DPDK patches and discussions
 help / color / mirror / Atom feed
From: "Burakov, Anatoly" <anatoly.burakov@intel.com>
To: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>,
	"dev@dpdk.org" <dev@dpdk.org>
Cc: "Dumitrescu, Cristian" <cristian.dumitrescu@intel.com>,
	"stephen@networkplumber.org" <stephen@networkplumber.org>
Subject: Re: [dpdk-dev] [PATCH v3 3/3] test: add tests for reciprocal based	division
Date: Mon, 4 Sep 2017 09:18:37 +0000	[thread overview]
Message-ID: <C6ECDF3AB251BE4894318F4E451236978229362E@IRSMSX109.ger.corp.intel.com> (raw)
In-Reply-To: <1504442189-4384-3-git-send-email-pbhagavatula@caviumnetworks.com>

Hi Pavan,

> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Pavan Nikhilesh
> Sent: Sunday, September 3, 2017 1:36 PM
> To: dev@dpdk.org
> Cc: Dumitrescu, Cristian <cristian.dumitrescu@intel.com>;
> stephen@networkplumber.org; Pavan Nikhilesh
> <pbhagavatula@caviumnetworks.com>
> Subject: [dpdk-dev] [PATCH v3 3/3] test: add tests for reciprocal based
> division
> 
> This commit provides a set of tests for verifying the correctness and
> performance of both unsigned 32 and 64bit reciprocal based division.
> 
> Signed-off-by: Pavan Nikhilesh <pbhagavatula@caviumnetworks.com>
> ---
>  test/test/Makefile                        |   2 +
>  test/test/test_reciprocal_division.c      | 109 ++++++++++++++++++
>  test/test/test_reciprocal_division_perf.c | 181
> ++++++++++++++++++++++++++++++
>  3 files changed, 292 insertions(+)
>  create mode 100644 test/test/test_reciprocal_division.c
>  create mode 100644 test/test/test_reciprocal_division_perf.c
> 
> diff --git a/test/test/Makefile b/test/test/Makefile index 42d9a49..6017862
> 100644
> --- a/test/test/Makefile
> +++ b/test/test/Makefile
> @@ -94,6 +94,8 @@ SRCS-y += test_cycles.c  SRCS-y += test_spinlock.c
> SRCS-y += test_memory.c  SRCS-y += test_memzone.c
> +SRCS-y += test_reciprocal_division.c
> +SRCS-y += test_reciprocal_division_perf.c
> 
>  SRCS-y += test_ring.c
>  SRCS-y += test_ring_perf.c
> diff --git a/test/test/test_reciprocal_division.c
> b/test/test/test_reciprocal_division.c
> new file mode 100644
> index 0000000..771ea64
> --- /dev/null
> +++ b/test/test/test_reciprocal_division.c
> @@ -0,0 +1,109 @@
> +/*
> + *   BSD LICENSE
> + *
> + *   Copyright (C) Cavium, Inc. 2017.
> + *
> + *   Redistribution and use in source and binary forms, with or without
> + *   modification, are permitted provided that the following conditions
> + *   are met:
> + *
> + *     * Redistributions of source code must retain the above copyright
> + *       notice, this list of conditions and the following disclaimer.
> + *     * Redistributions in binary form must reproduce the above copyright
> + *       notice, this list of conditions and the following disclaimer in
> + *       the documentation and/or other materials provided with the
> + *       distribution.
> + *     * Neither the name of Cavium, Inc nor the names of its
> + *       contributors may be used to endorse or promote products derived
> + *       from this software without specific prior written permission.
> + *
> + *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
> CONTRIBUTORS
> + *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT
> NOT
> + *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
> FITNESS FOR
> + *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
> COPYRIGHT
> + *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
> INCIDENTAL,
> + *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
> NOT
> + *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS
> OF USE,
> + *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
> AND ON ANY
> + *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR
> TORT
> + *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
> THE USE
> + *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH
> DAMAGE.
> + */
> +
> +#include "test.h"
> +
> +#include <stdio.h>
> +#include <unistd.h>
> +#include <inttypes.h>
> +
> +#include <rte_common.h>
> +#include <rte_cycles.h>
> +#include <rte_random.h>
> +#include <rte_reciprocal.h>
> +
> +#define MAX_ITERATIONS	1000000
> +#define DIVIDE_ITER		100
> +
> +static int
> +test_reciprocal_division(void)
> +{
> +	int i;
> +	int result = 0;
> +	uint32_t divisor_u32 = 0;
> +	uint32_t dividend_u32;
> +	uint32_t nresult_u32;
> +	uint32_t rresult_u32;
> +	uint64_t divisor_u64 = 0;
> +	uint64_t dividend_u64;
> +	uint64_t nresult_u64;
> +	uint64_t rresult_u64;
> +	struct rte_reciprocal_u32 reci_u32;
> +	struct rte_reciprocal_u64 reci_u64;
> +
