how to refactor nested for loop into smaller for loop assume each of them independent?

[Chris Angelico]

On Mon, Oct 10, 2016 at 4:57 AM, BartC <bc at freeuk.com> wrote:
> On 09/10/2016 18:33, Jussi Piitulainen wrote:
>>
>> Chris Angelico writes:
>>
>>> On Mon, Oct 10, 2016 at 12:26 AM, Dennis Lee Bieber wrote:
>>>>
>>>> {This response is delayed as I'm waiting for the program to complete so
>>>> I
>>>> can get the run time}
>>>> {Well... it's been near 24 hours and still merrily scrolling sums on my
>>>> console -- so I'm going to kill the running program}
>>>
>>>
>>> Eight BILLION of them. I wouldn't bother waiting for that.
>>
>>
>> Eight billion is a big number. I know from experience that computers can
>> compute a big number of computations in a short time. Therefore, I would
>> expect the computer to compute eight billion computations in a short
>> time. Maybe.
>
>
> Compute, maybe. But the OP is printing something out at each step. Printing
> to a console is very, very slow, compared with even with running CPython.
>
> On my machine, CPython might run through an empty loop at 10 to 20M
> iterations per second, but is some 4000 times slower if it has to print one
> "A" per line at each step.

Yeah. Quickest way to slow down a calculation is to naively add a
progress indicator to it.

(Actually, that's not quite fair. The quickest way is to add a
time-of-day call every iteration, in a misguided attempt to reduce
console output to "one every 0.1 seconds". That's crazy slow.)

> The OP might try redirecting the output to a file (assuming he doesn't want
> to read every single line of output as it's produced). That will be much
> faster, although it depends on the workload in the run() function being
> executed.
>
> Or only print every 1000th line or something if progress has to be
> monitored.

Yeah, if it's just for progress status. Of course, that does assume
that the run() function doesn't have anything particularly costly in
it. If it does, well, dis gonna take a while....

ChrisA