Swammerdami
Squadron Leader
There's only one thing anyone wants to hear from you, steve, And that is an explanation for the result you claimed:
Nothing did 10^10 iterations in 16 milliseconds. Steve is apparently unable to find the bug. Let's try to help him.
One possibility is that his loop looked like
But I don't think that's where steve went wrong. Presumably he had successfully coded a ten-billion iteration loop for his Mersenne Twister result. Having "ironed its bugs out", would he have "struck out for new frontier" for a similar loop?
There's another reason the second bug seems unlikely. The compiler will produce a warning when an option like "-Wall" is present. It seems safe to assume steve uses that option a lot.
Another possibility is that steve uses a smart optimizing compiler, and that that compiler could tell that the ten billion iterations would have no side effect ... so elided the whole loop altogether! I suppose it's possible that steve had never heard of such a thing, and is still wondering what happened.
But what is especially remarkable -- regardless of what steve's underlying bug was -- is that he has failed to acknowledge that there was a bug. What could he be thinking of? Is it possible that steve couldn't even grasp the absurdity of the timing he claimed? The result of the bug is plainly visible since Nothing did 10^10 iterations of anything in 16 milliseconds. Is he just hoping the bug will be forgotten? I've written over a million lines of code and am not ashamed to admit I couldn't count all the bugs on my digits even if I took my shoes off!
- - - - -- - - - - - -
I've thought of synopsizing some of my credentials in a Lounge thread. They're more a matter of shame for me than pride: With the genes I was dealt I should have been a math professor or start-up multimillionaire. But the genes were not enough, and the idiot part of my idiot-savant pathology took over. Still, there were two different V.P.s of Engineering who called me "the best microprogrammer in Silicon Valley." And two people called me the best circuit designer they'd ever met. (One of these had headed an IBM Research Lab, founded two companies, and ... was the best circuit designer I had ever met. I have over 30 U.S. patents, have written books, a newspaper column, and have published several peer-reviewed papers.
So I should just laugh at steve's pathetic efforts to insult me. But -- call me childish -- I do get irritated, so I set steve to ignore. If he ever explains the peculiar result quoted above, or manages to say something interesting or useful I hope someone will alert me.
Execution times C++ Mersenne Twister vs LG.
for 10^10 iterations C++ took 68 seconds, URAND LG took .016 seconds.
Nothing did 10^10 iterations in 16 milliseconds. Steve is apparently unable to find the bug. Let's try to help him.
One possibility is that his loop looked like
for (iter = 0; iter < (10^10); iter++)
orfor (iter = 0; iter < 10,000,000,000; iter++)
Although to a casual non-programmer these might appear to do 10 billion iterations, in fact they do zero. This despite that (10^10) and 10,000,000,000 appear to be synonyms of ten billion.But I don't think that's where steve went wrong. Presumably he had successfully coded a ten-billion iteration loop for his Mersenne Twister result. Having "ironed its bugs out", would he have "struck out for new frontier" for a similar loop?
There's another reason the second bug seems unlikely. The compiler will produce a warning when an option like "-Wall" is present. It seems safe to assume steve uses that option a lot.
Another possibility is that steve uses a smart optimizing compiler, and that that compiler could tell that the ten billion iterations would have no side effect ... so elided the whole loop altogether! I suppose it's possible that steve had never heard of such a thing, and is still wondering what happened.
But what is especially remarkable -- regardless of what steve's underlying bug was -- is that he has failed to acknowledge that there was a bug. What could he be thinking of? Is it possible that steve couldn't even grasp the absurdity of the timing he claimed? The result of the bug is plainly visible since Nothing did 10^10 iterations of anything in 16 milliseconds. Is he just hoping the bug will be forgotten? I've written over a million lines of code and am not ashamed to admit I couldn't count all the bugs on my digits even if I took my shoes off!
- - - - -- - - - - - -
I've thought of synopsizing some of my credentials in a Lounge thread. They're more a matter of shame for me than pride: With the genes I was dealt I should have been a math professor or start-up multimillionaire. But the genes were not enough, and the idiot part of my idiot-savant pathology took over. Still, there were two different V.P.s of Engineering who called me "the best microprogrammer in Silicon Valley." And two people called me the best circuit designer they'd ever met. (One of these had headed an IBM Research Lab, founded two companies, and ... was the best circuit designer I had ever met. I have over 30 U.S. patents, have written books, a newspaper column, and have published several peer-reviewed papers.
So I should just laugh at steve's pathetic efforts to insult me. But -- call me childish -- I do get irritated, so I set steve to ignore. If he ever explains the peculiar result quoted above, or manages to say something interesting or useful I hope someone will alert me.