From: Peng Yuan <yuan.peng@intel.com>
To: dts@dpdk.org
Cc: Peng Yuan <yuan.peng@intel.com>
Subject: [dts] [PATCH v2] test_plans: add test plan of bbdev
Date: Tue, 20 Mar 2018 11:02:12 +0800 [thread overview]
Message-ID: <1521514932-88552-1-git-send-email-yuan.peng@intel.com> (raw)
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
+==============================
+
+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.
+
+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.
+
+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.
+
+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
next reply other threads:[~2018-03-20 3:00 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-03-20 3:02 Peng Yuan [this message]
2018-03-27 8:19 ` Liu, Yong
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=1521514932-88552-1-git-send-email-yuan.peng@intel.com \
--to=yuan.peng@intel.com \
--cc=dts@dpdk.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).