Hi Stefan, In that case, wouldn't (shouldn't) it be possible to request a re-review of libraries alreayd accepted, provided there is good reason? In other words, as standards evolve, is there a mechanism by which libraries no longer up to par can be removed from boost/deprecated/marked as no longer officially endorsed, something like that (I can see that simply removing them would make people depending on them very unhappy). Best, Dee On Tue, Nov 3, 2009 at 2:16 AM, Stefan Strasser <strasser@uni-bremen.de> wrote:

Am Monday 02 November 2009 15:15:00 schrieb Hansjoerg:

...

Hello,

In my application I have a lot of small objects. For that I thought that it is the best to use an object pool..I have read the docu of object_pool...but I don't understand really how I have to use it..

I'd consider using another library, or allocate your objects in bulk yourself and, e.g. store the unused objects in a boost::intrusive::slist until they're needed. boost::object_pool has some major problems, e.g. linear complexity of free(). the library is a decade old and I don't think it would be accepted if it was reviewed as it is today. _______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

On Tuesday 03 November 2009 05:45:48 Stefan Strasser wrote:

Am Tuesday 03 November 2009 04:08:14 schrieb Diederick C. Niehorster: I actually considered that, but while recreating the functionality of Boost.Pool is pretty easy, given that Boost.Intrusive now exists, a generic pool allocation library is some effort. some things that come to mind is multiple allocation algorithms to meet different time complexity and memory overhead needs, pools that allow allocation of different-sized objects, maybe even a generic "allocation algorithm" library that is not limited to chunks of memory, but can be used for disk space and other arbitrary ranges, like allocating IDs.

I recently tried to use the std::allocator drop-in replacement from Boost.Pool in a massively multithreaded program running on a machine with a lot of cores. The result was very disappointing. I found my threads being serialized along the std::allocator interface (same problem as with the std::allocator from my STL). A nice feature for Boost.Pool would be the ability to scale in multithreaded environments (like for example the tbb::scaleable_allocator from the Intel Threading Building Blocks). KR, Frank

Zitat von Stefan Strasser <strasser@uni-bremen.de>:

Am Tuesday 03 November 2009 07:17:18 schrieb frank:

...

On Tuesday 03 November 2009 05:45:48 Stefan Strasser wrote:

...

Am Tuesday 03 November 2009 04:08:14 schrieb Diederick C. Niehorster: I actually considered that, but while recreating the functionality of Boost.Pool is pretty easy, given that Boost.Intrusive now exists, a generic pool allocation library is some effort. some things that come to mind is multiple allocation algorithms to meet different time complexity and memory overhead needs, pools that allow allocation of different-sized objects, maybe even a generic "allocation algorithm" library that is not limited to chunks of memory, but can be used for disk space and other arbitrary ranges, like allocating IDs.

I recently tried to use the std::allocator drop-in replacement from Boost.Pool in a massively multithreaded program running on a machine with a lot of cores. The result was very disappointing.

as is expected. you might want to look into using thread specific pools, that only synchronize when a new pool block is allocated.

could you please elaborate a bit on this? How would i do this with the std::allocator implemenation provided by Boost.Pool? KR, Frank

strasser@uni-bremen.de wrote:

Zitat von Frank Meerkoetter <frank@meerkoetter.org>:

...

...

as is expected. you might want to look into using thread specific pools, that only synchronize when a new pool block is allocated.

could you please elaborate a bit on this? How would i do this with the std::allocator implemenation provided by Boost.Pool?

the basic idea is to have an object pool for each thread, so that only when the pool runs out of memory std::new is called and a mutex lock is required. thread-local allocations can run concurrently.

you could use boost.pool for this, if none of your threads ever deallocate an object that was allocated by another thread (and thus be part of another pool). you`d have to write your own allocator I think. but boost.pool isn't suitable as a general purpose allocator anyway, because of O(n) free().

that`s another thing that Pool2 should support. some kind of hierarchial pool structure. when a pool runs out of memory it requests new memory from its parent pool, and also pools deallocations and puts those back into the right pool, so you don`t have the problem described above.

