Re: [PATCH RFC v2 00/10] SLUB percpu sheaves

2025-03-17 Thread Suren Baghdasaryan
On Mon, Mar 17, 2025 at 4:08 AM Vlastimil Babka wrote: > > On 3/14/25 18:10, Suren Baghdasaryan wrote: > > On Tue, Mar 4, 2025 at 11:08 AM Liam R. Howlett > > wrote: > >> > >> * Vlastimil Babka [250304 05:55]: > >> > On 2/25/25 21:26, Suren Baghdasaryan wrote: > >> > > On Mon, Feb 24, 2025 at 1

Re: [PATCH RFC v2 00/10] SLUB percpu sheaves

2025-03-17 Thread Vlastimil Babka
On 3/14/25 18:10, Suren Baghdasaryan wrote: > On Tue, Mar 4, 2025 at 11:08 AM Liam R. Howlett > wrote: >> >> * Vlastimil Babka [250304 05:55]: >> > On 2/25/25 21:26, Suren Baghdasaryan wrote: >> > > On Mon, Feb 24, 2025 at 1:12 PM Suren Baghdasaryan >> > > wrote: >> > >> >> > >> > >> > >> > >

Re: [PATCH RFC v2 00/10] SLUB percpu sheaves

2025-03-15 Thread Suren Baghdasaryan
On Tue, Mar 4, 2025 at 11:08 AM Liam R. Howlett wrote: > > * Vlastimil Babka [250304 05:55]: > > On 2/25/25 21:26, Suren Baghdasaryan wrote: > > > On Mon, Feb 24, 2025 at 1:12 PM Suren Baghdasaryan > > > wrote: > > >> > > >> > > > >> > > The values represent the total time it took to perform mm

Re: [PATCH RFC v2 00/10] SLUB percpu sheaves

2025-03-04 Thread Liam R. Howlett
* Vlastimil Babka [250304 05:55]: > On 2/25/25 21:26, Suren Baghdasaryan wrote: > > On Mon, Feb 24, 2025 at 1:12 PM Suren Baghdasaryan > > wrote: > >> > >> > > >> > > The values represent the total time it took to perform mmap syscalls, > >> > > less is > >> > > better. > >> > > > >> > > (1)

Re: [PATCH RFC v2 00/10] SLUB percpu sheaves

2025-03-04 Thread Suren Baghdasaryan
On Tue, Mar 4, 2025 at 2:55 AM Vlastimil Babka wrote: > > On 2/25/25 21:26, Suren Baghdasaryan wrote: > > On Mon, Feb 24, 2025 at 1:12 PM Suren Baghdasaryan > > wrote: > >> > >> > > >> > > The values represent the total time it took to perform mmap syscalls, > >> > > less is > >> > > better. >

Re: [PATCH RFC v2 00/10] SLUB percpu sheaves

2025-03-04 Thread Vlastimil Babka
On 2/25/25 21:26, Suren Baghdasaryan wrote: > On Mon, Feb 24, 2025 at 1:12 PM Suren Baghdasaryan wrote: >> >> > >> > > The values represent the total time it took to perform mmap syscalls, >> > > less is >> > > better. >> > > >> > > (1) baseline control >> > > Little core

Re: [PATCH RFC v2 00/10] SLUB percpu sheaves

2025-02-25 Thread Suren Baghdasaryan
On Mon, Feb 24, 2025 at 1:12 PM Suren Baghdasaryan wrote: > > On Mon, Feb 24, 2025 at 12:53 PM Vlastimil Babka wrote: > > > > On 2/24/25 02:36, Suren Baghdasaryan wrote: > > > On Sat, Feb 22, 2025 at 8:44 PM Suren Baghdasaryan > > > wrote: > > >> > > >> Don't know about this particular part but

Re: [PATCH RFC v2 00/10] SLUB percpu sheaves

2025-02-24 Thread Suren Baghdasaryan
On Mon, Feb 24, 2025 at 12:53 PM Vlastimil Babka wrote: > > On 2/24/25 02:36, Suren Baghdasaryan wrote: > > On Sat, Feb 22, 2025 at 8:44 PM Suren Baghdasaryan > > wrote: > >> > >> Don't know about this particular part but testing sheaves with maple > >> node cache and stress testing mmap/munmap

Re: [PATCH RFC v2 00/10] SLUB percpu sheaves

2025-02-24 Thread Vlastimil Babka
On 2/24/25 02:36, Suren Baghdasaryan wrote: > On Sat, Feb 22, 2025 at 8:44 PM Suren Baghdasaryan wrote: >> >> Don't know about this particular part but testing sheaves with maple >> node cache and stress testing mmap/munmap syscalls shows performance >> benefits as long as there is some delay to l

Re: [PATCH RFC v2 00/10] SLUB percpu sheaves

2025-02-23 Thread Suren Baghdasaryan
On Sun, Feb 23, 2025 at 5:36 PM Suren Baghdasaryan wrote: > > On Sat, Feb 22, 2025 at 8:44 PM Suren Baghdasaryan wrote: > > > > On Sat, Feb 22, 2025 at 4:19 PM Kent Overstreet > > wrote: > > > > > > On Fri, Feb 14, 2025 at 05:27:36PM +0100, Vlastimil Babka wrote: > > > > - Cheaper fast paths. Fo

Re: [PATCH RFC v2 00/10] SLUB percpu sheaves

2025-02-23 Thread Suren Baghdasaryan
On Sat, Feb 22, 2025 at 8:44 PM Suren Baghdasaryan wrote: > > On Sat, Feb 22, 2025 at 4:19 PM Kent Overstreet > wrote: > > > > On Fri, Feb 14, 2025 at 05:27:36PM +0100, Vlastimil Babka wrote: > > > - Cheaper fast paths. For allocations, instead of local double cmpxchg, > > > after Patch 5 it's

Re: [PATCH RFC v2 00/10] SLUB percpu sheaves

2025-02-22 Thread Suren Baghdasaryan
On Sat, Feb 22, 2025 at 4:19 PM Kent Overstreet wrote: > > On Fri, Feb 14, 2025 at 05:27:36PM +0100, Vlastimil Babka wrote: > > - Cheaper fast paths. For allocations, instead of local double cmpxchg, > > after Patch 5 it's preempt_disable() and no atomic operations. Same for > > freeing, which

Re: [PATCH RFC v2 00/10] SLUB percpu sheaves

2025-02-22 Thread Kent Overstreet
On Fri, Feb 14, 2025 at 05:27:36PM +0100, Vlastimil Babka wrote: > - Cheaper fast paths. For allocations, instead of local double cmpxchg, > after Patch 5 it's preempt_disable() and no atomic operations. Same for > freeing, which is normally a local double cmpxchg only for a short > term allo

Re: [PATCH RFC v2 00/10] SLUB percpu sheaves

2025-02-14 Thread Christoph Lameter (Ampere)
On Fri, 14 Feb 2025, Vlastimil Babka wrote: > - Cheaper fast paths. For allocations, instead of local double cmpxchg, > after Patch 5 it's preempt_disable() and no atomic operations. Same for > freeing, which is normally a local double cmpxchg only for a short > term allocations (so the same