"Richard Heathfield" <rjh@see.sig.invalid> wrote in message
news:Z76dnZQYG-54sb3VnZ2dnUVZ8q2dnZ2d@bt.com...
> Bartc said:
>
>>
>> "Richard Heathfield" <rjh@see.sig.invalid> wrote in message
>> news:YKmdnTjH8aINu73VnZ2dnUVZ8uKdnZ2d@bt.com...
>>> Bartc said:
>>>
> <snip>
>
>>>> It's possible C++ and Java have some start-up delays in which case you
>>>> cannot compare /runtimes/ reliably.
>>>
>>> Rubbish. It takes how long it takes. If WonderfulNewLanguage*= can
>>> process the data in a time comparable to, say, Java, *provided* you
>>> discount WonderfulNewLanguage*='s 20-minute startup delay, does that
>>> mean WonderfulNewLanguage*= is as good at Java over that "distance"?
>>> Clearly not. When you're racing for money, you don't give head starts.
>>
>> But the Java startup isn't 1200000ms; whatever it is, it's no more than
>> 58ms in this case, which is insignificant; probably less than the time it
>> took to press the Enter key or click the button to start the test.
>
> If it's insignificant, why bother to raise it in the first place?

Because you were testing a benchmark with such a small runtime that it
became significant! A 50ms startup wouldn't be an issue with anybody, in
practical terms. But in this case, it could have been obscuring actual
runtime differences. Perhaps Java was twice the speed of C, after startup,
but your test wouldn't show it!

Your figures also showed C++ to be 7x as slow as C, but in your other post
(where the timings were more substantial) it was only 8% slower.

--
Bartc