Zitat von Nigel Rantor <wiggly@wiggly.org>:

...

I would be really interested to hear more detailed critique of the current pool library from people who know/have used it.

there`s nothing "wrong" with boost.pool in the sense that it has errors.

That's why I carefully chose the word "critique" when asking for opinions. :-) I realise you're coming at it from a time -complexity and generality point of view.

but it only supports a very limited use case: pools of objects of the same type that are used to allocate objects individually but deallocated as a pool.

So would you really be suggesting augmenting the current set of facilities Pool provides then? I assume this since it seems to me (please correct me if I'm wrong) that the current Pool implementation (based on the docs) does a decent job of providing small-object allocation with minimal overhead in terms of memory complexity, even if it can be slow for de-allocating individual items rather than an entire pool.

here is (an outdated version of) the code I currently use, with O(1) allocation and O(1) deallocation. http://www.boostpro.com/vault/index.php?action=downloadfile&filename=intrusive_pool.hpp&directory=Memory&

(don`t use it, that version still had bugs)

but like boost.pool, it only supports a few (other) use cases, it is not a generic library.

here is an introduction to two allocation algorithms that pools could use to support arbitrary-sized objects: http://www.boost.org/doc/libs/1_38_0/doc/html/interprocess/allocators_contai...

Reading the above interprocess docs was another reason why I assume you're really talking about augmenting Pool. Those allocators appear like they'd have significant overhead for small objects, which Pool doesn't currently (With the same caveats about me reading the Pool docs correctly). So, if I understand correctly the wishlist is: - O(1) de-allocation. - Arbitrary sized allocations from a single pool. - Better performance for MT. I am tired so I'm going to be lazy and say that I can see how you could have multiple strategies in one library that allow people to pick their preference of time vs. space but I'm not immediately sure if you can make an allocation scheme that gives constant-time and space for all operations. Would need to actually spend some time thinking about it, seeing as it's not really my bag. Regards, Nigel

I want the following type of pool in the wishlist: Allowing to allocate misc-size memory blocks without an ability to delete them (I mean they will are deleted on pool destruction). I met some situations where such pool was really what needed.

I actually did this at one time w/ the boost::pool by making my requested_size 1 byte, then doing an ordered_malloc(num_bytes).

I'm not claiming that this is a good or a bad idea.  But it did work.

Heh. And I've written my own class-bicycle, allocating some fixed-size big chunks with boost::pool and then cutting needed blocks from them. But there are many stones and details, like memory aligning, or making object_pool from such pool. So, probably professionally-written solution will cover such problems.

Zitat von strasser@uni-bremen.de:

Zitat von Frank Meerkoetter <frank@meerkoetter.org>:

...

...

as is expected. you might want to look into using thread specific pools, that only synchronize when a new pool block is allocated.

could you please elaborate a bit on this? How would i do this with the std::allocator implemenation provided by Boost.Pool?

the basic idea is to have an object pool for each thread, so that only when the pool runs out of memory std::new is called and a mutex lock is required. thread-local allocations can run concurrently.

you could use boost.pool for this, if none of your threads ever deallocate an object that was allocated by another thread (and thus be part of another pool). you`d have to write your own allocator I think. but boost.pool isn't suitable as a general purpose allocator anyway, because of O(n) free().

Thanks for your explanation. The property that objects that are allocated in one thread must never be deallocated by a different thread is a show stopper for me. My current system is designed as a pipeline. KR, Frank

Diederick C. Niehorster wrote:

...

if we could put together a team for Pool2 I'd certainly take part in it, but it is too much effort for one person, considering that you can write something like that for your own needs, in a non-generic way, in a few hundred lines.

Hi Stefan,

I must say this is way beyond my reach (I'm not trained in programming or CS in general at all). Maybe there should be a place on the Boost website where these kind of projects or requests for team members could be put?

There must be more relatively old code that could use a refresh/thorough look.

I have not used pool yet but I'm currently working on a project that, based on profiling, might benefit from something like pool. I would be really interested to hear more detailed critique of the current pool library from people who know/have used it. At the moment I also have some amount of free time so might be interested in helping out if it sounds like something I could usefully work on. n