opposite terms just because it gives a more bombastic sound bite, yet makes no sense whatsoever, has no place in technical writing. We have words that mean certain things. Why not use them since it is just as easy to do so as it is to not do so?

February 9, 2009 3:10 PM PST jimdempseyatthecove Total Points: 77,429 Black Belt

While you are on the issue of ranting, my pet peeve is the use of bandwidth in computer-speak. Bandwidth is a radio or optical term to express a frequency range. Most computers run with a vary narrow bandwidth. Excepting for speed-step technology, (or dynamic over-clocking), the CPU bandwidth is very small. The performance of the system is frequency based, not bandwidth based. Work is performed with some proportionality to clock cycles and not the difference between the lowest cycle and highest cycle. So what is being said when someone claims to be using 30% of the system bandwidth? 30% of the processing capacity means something, 30% of a narrow bandwidth is meaningless. Jim Dempsey

February 9, 2009 4:10 PM PST rreis Total Points: 2,910 Brown Belt

sorry, can't agree. I see your point but I keep rewind in my head like a street conversation. "they have a thingy that just spends 30x times less energy than the other thingy" or ".. that spends 1/30th of"... I can see what is clearer (in terms of numbers) but also what best conveys the idea (in terms of shear saving). Maybe it's a question of context or just me that am not an english native speaker :)

February 9, 2009 10:09 PM PST Steve Paine

The other one that annoys me is when people say.... 'Its got a huge 5400mah battery' When mah has nothing to do with capacity unless you know the voltage!! Steve

February 10, 2009 5:16 AM PST fraggy Total Points: 1,020 Registered User

----------------- Take parallel speedup. If I say my parallel app has a speedup of 200%, how much time does it take relative to the serial execution? Does it take half the time (with 100% speedup being the same as serial time) or does it take 1/3rd the time (with 0% speedup being the same as serial time)? If it's the former, I should say 'I have a 2X speedup' (not two times less execution time). ----------------- This is the reason why I have stop using % in speedup : I'am afraid to lost readers !

February 13, 2009 9:33 AM PST Dweeberly

"I don't know half of you half as well as I should like; and I like less than half of you half as well as you deserve." - J. R. R. Tolkien

February 20, 2009 1:01 PM PST Clay Breshears (Intel) Total Points: 30,566 Black Belt

"In the fall of 1972 President Nixon announced that the rate of increase of inflation was decreasing. This was the first time a sitting president used the third derivative to advance his case for reelection." -- Hugo Rossi, "Mathematics Is an Edifice, Not a Toolbox," Notices of the AMS 43, no. 10, October 1996 Maybe "X times less" is attempting third derivative thinking, too?

March 23, 2009 9:01 AM PDT Clay Breshears (Intel) Total Points: 30,566 Black Belt

And another related thing (which really doesn't deserve a whole new blog), what's up with the use of superlatives like "Worst" and "Best"? I'm thinking about when these are applied to positive, neutral, and negative things. For example, for a positive or neutral job we know what it means to be the Best Teacher or Worst Cook. But what about a negative job, a position that carries negative connotations? What does it mean to be the Best Torturer or Worst Dictator (ala Parade Magazine)? In the fomer, does this mean he gets information the fastest and most efficient or with the most pain inflicted, or does this mean he applies the least painful methods (which is "best" for the prisoner being tortured)? Does the leader of a country described as the "Worst Dictator" do the a bad job of being a dictator and rule with honesty, humor, and benevolence? We need some new superlaitve to describe the Best of the Worst for negative things. "Berst"? So if you're the Berst Dictator, you are really hated and feared by your people and people around the world.

Trackbacks (0)

Leave a comment