Re: [UnifiedV4 00/16] The Unified slab allocator (V4)

[Wu Fengguang <fengguang.wu@xxxxxxxxx>]

[add CC to Alex: he is now in charge of kernel performance tests]

On Wed, Oct 06, 2010 at 11:01:35AM +0300, Pekka Enberg wrote:
> (Adding more people who've taken interest in slab performance in the
> past to CC.)
> 
> On Tue, Oct 5, 2010 at 9:57 PM, Christoph Lameter <cl@xxxxxxxxx> wrote:
> > V3->V4:
> > - Lots of debugging
> > - Performance optimizations (more would be good)...
> > - Drop per slab locking in favor of per node locking for
> > Âpartial lists (queuing implies freeing large amounts of objects
> > Âto per node lists of slab).
> > - Implement object expiration via reclaim VM logic.
> >
> > The following is a release of an allocator based on SLAB
> > and SLUB that integrates the best approaches from both allocators. The
> > per cpu queuing is like in SLAB whereas much of the infrastructure
> > comes from SLUB.
> >
> > After this patches SLUB will track the cpu cache contents
> > like SLAB attemped to. There are a number of architectural differences:
> >
> > 1. SLUB accurately tracks cpu caches instead of assuming that there
> > Â is only a single cpu cache per node or system.
> >
> > 2. SLUB object expiration is tied into the page reclaim logic. There
> > Â is no periodic cache expiration.
> >
> > 3. SLUB caches are dynamically configurable via the sysfs filesystem.
> >
> > 4. There is no per slab page metadata structure to maintain (aside
> > Â from the object bitmap that usually fits into the page struct).
> >
> > 5. Has all the resiliency and diagnostic features of SLUB.
> >
> > The unified allocator is a merging of SLUB with some queuing concepts from
> > SLAB and a new way of managing objects in the slabs using bitmaps. Memory
> > wise this is slightly more inefficient than SLUB (due to the need to place
> > large bitmaps --sized a few words--in some slab pages if there are more
> > than BITS_PER_LONG objects in a slab) but in general does not increase space
> > use too much.
> >
> > The SLAB scheme of not touching the object during management is adopted.
> > The unified allocator can efficiently free and allocate cache cold objects
> > without causing cache misses.
> >
> > Some numbers using tcp_rr on localhost
> >
> >
> > Dell R910 128G RAM, 64 processors, 4 NUMA nodes
> >
> > threads unified     slub      Âslab
> > 64 Â Â Â4141798 Â Â Â Â 3729037 Â Â Â Â 3884939
> > 128 Â Â 4146587 Â Â Â Â 3890993 Â Â Â Â 4105276
> > 192 Â Â 4003063 Â Â Â Â 3876570 Â Â Â Â 4110971
> > 256 Â Â 3928857 Â Â Â Â 3942806 Â Â Â Â 4099249
> > 320 Â Â 3922623 Â Â Â Â 3969042 Â Â Â Â 4093283
> > 384 Â Â 3827603 Â Â Â Â 4002833 Â Â Â Â 4108420
> > 448 Â Â 4140345 Â Â Â Â 4027251 Â Â Â Â 4118534
> > 512 Â Â 4163741 Â Â Â Â 4050130 Â Â Â Â 4122644
> > 576 Â Â 4175666 Â Â Â Â 4099934 Â Â Â Â 4149355
> > 640 Â Â 4190332 Â Â Â Â 4142570 Â Â Â Â 4175618
> > 704 Â Â 4198779 Â Â Â Â 4173177 Â Â Â Â 4193657
> > 768 Â Â 4662216 Â Â Â Â 4200462 Â Â Â Â 4222686
> 
> Are there any stability problems left? Have you tried other benchmarks
> (e.g. hackbench, sysbench)? Can we merge the series in smaller
> batches? For example, if we leave out the NUMA parts in the first
> stage, do we expect to see performance regressions?
> 
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo@xxxxxxxxxx  For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxxx  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>