> +	rte_srand(rte_rdtsc());
> +	printf("Validating unsigned 32bit division.\n");
> +	for (i = 0; i < MAX_ITERATIONS; i++) {
> +		/* Change divisor every DIVIDE_ITER iterations. */
> +		if (i % DIVIDE_ITER == 0) {
> +			divisor_u32 = rte_rand();
> +			reci_u32 = rte_reciprocal_value_u32(divisor_u32);
> +		}
> +
> +		dividend_u32 = rte_rand();
> +		nresult_u32 = dividend_u32 / divisor_u32;
> +		rresult_u32 = rte_reciprocal_divide_u32(dividend_u32,
> +				&reci_u32);
> +		if (nresult_u32 != rresult_u32) {
> +			printf("Division failed, expected %"PRIu32" "
> +				   "result %"PRIu32"",
> +					nresult_u32, rresult_u32);
> +			result = 1;
> +			break;
> +		}
> +	}
> +
> +	printf("Validating unsigned 64bit division.\n");
> +	for (i = 0; i < MAX_ITERATIONS; i++) {
> +		/* Change divisor every DIVIDE_ITER iterations. */
> +		if (i % DIVIDE_ITER == 0) {
> +			divisor_u64 = rte_rand();
> +			reci_u64 = rte_reciprocal_value_u64(divisor_u64);
> +		}
> +
> +		dividend_u64 = rte_rand();
> +		nresult_u64 = dividend_u64 / divisor_u64;
> +		rresult_u64 = rte_reciprocal_divide_u64(dividend_u64,
> +				&reci_u64);
> +		if (nresult_u64 != rresult_u64) {
> +			printf("Division failed, expected %"PRIu64" "
> +				   "result %"PRIu64"",
> +					nresult_u64, rresult_u64);
> +			result = 1;
> +			break;
> +		}
> +	}
> +
> +	return result;
> +}
> +
> +REGISTER_TEST_COMMAND(reciprocal_division, test_reciprocal_division);
> diff --git a/test/test/test_reciprocal_division_perf.c
> b/test/test/test_reciprocal_division_perf.c
> new file mode 100644
> index 0000000..31f069d
> --- /dev/null
> +++ b/test/test/test_reciprocal_division_perf.c
> @@ -0,0 +1,181 @@
> +/*
> + *   BSD LICENSE
> + *
> + *   Copyright (C) Cavium, Inc. 2017.
> + *
> + *   Redistribution and use in source and binary forms, with or without
> + *   modification, are permitted provided that the following conditions
> + *   are met:
> + *
> + *     * Redistributions of source code must retain the above copyright
> + *       notice, this list of conditions and the following disclaimer.
> + *     * Redistributions in binary form must reproduce the above copyright
> + *       notice, this list of conditions and the following disclaimer in
> + *       the documentation and/or other materials provided with the
> + *       distribution.
> + *     * Neither the name of Cavium, Inc nor the names of its
> + *       contributors may be used to endorse or promote products derived
> + *       from this software without specific prior written permission.
> + *
> + *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
> CONTRIBUTORS
> + *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT
> NOT
> + *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
> FITNESS FOR
> + *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
> COPYRIGHT
> + *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
> INCIDENTAL,
> + *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
> NOT
> + *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS
> OF USE,
> + *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
> AND ON ANY
> + *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR
> TORT
> + *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
> THE USE
> + *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH
> DAMAGE.
> + */
> +
> +#include "test.h"
> +
> +#include <stdio.h>
> +#include <unistd.h>
> +#include <inttypes.h>
> +
> +#include <rte_common.h>
> +#include <rte_cycles.h>
> +#include <rte_random.h>
> +#include <rte_reciprocal.h>
> +
> +#define MAX_ITERATIONS	1000000
> +#define DIVIDE_ITER		100
> +
> +static int
> +test_reciprocal_division_perf(void)
> +{
> +	int result = 0;
> +	uint32_t divisor_u32 = 0;
> +	uint32_t dividend_u32;
> +	uint32_t nresult_u32;
> +	uint32_t rresult_u32;
> +	uint64_t divisor_u64 = 0;
> +	uint64_t dividend_u64;
> +	uint64_t nresult_u64;
> +	uint64_t rresult_u64;
> +	uint64_t cur_cyc;
> +	uint64_t tot_cyc_n = 0;
> +	uint64_t tot_cyc_r = 0;
> +	uint64_t i;
> +	struct rte_reciprocal_u32 reci_u32 = {0};
> +	struct rte_reciprocal_u64 reci_u64 = {0};
> +
> +	rte_srand(rte_rdtsc());
> +	printf("Validating unsigned 32bit division.\n");
> +	for (i = 0; i < MAX_ITERATIONS; i++) {
> +		/* Change divisor every DIVIDE_ITER iterations. */
> +		if (i % DIVIDE_ITER == 0) {
> +			divisor_u32 = rte_rand();
> +			reci_u32 = rte_reciprocal_value_u32(divisor_u32);
> +		}
> +
> +		dividend_u32 = rte_rand();
> +		cur_cyc = rte_rdtsc();
> +		nresult_u32 = dividend_u32 / divisor_u32;
> +		tot_cyc_n += rte_rdtsc() - cur_cyc;
> +
> +		cur_cyc = rte_rdtsc();

