DPDK patches and discussions
 help / color / mirror / Atom feed
From: Olivier Matz <olivier.matz@6wind.com>
To: "Morten Brørup" <mb@smartsharesystems.com>
Cc: andrew.rybchenko@oktetlabs.ru, jerinj@marvell.com,
	thomas@monjalon.net, bruce.richardson@intel.com, dev@dpdk.org
Subject: Re: [PATCH] mempool: cache align mempool cache objects
Date: Thu, 27 Oct 2022 13:42:40 +0200	[thread overview]
Message-ID: <Y1pusD3PMQF2Ijdr@platinum> (raw)
In-Reply-To: <98CBD80474FA8B44BF855DF32C47DC35D87450@smartserver.smartshare.dk>

On Thu, Oct 27, 2022 at 11:22:07AM +0200, Morten Brørup wrote:
> > From: Olivier Matz [mailto:olivier.matz@6wind.com]
> > Sent: Thursday, 27 October 2022 10.35
> > 
> > Hi Morten,
> > 
> > On Wed, Oct 26, 2022 at 04:44:36PM +0200, Morten Brørup wrote:
> > > Add __rte_cache_aligned to the objs array.
> > >
> > > It makes no difference in the general case, but if get/put operations
> > are
> > > always 32 objects, it will reduce the number of memory (or last level
> > > cache) accesses from five to four 64 B cache lines for every get/put
> > > operation.
> > >
> > > For readability reasons, an example using 16 objects follows:
> > >
> > > Currently, with 16 objects (128B), we access to 3
> > > cache lines:
> > >
> > >       ┌────────┐
> > >       │len     │
> > > cache │********│---
> > > line0 │********│ ^
> > >       │********│ |
> > >       ├────────┤ | 16 objects
> > >       │********│ | 128B
> > > cache │********│ |
> > > line1 │********│ |
> > >       │********│ |
> > >       ├────────┤ |
> > >       │********│_v_
> > > cache │        │
> > > line2 │        │
> > >       │        │
> > >       └────────┘
> > >
> > > With the alignment, it is also 3 cache lines:
> > >
> > >       ┌────────┐
> > >       │len     │
> > > cache │        │
> > > line0 │        │
> > >       │        │
> > >       ├────────┤---
> > >       │********│ ^
> > > cache │********│ |
> > > line1 │********│ |
> > >       │********│ |
> > >       ├────────┤ | 16 objects
> > >       │********│ | 128B
> > > cache │********│ |
> > > line2 │********│ |
> > >       │********│ v
> > >       └────────┘---
> > >
> > > However, accessing the objects at the bottom of the mempool cache is
> > a
> > > special case, where cache line0 is also used for objects.
> > >
> > > Consider the next burst (and any following bursts):
> > >
> > > Current:
> > >       ┌────────┐
> > >       │len     │
> > > cache │        │
> > > line0 │        │
> > >       │        │
> > >       ├────────┤
> > >       │        │
> > > cache │        │
> > > line1 │        │
> > >       │        │
> > >       ├────────┤
> > >       │        │
> > > cache │********│---
> > > line2 │********│ ^
> > >       │********│ |
> > >       ├────────┤ | 16 objects
> > >       │********│ | 128B
> > > cache │********│ |
> > > line3 │********│ |
> > >       │********│ |
> > >       ├────────┤ |
> > >       │********│_v_
> > > cache │        │
> > > line4 │        │
> > >       │        │
> > >       └────────┘
> > > 4 cache lines touched, incl. line0 for len.
> > >
> > > With the proposed alignment:
> > >       ┌────────┐
> > >       │len     │
> > > cache │        │
> > > line0 │        │
> > >       │        │
> > >       ├────────┤
> > >       │        │
> > > cache │        │
> > > line1 │        │
> > >       │        │
> > >       ├────────┤
> > >       │        │
> > > cache │        │
> > > line2 │        │
> > >       │        │
> > >       ├────────┤
> > >       │********│---
> > > cache │********│ ^
> > > line3 │********│ |
> > >       │********│ | 16 objects
> > >       ├────────┤ | 128B
> > >       │********│ |
> > > cache │********│ |
> > > line4 │********│ |
> > >       │********│_v_
> > >       └────────┘
> > > Only 3 cache lines touched, incl. line0 for len.
> > 
> > I understand your logic, but are we sure that having an application
> > that
> > works with bulks of 32 means that the cache will stay aligned to 32
> > elements for the whole life of the application?
> > 
> > In an application, the alignment of the cache can change if you have
> > any of:
> > - software queues (reassembly for instance)
> > - packet duplication (bridge, multicast)
> > - locally generated packets (keepalive, control protocol)
> > - pipeline to other cores
> > 
> > Even with testpmd, which work by bulk of 32, I can see that the size
> > of the cache filling is not aligned to 32. Right after starting the
> > application, we already have this:
> > 
> >   internal cache infos:
> >     cache_size=250
> >     cache_count[0]=231
> > 
> > This is probably related to the hw rx rings size, number of queues,
> > number of ports.
> > 
> > The "250" default value for cache size in testpmd is questionable, but
> > with --mbcache=256, the behavior is similar.
> > 
> > Also, when we transmit to a NIC, the mbufs are not returned immediatly
> > to the pool, they may stay in the hw tx ring during some time, which is
> > a driver decision.
> > 
> > After processing traffic on cores 8 and 24 with this testpmd, I get:
> >     cache_count[0]=231
> >     cache_count[8]=123
> >     cache_count[24]=122
> > 
> > In my opinion, it is not realistic to think that the mempool cache will
> > remain aligned to cachelines. In these conditions, it looks better to
> > keep the structure packed to avoid wasting memory.
> 
> I agree that is a special use case to only access the mempool cache in
> bursts of 32 objects, so the accesses are always cache line
> aligned. (Generalized, the burst size must not be 32; a burst size
> that is a multiple of RTE_CACHE_LINE_SIZE/sizeof(void*), i.e. a burst
> size of 8 on a 64-bit architecture, will do.)

