DPDK patches and discussions
 help / color / mirror / Atom feed
From: "Mattias Rönnblom" <hofors@lysator.liu.se>
To: "Morten Brørup" <mb@smartsharesystems.com>,
	olivier.matz@6wind.com, andrew.rybchenko@oktetlabs.ru,
	mattias.ronnblom@ericsson.com, stephen@networkplumber.org,
	jerinj@marvell.com, bruce.richardson@intel.com
Cc: thomas@monjalon.net, dev@dpdk.org
Subject: Re: [PATCH v4 3/3] mempool: use cache for frequently updated stats
Date: Mon, 7 Nov 2022 08:30:40 +0100	[thread overview]
Message-ID: <532b74d2-37f3-7f0d-f7bc-f8f7034b6559@lysator.liu.se> (raw)
In-Reply-To: <20221104120329.1219-3-mb@smartsharesystems.com>

On 2022-11-04 13:03, Morten Brørup wrote:
> When built with stats enabled (RTE_LIBRTE_MEMPOOL_STATS defined), the
> performance of mempools with caches is improved as follows.
> 
> When accessing objects in the mempool, either the put_bulk and put_objs or
> the get_success_bulk and get_success_objs statistics counters are likely
> to be incremented.
> 
> By adding an alternative set of these counters to the mempool cache
> structure, accessing the dedicated statistics structure is avoided in the
> likely cases where these counters are incremented.
> 
> The trick here is that the cache line holding the mempool cache structure
> is accessed anyway, in order to access the 'len' or 'flushthresh' fields.
> Updating some statistics counters in the same cache line has lower
> performance cost than accessing the statistics counters in the dedicated
> statistics structure, which resides in another cache line.
> 
> mempool_perf_autotest with this patch shows the following improvements in
> rate_persec.
> 
> The cost of enabling mempool stats (without debug) after this patch:
> -6.8 % and -6.7 %, respectively without and with cache.
> 
> v4:
> * Fix checkpatch warnings:
>    A couple of typos in the patch description.
>    The macro to add to a mempool cache stat variable should not use
>    do {} while (0). Personally, I would tend to disagree with this, but
>    whatever keeps the CI happy.
> v3:
> * Don't update the description of the RTE_MEMPOOL_STAT_ADD macro.
>    This change belongs in the first patch of the series.
> v2:
> * Move the statistics counters into a stats structure.
> 
> Signed-off-by: Morten Brørup <mb@smartsharesystems.com>
> ---
>   lib/mempool/rte_mempool.c |  9 ++++++
>   lib/mempool/rte_mempool.h | 66 ++++++++++++++++++++++++++++++++-------
>   2 files changed, 64 insertions(+), 11 deletions(-)
> 
> diff --git a/lib/mempool/rte_mempool.c b/lib/mempool/rte_mempool.c
> index e6208125e0..a18e39af04 100644
> --- a/lib/mempool/rte_mempool.c
> +++ b/lib/mempool/rte_mempool.c
> @@ -1286,6 +1286,15 @@ rte_mempool_dump(FILE *f, struct rte_mempool *mp)
>   		sum.get_success_blks += mp->stats[lcore_id].get_success_blks;
>   		sum.get_fail_blks += mp->stats[lcore_id].get_fail_blks;
>   	}
> +	if (mp->cache_size != 0) {
> +		/* Add the statistics stored in the mempool caches. */
> +		for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++) {
> +			sum.put_bulk += mp->local_cache[lcore_id].stats.put_bulk;
> +			sum.put_objs += mp->local_cache[lcore_id].stats.put_objs;
> +			sum.get_success_bulk += mp->local_cache[lcore_id].stats.get_success_bulk;
> +			sum.get_success_objs += mp->local_cache[lcore_id].stats.get_success_objs;
> +		}
> +	}
>   	fprintf(f, "  stats:\n");
>   	fprintf(f, "    put_bulk=%"PRIu64"\n", sum.put_bulk);
>   	fprintf(f, "    put_objs=%"PRIu64"\n", sum.put_objs);
> diff --git a/lib/mempool/rte_mempool.h b/lib/mempool/rte_mempool.h
> index abfe34c05f..e6eb573739 100644
> --- a/lib/mempool/rte_mempool.h
> +++ b/lib/mempool/rte_mempool.h
> @@ -86,6 +86,19 @@ 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 */
> +#ifdef RTE_LIBRTE_MEMPOOL_STATS
> +	uint32_t unused;
> +	/*
> +	 * Alternative location for the most frequently updated mempool statistics (per-lcore),
> +	 * providing faster update access when using a mempool cache.
