From: "Liu, Yong" <yong.liu@intel.com>
To: "Peng, Yuan" <yuan.peng@intel.com>, "dts@dpdk.org" <dts@dpdk.org>
Cc: "Peng, Yuan" <yuan.peng@intel.com>
Subject: Re: [dts] [PATCH v2] test_plans: add test plan of bbdev
Date: Tue, 27 Mar 2018 08:19:47 +0000 [thread overview]
Message-ID: <86228AFD5BCD8E4EBFD2B90117B5E81E62FCB595@SHSMSX103.ccr.corp.intel.com> (raw)
In-Reply-To: <1521514932-88552-1-git-send-email-yuan.peng@intel.com>
Yuan, few comments are inline.
> -----Original Message-----
> From: dts [mailto:dts-bounces@dpdk.org] On Behalf Of Peng Yuan
> Sent: Tuesday, March 20, 2018 11:02 AM
> To: dts@dpdk.org
> Cc: Peng, Yuan <yuan.peng@intel.com>
> Subject: [dts] [PATCH v2] test_plans: add test plan of bbdev
>
> BBDEV allow DPDK as a 3gpp Layer 1 signal processing acceleration solution.
>
> Signed-off-by: Peng Yuan <yuan.peng@intel.com>
>
> diff --git a/test_plans/bbdev_test_plan.rst
> b/test_plans/bbdev_test_plan.rst
> new file mode 100644
> index 0000000..93ba6a7
> --- /dev/null
> +++ b/test_plans/bbdev_test_plan.rst
> @@ -0,0 +1,203 @@
> +.. Copyright (c) <2018>, Intel Corporation
> + All rights reserved.
> +
> + 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 Intel Corporation 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.
> +
> +=============================================================
> +Wireless device for ICX-D (bbdev) for Turbo decoding/encoding
> +=============================================================
> +Description
> +===========
> +
> + The Wireless Baseband library provides a common programming framework
> that
> + abstracts HW accelerators based on FPGA and/or Fixed Function
> Accelerators
> + that assist with 3gpp Physical Layer processing. Furthermore, it
> decouples
> + the application from the compute-intensive wireless functions by
> abstracting
> + their optimized libraries to appear as virtual bbdev devices.
> + The functional scope of the BBDEV library are those functions in
> relation to
> + the 3gpp Layer 1 signal processing (channel coding, modulation, ...).
> + The framework currently only supports Turbo Code FEC function.
> + The Wireless Baseband library follows the same ideology of DPDK's
> Ethernet
> + Device and Crypto Device frameworks. Wireless Baseband provides a
> generic
> + acceleration abstraction framework which supports both physical
> (hardware)
> + and virtual (software) wireless acceleration functions.
> + Now we only support virtual device, In the case of the hardware, of
> course
> + there will be change in results. This is the aim of having hardware
> + acceleration to get the latency and throughput improved.
> +
> +Prerequisites
> +=============
> +
> +1. OS and Hardware
> + (a) An AVX2 supporting machine
> + (b) Windriver TS 2 or CentOS 7 operating systems
> + (Fedora 25 and Ubuntu 16.04 is ok.)
> + (c) Intel ICC compiler installed(GCC compiler can be used too.)
> +
> +2. FlexRAN SDK Libraries
> + To build DPDK with the *turbo_sw* PMD the user is required to download
> + the export controlled ``FlexRAN SDK`` Libraries.
> + An account at Intel Resource Design Center needs to be registered from
> + https://www.intel.com/content/www/us/en/design/resource-design-
> center.html
> + Direct download link from Intel RDC ->
> + https://cdrd.intel.com/v1/dl/getContent/575367
> + After download is complete, the user needs to unpack and compile on
> their
> + system before building DPDK.
> + You can refer to the file dpdk/doc/guides/bbdevs/turbo_sw.rst.
> +
> +3. PMD setting
> + Current BBDEV framework is en-suited with two vdev PMD drivers:
> + null and turbo_sw.
> + 1) Null PMD is similar to cryptodev Null PMD, which is an empty driver
> to
> + measure the overhead added by the framework.
> + 2) Turbo_sw is a sw-only driver wrapper for FlexRAN SDK optimized
> Turbo
> + coding libraries.
> + It can be enabled by setting
> +
> + ``CONFIG_RTE_LIBRTE_PMD_BBDEV_TURBO_SW=y``
> +
> + It is located in the build configuration file ``common_base``.
> +
> +4. Test tool
> + A test suite for BBDEV is packaged with the framework to ease the
> + validation needs for various functions and use cases.
> + The tool to use for validation and testing is called: test-bbdev,
> + that is packaged with test vectors that are ready-to-use.
> + Test-bbdev tool is located at this location:
> + app/test-bbdev/
> + The command-line options you can refer to:
> + dpdk/doc/guides/tools/testbbdev.rst
> +
> +Test case 1: bbdev null device
> +==============================
> +
Yuan, please remove test case index which is not needed.