Is there a real situation where it happens to always have read/write
accesses per bulks of 32? From what I see in my quick test, it is not
the case, even with testpmd.

> Adding a hole of 52 byte per mempool cache is nothing, considering
> that the mempool cache already uses 8 KB (RTE_MEMPOOL_CACHE_MAX_SIZE *
> 2 * sizeof(void*) = 1024 * 8 byte) for the objects.
>
> Also - assuming that memory allocations are cache line aligned - the
> 52 byte of unused memory cannot be used regardless if they are before
> or after the objects. Instead of having 52 B unused after the objects,
> we might as well have a hole of 52 B unused before the objects. In
> other words: There is really no downside to this.

Correct, the memory waste argument to nack the patch is invalid.

> Jerin also presented a separate argument for moving the objects to
> another cache line than the len field: The risk for "load-after-store
> stall" when loading the len field after storing objects in cache line0
> [1].
> 
> [1]: http://inbox.dpdk.org/dev/CALBAE1P4zFYdLwoQukn5Q-V-nTvc_UBWmWjhaV2uVBXQRytSSA@mail.gmail.com/

I'll be prudent on this justification without numbers. The case where we
access to the objects of the first cache line (among several KB) is
maybe not that frequent.

> A new idea just popped into my head: The hot debug statistics
> counters (put_bulk, put_objs, get_success_bulk, get_success_objs)
> could be moved to this free space, reducing the need to touch another
> cache line for debug counters. I haven’t thought this idea through
> yet; it might conflict with Jerin's comment.

Yes, but since the stats are only enabled when RTE_LIBRTE_MEMPOOL_DEBUG
is set, it won't have any impact on non-debug builds.


Honnestly, I find it hard to convince myself that it is a real
optimization. I don't see any reason why it would be slower though. So
since we already broke the mempool cache struct ABI in a previous
commit, and since it won't consume more memory, I'm ok to include that
patch. It would be great to have numbers to put some weight in the
balance.