> +	 */
> +	struct {
> +		uint64_t put_bulk;          /**< Number of puts. */
> +		uint64_t put_objs;          /**< Number of objects successfully put. */
> +		uint64_t get_success_bulk;  /**< Successful allocation number. */
> +		uint64_t get_success_objs;  /**< Objects successfully allocated. */
> +	} stats;                        /**< Statistics */
> +#endif
>   	/**
>   	 * Cache objects
>   	 *
> @@ -319,6 +332,22 @@ struct rte_mempool {
>   #define RTE_MEMPOOL_STAT_ADD(mp, name, n) do {} while (0)
>   #endif
>   
> +/**
> + * @internal When stats is enabled, store some statistics.
> + *
> + * @param cache
> + *   Pointer to the memory pool cache.
> + * @param name
> + *   Name of the statistics field to increment in the memory pool cache.
> + * @param n
> + *   Number to add to the statistics.
> + */
> +#ifdef RTE_LIBRTE_MEMPOOL_STATS
> +#define RTE_MEMPOOL_CACHE_STAT_ADD(cache, name, n) (cache)->stats.name += n
> +#else
> +#define RTE_MEMPOOL_CACHE_STAT_ADD(cache, name, n) do {} while (0)
> +#endif
> +
>   /**
>    * @internal Calculate the size of the mempool header.
>    *
> @@ -1333,13 +1362,17 @@ rte_mempool_do_generic_put(struct rte_mempool *mp, void * const *obj_table,
>   {
>   	void **cache_objs;
>   
> +	/* No cache provided */
> +	if (unlikely(cache == NULL))
> +		goto driver_enqueue;
> +
>   	/* increment stat now, adding in mempool always success */
> -	RTE_MEMPOOL_STAT_ADD(mp, put_bulk, 1);
> -	RTE_MEMPOOL_STAT_ADD(mp, put_objs, n);
> +	RTE_MEMPOOL_CACHE_STAT_ADD(cache, put_bulk, 1);
> +	RTE_MEMPOOL_CACHE_STAT_ADD(cache, put_objs, n);
>   
> -	/* No cache provided or the request itself is too big for the cache */
> -	if (unlikely(cache == NULL || n > cache->flushthresh))
> -		goto driver_enqueue;
> +	/* The request itself is too big for the cache */
> +	if (unlikely(n > cache->flushthresh))
> +		goto driver_enqueue_stats_incremented;
>   
>   	/*
>   	 * The cache follows the following algorithm:
> @@ -1364,6 +1397,12 @@ rte_mempool_do_generic_put(struct rte_mempool *mp, void * const *obj_table,
>   
>   driver_enqueue:
>   
> +	/* increment stat now, adding in mempool always success */
> +	RTE_MEMPOOL_STAT_ADD(mp, put_bulk, 1);
> +	RTE_MEMPOOL_STAT_ADD(mp, put_objs, n);
> +
> +driver_enqueue_stats_incremented:
> +
>   	/* push objects to the backend */
>   	rte_mempool_ops_enqueue_bulk(mp, obj_table, n);
>   }
> @@ -1470,8 +1509,8 @@ rte_mempool_do_generic_get(struct rte_mempool *mp, void **obj_table,
>   	if (remaining == 0) {
>   		/* The entire request is satisfied from the cache. */
>   
> -		RTE_MEMPOOL_STAT_ADD(mp, get_success_bulk, 1);
> -		RTE_MEMPOOL_STAT_ADD(mp, get_success_objs, n);
> +		RTE_MEMPOOL_CACHE_STAT_ADD(cache, get_success_bulk, 1);
> +		RTE_MEMPOOL_CACHE_STAT_ADD(cache, get_success_objs, n);
>   
>   		return 0;
>   	}
> @@ -1500,8 +1539,8 @@ rte_mempool_do_generic_get(struct rte_mempool *mp, void **obj_table,
>   
>   	cache->len = cache->size;
>   
> -	RTE_MEMPOOL_STAT_ADD(mp, get_success_bulk, 1);
> -	RTE_MEMPOOL_STAT_ADD(mp, get_success_objs, n);
> +	RTE_MEMPOOL_CACHE_STAT_ADD(cache, get_success_bulk, 1);
> +	RTE_MEMPOOL_CACHE_STAT_ADD(cache, get_success_objs, n);
>   
>   	return 0;
>   
> @@ -1523,8 +1562,13 @@ rte_mempool_do_generic_get(struct rte_mempool *mp, void **obj_table,
>   		RTE_MEMPOOL_STAT_ADD(mp, get_fail_bulk, 1);
>   		RTE_MEMPOOL_STAT_ADD(mp, get_fail_objs, n);
>   	} else {
