Yes, I did. The arrays are faster in 90% of cases. The only time I got the
HashSet to be faster was when I was caching the entire marker hierarchy and
the hierarchy was more than 3 levels deep. That is definitely not the most
common case.

Also, I think the Marker... localParents parameter might be more performant
as Marker[] localParents


On Mon, Apr 21, 2014 at 10:07 PM, Matt Sicker <[email protected]> wrote:

> Also, did you compare the performance of using an array versus a HashSet?
>
>
> On 21 April 2014 19:44, Matt Sicker <[email protected]> wrote:
>
>> Here's what I'm changing that contains method to:
>>
>>         private static boolean contains(final Marker parent, final
>> Marker... localParents) {
>>             //noinspection ForLoopReplaceableByForEach
>>             for (int i = 0, localParentsLength = localParents.length; i <
>> localParentsLength; i++) {
>>                 final Marker marker = localParents[i];
>>                 if (marker == parent) {
>>                     return true;
>>                 }
>>             }
>>             return false;
>>         }
>>
>>
>>
>> On 21 April 2014 19:42, Matt Sicker <[email protected]> wrote:
>>
>>> Which brings up another issue with markers. In MarkerManager, we have a
>>> volatile array of Markers. Here's the message from IntelliJ:
>>>
>>> Reports array fields which are declared as *volatile*. Such fields may
>>> be confusing, as accessing the array itself follows the rules for
>>> *volatile* fields, but accessing the array's contents does not. If such
>>> volatile access is needed to array contents, the JDK5.0
>>> *java.util.concurrent.atomic* classes should be used instead.
>>>
>>> Is this relevant here?
>>>
>>>
>>> On 21 April 2014 19:37, Matt Sicker <[email protected]> wrote:
>>>
>>>> 1) that would be my bad. I usually replace those with foreach loops
>>>> where possible. It's usually good to comment in those cases. I'll revert
>>>> that and comment.
>>>>
>>>> 2) that makes more sense than the exception
>>>>
>>>>
>>>> On 21 April 2014 18:46, Bruce Brouwer <[email protected]> wrote:
>>>>
>>>>> I saw that some small changes were being made to the Markers. I had a
>>>>> few thoughts regarding them:
>>>>>
>>>>> 1) Use of array iterator instead of indexed for loop.
>>>>> for (Marker marker : localParents)
>>>>> instead of
>>>>> for (int i = 0; i < localParents.length; i++)
>>>>>
>>>>> When I was doing my performance benchmarks, I was finding the latter
>>>>> to be faster. I'm guessing this is simply because a new Iterable object
>>>>> needs to be created to iterate over the array.
>>>>>
>>>>> For most methods, such as add, remove, this was not a big deal. But
>>>>> for the isInstanceOf and checkParent methods, we want those to be as fast
>>>>> as possible.
>>>>>
>>>>> 2) isInstanceOf(String markerName)
>>>>> Instead of throwing an IllegalArgumentException when a marker of name
>>>>> markerName doesn't exist, why don't we simply return false? I don't want 
>>>>> an
>>>>> IllegalArgumentException to happen because I'm testing a markerName.
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>>
>>>>> Bruce Brouwer
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> Matt Sicker <[email protected]>
>>>>
>>>
>>>
>>>
>>> --
>>> Matt Sicker <[email protected]>
>>>
>>
>>
>>
>> --
>> Matt Sicker <[email protected]>
>>
>
>
>
> --
> Matt Sicker <[email protected]>
>



-- 

Bruce Brouwer

Reply via email to