> 
> > 
> > Olivier
> > 
> > 
> > >
> > > Credits go to Olivier Matz for the nice ASCII graphics.
> > >
> > > Signed-off-by: Morten Brørup <mb@smartsharesystems.com>
> > > ---
> > >  lib/mempool/rte_mempool.h | 6 ++++--
> > >  1 file changed, 4 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/lib/mempool/rte_mempool.h b/lib/mempool/rte_mempool.h
> > > index 1f5707f46a..3725a72951 100644
> > > --- a/lib/mempool/rte_mempool.h
> > > +++ b/lib/mempool/rte_mempool.h
> > > @@ -86,11 +86,13 @@ struct rte_mempool_cache {
> > >  	uint32_t size;	      /**< Size of the cache */
> > >  	uint32_t flushthresh; /**< Threshold before we flush excess
> > elements */
> > >  	uint32_t len;	      /**< Current cache count */
> > > -	/*
> > > +	/**
> > > +	 * Cache objects
> > > +	 *
> > >  	 * Cache is allocated to this size to allow it to overflow in
> > certain
> > >  	 * cases to avoid needless emptying of cache.
> > >  	 */
> > > -	void *objs[RTE_MEMPOOL_CACHE_MAX_SIZE * 2]; /**< Cache objects */
> > > +	void *objs[RTE_MEMPOOL_CACHE_MAX_SIZE * 2] __rte_cache_aligned;
> > >  } __rte_cache_aligned;
> > >
> > >  /**
> > > --
> > > 2.17.1
> > >
> 

  reply	other threads:[~2022-10-27 11:42 UTC|newest]