> +Executing bbdev null device with *bbdev_vector_null.data* helps in
> measuring
> +the overhead introduced by the bbdev framework::
> +
> + ./test-bbdev.py -e="--vdev=bbdev_null0"
> + -v ./test_vectors/bbdev_vector_null.data
> +
> +bbdev_null device does not have to be defined explicitly as it is created
> +by default. so the command-line can be written as::
> +
> + ./test-bbdev.py -v ./test_vectors/bbdev_vector_null.data
> +
> +the case only cover unittest, all the other cases are skipped.
> +
> +Test case 2: Turbo encoding validation
> +======================================
> +
> +It runs **validation** test for Turbo encode vector file
> +Number of operations to process on device is set to 64
> +and operations timeout is set to 120s
> +and enqueue/dequeue burst size is set to 8 and to 32.
> +Moreover a bbdev (*turbo_sw*) device will be created::
> +
> + ./test-bbdev.py -p ../../x86_64-native-linuxapp-icc/app/testbbdev
> + -e="--vdev=turbo_sw" -t 120 -c validation
> + -v ./test_vectors/bbdev_vector_te_default.data -n 64 -b 8 32
> +
> +you can check the result from the detailed printing, and compare
> +the two results from different burst size.
> +
Yuan, could you please give one simple sample for the result?
And please also point out what kind of check is needed.
> +Test case 3: Turbo decoding validation
> +======================================
> +
> +It runs **validation** test for Turbo decode vector file
> +we use default options::
> +
> + ./test-bbdev.py -e="--vdev=turbo_sw" -c validation
> + -v ./test_vectors/bbdev_vector_td_default.data
> +
> +you can check the result from the detailed printing.
Same as previous one.
> +
> +Test case 4: Turbo encoding latency
> +===================================
> +
> +It runs **latency** test for Turbo encode vector file::
> +
> + ./test-bbdev.py -e="--vdev=turbo_sw" -c latency
> + -v ./test_vectors/bbdev_vector_te_default.data -n 512 -b 64 512
> +
> +Test calculates three latency metrics:
> +1) offload_latency_tc
> +2) offload_latency_empty_q_tc
> +3) operation_latency_tc
> +You can compare the three latency from different burst size.
Same as previous one.
> +
> +Test case 5: Turbo decoding latency
> +===================================
> +
> +It runs **latency** test for Turbo decode vector file::
> +
> + ./test-bbdev.py -e="--vdev=turbo_sw" -c latency
> + -v ./test_vectors/bbdev_vector_td_default.data -n 512 -b 64
> +
> + ./test-bbdev.py -e="--vdev=turbo_sw" -c latency
> + -v ./test_vectors/bbdev_vector_td_default.data -n 128 -b 64
> +
> +You can compare the three latency from different number of operations.
> +
> +Test case 6: Turbo encoding throughput
> +======================================
> +
> +It runs **throughput** test for Turbo encode vector file::
> +
> + ./test-bbdev.py -e="--vdev=turbo_sw" -c throughput
> + -v ./test_vectors/bbdev_vector_te_default.data -n 64 -l 16 -b 64
> +
> + ./test-bbdev.py -e="--vdev=turbo_sw" -c throughput
> + -v ./test_vectors/bbdev_vector_te_default.data -n 64 -l 8 -b 64
> +
> +You can compare the turbo encode throughput from different number of
> lcores.
> +then different burst size::
> +
> + ./test-bbdev.py -e="--vdev=turbo_sw" -c throughput
> + -v ./test_vectors/bbdev_vector_te_default.data -n 64 -l 16 -b 64
> +
> + ./test-bbdev.py -e="--vdev=turbo_sw" -c throughput
> + -v ./test_vectors/bbdev_vector_te_default.data -n 64 -l 16 -b 32
> +
> +You can compare the turbo encode throughput from different burst size.
> +then different number of operations::
> +
> + ./test-bbdev.py -e="--vdev=turbo_sw" -c throughput
> + -v ./test_vectors/bbdev_vector_te_default.data -n 64 -l 16 -b 32
> +
> + ./test-bbdev.py -e="--vdev=turbo_sw" -c throughput
> + -v ./test_vectors/bbdev_vector_te_default.data -n 32 -l 16 -b 32
> +
> +You can compare the turbo encode throughput from different number of
> +operations.
> +
> +Test case 7: Turbo decoding throughput
> +======================================
> +
> +It runs **throughput** test for Turbo decode vector file.
> +you can run the three subcases of test case 6 which just needed to
> +change the test vector file to "bbdev_vector_td_default.data", and
> +compare the results.
> --
> 2.5.0
prev parent reply other threads:[~2018-03-27 8:19 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-03-20 3:02 Peng Yuan
2018-03-27 8:19 ` Liu, Yong [this message]
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=86228AFD5BCD8E4EBFD2B90117B5E81E62FCB595@SHSMSX103.ccr.corp.intel.com \
--to=yong.liu@intel.com \
--cc=dts@dpdk.org \
--cc=yuan.peng@intel.com \
/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).