[External] : Re: JMH results for IndexedLinkedList

[Stuart Marks]

On 7/10/22 8:25 AM, Tagir Valeev wrote:
> Hello!
> 
> On Sun, Jul 10, 2022 at 6:33 AM Rodion Efremov <coderodd3 at gmail.com> wrote:
>> I am interested in
>> https://bugs.openjdk.org/browse/JDK-8143850
>>
>> Is there a way for becoming an assignee for the above issue? If yes, how do I proceed and what is the schedule?
> 
> Stuart Marks might be the right person to ask. Adding to CC.

Thanks Tagir.

There is already a considerable implementation overlap between ArrayList and 
ArrayDeque that seems difficult to unify. I'm loathe to add yet another data 
structure implementation that's functionally similar to the existing ones but has a 
bunch of subtle performance and usage tradeoffs. Doing this would add to the 
maintenance cost of the JDK and also make it more difficult for programmers to 
choose the "right" data structure for their application.

Already there is a problem with ArrayList and ArrayDeque being separate classes. 
ArrayDeque "fixes" the performance problem with addition/removal at the front of an 
ArrayList; yet last time I checked, ArrayDeque was used only about 1% as frequently 
as ArrayList. Maybe this is because people don't know about it, or maybe because 
people are uncomfortable losing List-like indexed access with ArrayDeque.

Regarding JDK-8143850 there is a twisty maze through the constraints of avoiding 
adding too much code, avoiding too much API surface area, and maintaining 
performance. I have a prototype of ArrayList that makes it double-ended, which seems 
promising, to me at least. (This is independent of SequencedCollection -- 
JDK-8266572 -- which is an API move, not an implementation move.) But doing this 
necessarily slows down append-only operations, while providing a performance boon 
for operations at the front for all but the smallest lists. The tradeoffs here are 
quite subtle.

In any case I think the best path forward for "new" data structures with interesting 
performance characteristics is to publish an artifact on Maven central. This will 
allow people who have specialized performance needs to use it.

s'marks

> 
> With best regards,
> Tagir Valeev.
> 
>>
>> Best regards,
>> rodde
>>
>> la 9.7.2022 klo 22.33 Tagir Valeev <amaembo at gmail.com> kirjoitti:
>>>
>>> Note that nobody these days cares about LinkedList. Use-cases where LinkedList outperforms careful use of ArrayList or ArrayDeque are next to none. So saying that your data structure is better than LinkedList is totally not a reason to add it to JDK. It should be better than ArrayList and ArrayDeque.
>>>
>>> Having a single data structure that provides list and deque interface is a reasonable idea. However it would be much simpler to retrofit existing data structure like ArrayDeque, rather than create a new data structure. Here's an issue for this:
>>> https://bugs.openjdk.org/browse/JDK-8143850
>>>
>>> There were also discussions to enhance collections in general, adding more useful methods like getFirst() or removeLast() to ArrayList, etc. See for details:
>>> https://bugs.openjdk.org/browse/JDK-8266572
>>>
>>> To conclude, the idea of adding one more collection implementation looks questionable to me. It will add more confusion when people need to select which collection fits their needs better. It will require more learning. This could be justified if there are clear benefits in using it in real world problems, compared to existing collections. But so far I don't see the examples of such problems.
>>>
>>> With best regards,
>>> Tagir Valeev
>>>
>>> сб, 9 июл. 2022 г., 11:22 Rodion Efremov <coderodd3 at gmail.com>:
>>>>
>>>> Hello,
>>>>
>>>> My benchmarking suggests, that, if nothing else, my IndexedLinkedList outperforms gracefully the java.util.LinkedList, so the use case should be the same (List<E> + Deque<E> -interfaces) for both of the aforementioned data structures.
>>>>
>>>> Best regards,
>>>> rodde
>>>>
>>>>
>>>> On Sat, Jul 9, 2022 at 11:19 AM Tagir Valeev <amaembo at gmail.com> wrote:
>>>>>
>>>>> Hello!
>>>>>
>>>>> Are there real world problems/use cases where IndexedLinkedList would be preferred in terms of CPU/memory usage over ArrayList?
>>>>>
>>>>> сб, 9 июл. 2022 г., 07:18 Rodion Efremov <coderodd3 at gmail.com>:
>>>>>>
>>>>>> Data structure repo:
>>>>>> https://urldefense.com/v3/__https://github.com/coderodde/IndexedLinkedList__;!!ACWV5N9M2RV99hQ!LBLiZkI0ZS8UoU_DXtnXCXGFsFzBUd3q7qyok0L3vUbTQXbnDnXoyb8GsZUyyt2pnCjra9VSKVLJG8h-$
>>>>>>
>>>>>> Benchmark repo:
>>>>>> https://urldefense.com/v3/__https://github.com/coderodde/IndexedLinkedListBenchmark__;!!ACWV5N9M2RV99hQ!LBLiZkI0ZS8UoU_DXtnXCXGFsFzBUd3q7qyok0L3vUbTQXbnDnXoyb8GsZUyyt2pnCjra9VSKafWnZNi$
>>>>>>
>>>>>> I have profiled my data structure and it seems it’s more performant than java.util.LinkedList or TreeList, if nothing else.
>>>>>>
>>>>>> So, is there any chance of including IndexedLinkedList to JDK?
>>>>>>
>>>>>> Best regards,
>>>>>> rodde