[squeak-dev] Re: The defaullt implementation of isEmpty might do
too much work
H. Hirzel
hannes.hirzel at gmail.com
Thu Oct 27 17:20:15 UTC 2016
Yes, in this case a series of tests across collection classes actually
would be a proof. A single point measurement is not a statistically
relevant result.
--Hannes
On 10/27/16, Chris Muller <ma.chris.m at gmail.com> wrote:
>> | b |
>> b := (1 to: 10000) asBag.
>> [b isEmpty] bench
>>
>> before: : '5,540 per second. 181 microseconds per run.'
>> now: '167,000 per second. 6 microseconds per run.'
>>
>> In my book that's a speedup worth having.
>
> You're depending on the superclass implementation in Collection for
> that speedup. What did you say about that yesterday?
>
> Honestly, Bert, it almost feels like you're being intellectually
> dishonest. You chose the only one that got sped up and ignored all
> the ones that got slowed down. AND, you also ignored all the other
> points (again) which render the above point irrelevant. Sad.
>
> I give up. You "win".
>
>
More information about the Squeak-dev
mailing list
|