From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mails.dpdk.org (mails.dpdk.org [217.70.189.124]) by inbox.dpdk.org (Postfix) with ESMTP id 36E5343E98; Thu, 18 Apr 2024 02:10:58 +0200 (CEST) Received: from mails.dpdk.org (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id 8DE6B402CC; Thu, 18 Apr 2024 02:10:57 +0200 (CEST) Received: from linux.microsoft.com (linux.microsoft.com [13.77.154.182]) by mails.dpdk.org (Postfix) with ESMTP id 5405B40272 for ; Thu, 18 Apr 2024 02:10:56 +0200 (CEST) Received: by linux.microsoft.com (Postfix, from userid 1086) id A619420FD4C4; Wed, 17 Apr 2024 17:10:55 -0700 (PDT) DKIM-Filter: OpenDKIM Filter v2.11.0 linux.microsoft.com A619420FD4C4 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.microsoft.com; s=default; t=1713399055; bh=F0v8qCd4jQ5HvWK0edqwhr+z9lOfjh6vN/+tw52TGU0=; h=Date:From:To:Cc:Subject:References:In-Reply-To:From; b=aCawvPQpQeKtdI/fQjmIwp7fUNFSucJaw9qphSZJDThBInIRWqujPIKfWW3rynA4v bOEzJ/RpN2LKsDmCBhbcQmqpe11+8Icqbkf1vsh9YJ9DsuWhI1ppVzGxsxzfQjJl8Z DgtY78kuzQXU8RVCgzZDISFYY1wQtbv5NJRkJW+o= Date: Wed, 17 Apr 2024 17:10:55 -0700 From: Tyler Retzlaff To: Stephen Hemminger Cc: dev@dpdk.org, Reshma Pattan Subject: Re: [PATCH v3 3/5] latencystats: do not use floating point Message-ID: <20240418001055.GC26881@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net> References: <0240408195036.182545-1-stephen@networkplumber.org> <20240417170908.76701-1-stephen@networkplumber.org> <20240417170908.76701-4-stephen@networkplumber.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20240417170908.76701-4-stephen@networkplumber.org> User-Agent: Mutt/1.5.21 (2010-09-15) X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org On Wed, Apr 17, 2024 at 10:07:25AM -0700, Stephen Hemminger wrote: > The cycle counts do not need to be stored as floating point. > Instead keep track of latency in cycles, and convert to > nanoseconds when read. > > Change Exponential Weighted Moving Average weight from .2 to .25 > to avoid use of floating point for that. > > Signed-off-by: Stephen Hemminger > --- > lib/latencystats/rte_latencystats.c | 37 +++++++++++------------------ > 1 file changed, 14 insertions(+), 23 deletions(-) > > diff --git a/lib/latencystats/rte_latencystats.c b/lib/latencystats/rte_latencystats.c > index fe8c3c563a..11bd0ea4ae 100644 > --- a/lib/latencystats/rte_latencystats.c > +++ b/lib/latencystats/rte_latencystats.c > @@ -47,10 +47,10 @@ static uint64_t timer_tsc; > static uint64_t prev_tsc; > > struct rte_latency_stats { > - float min_latency; /**< Minimum latency in nano seconds */ > - float avg_latency; /**< Average latency in nano seconds */ > - float max_latency; /**< Maximum latency in nano seconds */ > - float jitter; /** Latency variation */ > + uint64_t min_latency; /**< Minimum latency */ > + uint64_t avg_latency; /**< Average latency */ > + uint64_t max_latency; /**< Maximum latency */ > + uint64_t jitter; /** Latency variation */ > rte_spinlock_t lock; /** Latency calculation lock */ > }; > > @@ -82,13 +82,12 @@ int32_t > rte_latencystats_update(void) > { > unsigned int i; > - float *stats_ptr = NULL; > uint64_t values[NUM_LATENCY_STATS] = {0}; > int ret; > > for (i = 0; i < NUM_LATENCY_STATS; i++) { > - stats_ptr = RTE_PTR_ADD(glob_stats, > - lat_stats_strings[i].offset); > + const uint64_t *stats_ptr = RTE_PTR_ADD(glob_stats, > + lat_stats_strings[i].offset); > values[i] = floor(*stats_ptr / latencystat_cycles_per_ns()); will this need a cast to uint64_t? > } > > @@ -105,11 +104,10 @@ static void > rte_latencystats_fill_values(struct rte_metric_value *values) > { > unsigned int i; > - float *stats_ptr = NULL; > > for (i = 0; i < NUM_LATENCY_STATS; i++) { > - stats_ptr = RTE_PTR_ADD(glob_stats, > - lat_stats_strings[i].offset); > + const uint64_t *stats_ptr = RTE_PTR_ADD(glob_stats, > + lat_stats_strings[i].offset); > values[i].key = i; > values[i].value = floor(*stats_ptr / latencystat_cycles_per_ns()); > } > @@ -156,15 +154,8 @@ calc_latency(uint16_t pid __rte_unused, > void *_ __rte_unused) > { > unsigned int i; > - uint64_t now; > - float latency; > - static float prev_latency; > - /* > - * Alpha represents degree of weighting decrease in EWMA, > - * a constant smoothing factor between 0 and 1. The value > - * is used below for measuring average latency. > - */ > - const float alpha = 0.2; > + uint64_t now, latency; > + static uint64_t prev_latency; > if this is merged i can abandon this series https://patchwork.dpdk.org/project/dpdk/list/?series=31747 Acked-by: Tyler Retzlaff