Is there any particular reason for calling rte_rdtsc() twice in a row? Instead of, say, doing something like this:

start = rdtsc();
// do something
split = rdtsc();
// do something else
end = rdtsc();
// figure out which cycle value goes where

> +		rresult_u32 = rte_reciprocal_divide_u32(dividend_u32,
> +				&reci_u32);
> +		tot_cyc_r += rte_rdtsc() - cur_cyc;
> +		if (nresult_u32 != rresult_u32) {
> +			printf("Division failed, expected %"PRIu32" "
> +					"result %"PRIu32"",
> +					nresult_u32, rresult_u32);
> +			result = 1;
> +			break;
> +		}
> +	}
> +	printf("32bit Division results:\n");
> +	printf("Total number of cycles normal division     : %"PRIu64"\n",
> +			tot_cyc_n);
> +	printf("Total number of cycles reciprocal division : %"PRIu64"\n",
> +			tot_cyc_r);
> +	printf("Cycles per division(normal) : %3.2f\n",
> +			((double)tot_cyc_n)/i);
> +	printf("Cycles per division(normal) : %3.2f\n\n",
> +			((double)tot_cyc_r)/i);
> +
> +	tot_cyc_n = 0;
> +	tot_cyc_r = 0;
> +	printf("Validating unsigned 64bit division.\n");
> +
> +	for (i = 0; i < MAX_ITERATIONS; i++) {
> +		/* Change divisor every DIVIDE_ITER iterations. */
> +		if (i % DIVIDE_ITER == 0) {
> +			divisor_u64 = rte_rand();
> +			reci_u64 = rte_reciprocal_value_u64(divisor_u64);
> +		}
> +
> +		dividend_u64 = rte_rand();
> +		cur_cyc = rte_rdtsc();
> +		nresult_u64 = dividend_u64 / divisor_u64;
> +		tot_cyc_n += rte_rdtsc() - cur_cyc;
> +
> +		cur_cyc = rte_rdtsc();