> -		RTE_MEMPOOL_STAT_ADD(mp, get_success_bulk, 1);
> -		RTE_MEMPOOL_STAT_ADD(mp, get_success_objs, n);
> +		if (likely(cache != NULL)) {
> +			RTE_MEMPOOL_CACHE_STAT_ADD(cache, get_success_bulk, 1);
> +			RTE_MEMPOOL_CACHE_STAT_ADD(cache, get_success_objs, n);
> +		} else {
> +			RTE_MEMPOOL_STAT_ADD(mp, get_success_bulk, 1);
> +			RTE_MEMPOOL_STAT_ADD(mp, get_success_objs, n);
> +		}
>   	}
>   
>   	return ret;

For the series,
Reviewed-By: Mattias Rönnblom <mattias.ronnblom@ericsson.com>

  parent reply	other threads:[~2022-11-07  7:30 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-30 11:54 [PATCH] mempool: split statistics from debug Morten Brørup
2022-10-30 14:04 ` Morten Brørup
2022-10-30 16:12   ` Stephen Hemminger
2022-10-30 20:29     ` Morten Brørup
2022-10-31 11:26 ` [PATCH v2 1/3] " Morten Brørup
2022-10-31 11:26   ` [PATCH v2 2/3] mempool: include non-DPDK threads in statistics Morten Brørup
2022-11-02  7:52     ` Mattias Rönnblom
2022-11-02  9:09       ` Morten Brørup
2022-11-02 15:19         ` Stephen Hemminger
2022-11-02 15:37           ` Morten Brørup
2022-11-02 17:53         ` Mattias Rönnblom
2022-11-03  8:59           ` Morten Brørup
2022-11-04  8:58             ` Mattias Rönnblom
2022-11-04 10:01               ` Morten Brørup
2022-11-07  7:26                 ` Mattias Rönnblom
2022-11-07  8:56                   ` Morten Brørup
2022-10-31 11:26   ` [PATCH v2 3/3] mempool: use cache for frequently updated statistics Morten Brørup
2022-11-02  8:01     ` Mattias Rönnblom
2022-11-02  9:29       ` Morten Brørup
2022-11-02 17:55         ` Mattias Rönnblom
2022-11-04 11:17   ` [PATCH v3 1/3] mempool: split stats from debug Morten Brørup
2022-11-04 11:17     ` [PATCH v3 2/3] mempool: add stats for unregistered non-EAL threads Morten Brørup
2022-11-04 11:17     ` [PATCH v3 3/3] mempool: use cache for frequently updated stats Morten Brørup
2022-11-04 12:03     ` [PATCH v4 1/3] mempool: split stats from debug Morten Brørup
2022-11-04 12:03       ` [PATCH v4 2/3] mempool: add stats for unregistered non-EAL threads Morten Brørup
2022-11-06 11:34         ` Andrew Rybchenko
2022-11-04 12:03       ` [PATCH v4 3/3] mempool: use cache for frequently updated stats Morten Brørup
2022-11-06 11:40         ` Andrew Rybchenko
2022-11-06 11:50           ` Morten Brørup
2022-11-06 11:59             ` Andrew Rybchenko
2022-11-06 12:16               ` Morten Brørup
2022-11-07  7:30         ` Mattias Rönnblom [this message]
2022-11-08  9:20         ` Konstantin Ananyev
2022-11-08 11:21           ` Morten Brørup
2022-11-06 11:32       ` [PATCH v4 1/3] mempool: split stats from debug Andrew Rybchenko
2022-11-09 18:18       ` [PATCH v5 " Morten Brørup
2022-11-09 18:18         ` [PATCH v5 2/3] mempool: add stats for unregistered non-EAL threads Morten Brørup
2022-11-09 18:18         ` [PATCH v5 3/3] mempool: use cache for frequently updated stats Morten Brørup
2022-11-10 16:36           ` 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=532b74d2-37f3-7f0d-f7bc-f8f7034b6559@lysator.liu.se \
    --to=hofors@lysator.liu.se \
    --cc=andrew.rybchenko@oktetlabs.ru \
    --cc=bruce.richardson@intel.com \
    --cc=dev@dpdk.org \
    --cc=jerinj@marvell.com \
    --cc=mattias.ronnblom@ericsson.com \
    --cc=mb@smartsharesystems.com \
    --cc=olivier.matz@6wind.com \
    --cc=stephen@networkplumber.org \
    --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).