Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1441401

Re: [RFC PATCH v2 6/7] lib/persubnode: Introducing a simple per-subnode APIs

Show all headers | View raw

Hello,

On Mon, Jul 11, 2016 at 01:32:11PM -0400, Waiman Long wrote:
> The percpu APIs are extensively used in the Linux kernel to reduce
> cacheline contention and improve performance. For some use cases, the
> percpu APIs may be too fine-grain for distributed resources whereas
> a per-node based allocation may be too coarse as we can have dozens
> of CPUs in a NUMA node in some high-end systems.
> 
> This patch introduces a simple per-subnode APIs where each of the
> distributed resources will be shared by only a handful of CPUs within
> a NUMA node. The per-subnode APIs are built on top of the percpu APIs
> and hence requires the same amount of memory as if the percpu APIs
> are used. However, it helps to reduce the total number of separate
> resources that needed to be managed. As a result, it can speed up code
> that need to iterate all the resources compared with using the percpu
> APIs. Cacheline contention, however, will increases slightly as each
> resource is shared by more than one CPU. As long as the number of CPUs
> in each subnode is small, the performance impact won't be significant.
> 
> In this patch, at most 2 sibling groups can be put into a subnode. For
> an x86-64 CPU, at most 4 CPUs will be in a subnode when HT is enabled
> and 2 when it is not.

I understand that there's a trade-off between local access and global
traversing and you're trying to find a sweet spot between the two, but
this seems pretty arbitrary.  What's the use case?  What are the
numbers?  Why are global traversals often enough to matter so much?

Thanks.

-- 
tejun

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

[PATCH v2 0/7] vfs: Use dlock list for SB's s_inodes list Waiman Long <Waiman.Long@hpe.com> - 2016-07-11 19:40 +0200
  [RFC PATCH v2 7/7] lib/dlock-list: Use the per-subnode APIs for managing lists Waiman Long <Waiman.Long@hpe.com> - 2016-07-11 19:40 +0200
  [PATCH v2 1/7] lib/dlock-list: Distributed and lock-protected lists Waiman Long <Waiman.Long@hpe.com> - 2016-07-11 19:40 +0200
    Re: [PATCH v2 1/7] lib/dlock-list: Distributed and lock-protected  lists Tejun Heo <tj@kernel.org> - 2016-07-13 18:20 +0200
      Re: [PATCH v2 1/7] lib/dlock-list: Distributed and lock-protected  lists Tejun Heo <tj@kernel.org> - 2016-07-14 14:00 +0200
        Re: [PATCH v2 1/7] lib/dlock-list: Distributed and lock-protected  lists Jan Kara <jack@suse.cz> - 2016-07-14 16:40 +0200
          Re: [PATCH v2 1/7] lib/dlock-list: Distributed and lock-protected  lists Tejun Heo <tj@kernel.org> - 2016-07-14 17:00 +0200
        Re: [PATCH v2 1/7] lib/dlock-list: Distributed and lock-protected  lists Tejun Heo <tj@kernel.org> - 2016-07-14 19:50 +0200
  [PATCH v2 2/7] lib/dlock-list: Add __percpu modifier for parameters Waiman Long <Waiman.Long@hpe.com> - 2016-07-11 19:40 +0200
    Re: [PATCH v2 2/7] lib/dlock-list: Add __percpu modifier for  parameters Tejun Heo <tj@kernel.org> - 2016-07-13 18:20 +0200
  [PATCH v2 5/7] vfs: Use dlock list for superblock's inode list Waiman Long <Waiman.Long@hpe.com> - 2016-07-11 19:40 +0200
  [PATCH v2 3/7] fsnotify: Simplify inode iteration on umount Waiman Long <Waiman.Long@hpe.com> - 2016-07-11 19:40 +0200
  [RFC PATCH v2 6/7] lib/persubnode: Introducing a simple per-subnode APIs Waiman Long <Waiman.Long@hpe.com> - 2016-07-11 19:40 +0200
    Re: [RFC PATCH v2 6/7] lib/persubnode: Introducing a simple  per-subnode APIs Boqun Feng <boqun.feng@gmail.com> - 2016-07-12 05:20 +0200
    Re: [RFC PATCH v2 6/7] lib/persubnode: Introducing a simple  per-subnode APIs Tejun Heo <tj@kernel.org> - 2016-07-12 16:30 +0200
      Re: [RFC PATCH v2 6/7] lib/persubnode: Introducing a simple  per-subnode APIs Tejun Heo <tj@kernel.org> - 2016-07-12 21:00 +0200

csiph-web