Same as above

> +		rresult_u64 = rte_reciprocal_divide_u64(dividend_u64,
> +				&reci_u64);
> +		tot_cyc_r += rte_rdtsc() - cur_cyc;
> +		if (nresult_u64 != rresult_u64) {
> +			printf("Division failed, expected %"PRIu64" "
> +					"result %"PRIu64"",
> +					nresult_u64, rresult_u64);
> +			result = 1;
> +			break;
> +		}
> +	}
> +	printf("64bit Division results:\n");
> +	printf("Total number of cycles normal division     : %"PRIu64"\n",
> +			tot_cyc_n);
> +	printf("Total number of cycles reciprocal division : %"PRIu64"\n",
> +			tot_cyc_r);
> +	printf("Cycles per division(normal) : %3.2f\n",
> +			((double)tot_cyc_n)/i);
> +	printf("Cycles per division(normal) : %3.2f\n\n",
> +			((double)tot_cyc_r)/i);
> +
> +	printf("Validating unsigned 64bit division with 32bit divisor.\n");
> +	for (i = 0; i < MAX_ITERATIONS; i++) {
> +		/* Change divisor every DIVIDE_ITER iterations. */
> +		if (i % DIVIDE_ITER == 0) {
> +			divisor_u64 = rte_rand() >> 32;
> +			reci_u64 = rte_reciprocal_value_u64(divisor_u64);
> +		}
> +
> +		dividend_u64 = rte_rand();
> +		cur_cyc = rte_rdtsc();
> +		nresult_u64 = dividend_u64 / divisor_u64;
> +		tot_cyc_n += rte_rdtsc() - cur_cyc;
> +
> +		cur_cyc = rte_rdtsc();
> +		rresult_u64 = rte_reciprocal_divide_u64(dividend_u64,
> +				&reci_u64);
> +		tot_cyc_r += rte_rdtsc() - cur_cyc;
> +		if (nresult_u64 != rresult_u64) {
> +			printf("Division failed, expected %"PRIu64" "
> +					"result %"PRIu64"",
> +					nresult_u64, rresult_u64);
> +			result = 1;
> +			break;
> +		}
> +	}
> +	printf("64bit Division results:\n");
> +	printf("Total number of cycles normal division     : %"PRIu64"\n",
> +			tot_cyc_n);
> +	printf("Total number of cycles reciprocal division : %"PRIu64"\n",
> +			tot_cyc_r);
> +	printf("Cycles per division(normal) : %3.2f\n",
> +			((double)tot_cyc_n)/i);
> +	printf("Cycles per division(normal) : %3.2f\n",
> +			((double)tot_cyc_r)/i);
> +
> +	tot_cyc_n = 0;
> +	tot_cyc_r = 0;
> +
> +	return result;
> +}
> +
> +REGISTER_TEST_COMMAND(reciprocal_division_perf,
> +test_reciprocal_division_perf);
> --
> 2.7.4

  reply	other threads:[~2017-09-04  9:18 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-09-03 12:36 [dpdk-dev] [PATCH v3 1/3] eal: introduce integer divide through reciprocal Pavan Nikhilesh
2017-09-03 12:36 ` [dpdk-dev] [PATCH v3 2/3] eal: add u64 bit variant for reciprocal Pavan Nikhilesh
2017-09-04 14:34   ` Burakov, Anatoly
2017-09-04 14:55     ` Pavan Nikhilesh Bhagavatula
2017-09-03 12:36 ` [dpdk-dev] [PATCH v3 3/3] test: add tests for reciprocal based division Pavan Nikhilesh
2017-09-04  9:18   ` Burakov, Anatoly [this message]
2017-09-04  9:49     ` Pavan Nikhilesh Bhagavatula
2017-09-04 14:01 ` [dpdk-dev] [PATCH v3 1/3] eal: introduce integer divide through reciprocal Burakov, Anatoly

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=C6ECDF3AB251BE4894318F4E451236978229362E@IRSMSX109.ger.corp.intel.com \
    --to=anatoly.burakov@intel.com \
    --cc=cristian.dumitrescu@intel.com \
    --cc=dev@dpdk.org \
    --cc=pbhagavatula@caviumnetworks.com \
    --cc=stephen@networkplumber.org \
    /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).