Thread overview: 85+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-12-26 15:34 [RFC] mempool: rte_mempool_do_generic_get optimizations Morten Brørup
2022-01-06 12:23 ` [PATCH] mempool: optimize incomplete cache handling Morten Brørup
2022-01-06 16:55   ` Jerin Jacob
2022-01-07  8:46     ` Morten Brørup
2022-01-10  7:26       ` Jerin Jacob
2022-01-10 10:55         ` Morten Brørup
2022-01-14 16:36 ` [PATCH] mempool: fix get objects from mempool with cache Morten Brørup
2022-01-17 17:35   ` Bruce Richardson
2022-01-18  8:25     ` Morten Brørup
2022-01-18  9:07       ` Bruce Richardson
2022-01-24 15:38   ` Olivier Matz
2022-01-24 16:11     ` Olivier Matz
2022-01-28 10:22     ` Morten Brørup
2022-01-17 11:52 ` [PATCH] mempool: optimize put objects to " Morten Brørup
2022-01-19 14:52 ` [PATCH v2] mempool: fix " Morten Brørup
2022-01-19 15:03 ` [PATCH v3] " Morten Brørup
2022-01-24 15:39   ` Olivier Matz
2022-01-28  9:37     ` Morten Brørup
2022-02-02  8:14 ` [PATCH v2] mempool: fix get objects from " Morten Brørup
2022-06-15 21:18   ` Morten Brørup
2022-09-29 10:52     ` Morten Brørup
2022-10-04 12:57   ` Andrew Rybchenko
2022-10-04 15:13     ` Morten Brørup
2022-10-04 15:58       ` Andrew Rybchenko
2022-10-04 18:09         ` Morten Brørup
2022-10-06 13:43       ` Aaron Conole
2022-10-04 16:03   ` Morten Brørup
2022-10-04 16:36   ` Morten Brørup
2022-10-04 16:39   ` Morten Brørup
2022-02-02 10:33 ` [PATCH v4] mempool: fix mempool cache flushing algorithm Morten Brørup
2022-04-07  9:04   ` Morten Brørup
2022-04-07  9:14     ` Bruce Richardson
2022-04-07  9:26       ` Morten Brørup
2022-04-07 10:32         ` Bruce Richardson
2022-04-07 10:43           ` Bruce Richardson
2022-04-07 11:36             ` Morten Brørup
2022-10-04 20:01   ` Morten Brørup
2022-10-09 11:11   ` [PATCH 1/2] mempool: check driver enqueue result in one place Andrew Rybchenko
2022-10-09 11:11     ` [PATCH 2/2] mempool: avoid usage of term ring on put Andrew Rybchenko
2022-10-09 13:08       ` Morten Brørup
2022-10-09 13:14         ` Andrew Rybchenko
2022-10-09 13:01     ` [PATCH 1/2] mempool: check driver enqueue result in one place Morten Brørup
2022-10-09 13:19   ` [PATCH v4] mempool: fix mempool cache flushing algorithm Andrew Rybchenko
2022-10-04 12:53 ` [PATCH v3] mempool: fix get objects from mempool with cache Andrew Rybchenko
2022-10-04 14:42   ` Morten Brørup
2022-10-07 10:44 ` [PATCH v4] " Andrew Rybchenko
2022-10-08 20:56   ` Thomas Monjalon
2022-10-11 20:30     ` Copy-pasted code should be updated Morten Brørup
2022-10-11 21:47       ` Honnappa Nagarahalli
2022-10-30  8:44         ` Morten Brørup
2022-10-30 22:50           ` Honnappa Nagarahalli
2022-10-14 14:01     ` [PATCH v4] mempool: fix get objects from mempool with cache Olivier Matz
2022-10-09 13:37 ` [PATCH v6 0/4] mempool: fix mempool cache flushing algorithm Andrew Rybchenko
2022-10-09 13:37   ` [PATCH v6 1/4] mempool: check driver enqueue result in one place Andrew Rybchenko
2022-10-09 13:37   ` [PATCH v6 2/4] mempool: avoid usage of term ring on put Andrew Rybchenko
2022-10-09 13:37   ` [PATCH v6 3/4] mempool: fix cache flushing algorithm Andrew Rybchenko
2022-10-09 14:31     ` Morten Brørup
2022-10-09 14:51       ` Andrew Rybchenko
2022-10-09 15:08         ` Morten Brørup
2022-10-14 14:01           ` Olivier Matz
2022-10-14 15:57             ` Morten Brørup
2022-10-14 19:50               ` Olivier Matz
2022-10-15  6:57                 ` Morten Brørup
2022-10-18 16:32                   ` Jerin Jacob
2022-10-09 13:37   ` [PATCH v6 4/4] mempool: flush cache completely on overflow Andrew Rybchenko
2022-10-09 14:44     ` Morten Brørup
2022-10-14 14:01       ` Olivier Matz
2022-10-10 15:21   ` [PATCH v6 0/4] mempool: fix mempool cache flushing algorithm Thomas Monjalon
2022-10-11 19:26     ` Morten Brørup
2022-10-26 14:09     ` Thomas Monjalon
2022-10-26 14:26       ` Morten Brørup
2022-10-26 14:44         ` [PATCH] mempool: cache align mempool cache objects Morten Brørup
2022-10-26 19:44           ` Andrew Rybchenko
2022-10-27  8:34           ` Olivier Matz
2022-10-27  9:22             ` Morten Brørup
2022-10-27 11:42               ` Olivier Matz [this message]
2022-10-27 12:11                 ` Morten Brørup
2022-10-27 15:20                   ` Olivier Matz
2022-10-28  6:35           ` [PATCH v3 1/2] " Morten Brørup
2022-10-28  6:35             ` [PATCH v3 2/2] mempool: optimized debug statistics Morten Brørup
2022-10-28  6:41           ` [PATCH v4 1/2] mempool: cache align mempool cache objects Morten Brørup
2022-10-28  6:41             ` [PATCH v4 2/2] mempool: optimized debug statistics Morten Brørup
2022-10-30  9:09               ` Morten Brørup
2022-10-30  9:16                 ` Thomas Monjalon
2022-10-30  9:17             ` [PATCH v4 1/2] mempool: cache align mempool cache objects Thomas Monjalon

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=Y1pusD3PMQF2Ijdr@platinum \
    --to=olivier.matz@6wind.com \
    --cc=andrew.rybchenko@oktetlabs.ru \
    --cc=bruce.richardson@intel.com \
    --cc=dev@dpdk.org \
    --cc=jerinj@marvell.com \
    --cc=mb@smartsharesystems.com \
    --cc=thomas@monjalon.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).