[dts] [PATCH v2] test_plans/unit_test_eal: add hash relative unit test plan
Tu, Lijuan
lijuan.tu at intel.com
Mon Jan 7 09:19:03 CET 2019
Applied, thanks
> -----Original Message-----
> From: dts [mailto:dts-bounces at dpdk.org] On Behalf Of Xueqin Lin
> Sent: Friday, December 28, 2018 5:05 AM
> To: dts at dpdk.org
> Cc: Lin, Xueqin <xueqin.lin at intel.com>
> Subject: [dts] [PATCH v2] test_plans/unit_test_eal: add hash relative unit test
> plan
>
> Signed-off-by: Xueqin Lin <xueqin.lin at intel.com>
>
> ---
> test_plans/unit_tests_eal_test_plan.rst | 84
> +++++++++++++++++++++++++++++++++
> 1 file changed, 84 insertions(+)
>
> diff --git a/test_plans/unit_tests_eal_test_plan.rst
> b/test_plans/unit_tests_eal_test_plan.rst
> index 201fd30..33b0f14 100644
> --- a/test_plans/unit_tests_eal_test_plan.rst
> +++ b/test_plans/unit_tests_eal_test_plan.rst
> @@ -313,3 +313,87 @@ To Be Filled
> Link_bonding
> ============
> To Be Filled
> +
> +Hash
> +====
> +This does unit function test for hash features:
> +
> +- Average table utilization when disable extendable table function
> +- Average table utilization when enable extendable table function,
> + check could reach 100% utilization
> +
> +
> +Hash_perf
> +=========
> +This does the performance test with a single thread, including the
> +cases with and without extendable table:
> +
> +- Measure cycles for add, lookup, lookup_bulk, delete
> +- With/without pre-computed hash values
> +- For different key lengths
> +
> +
> +Hash_functions
> +==============
> +This does unit test for hash functions:
> +
> +- Measure cycles for hashing
> +- Jhash vs rte_hash_crc
> +- For different key lenthgs, seeds
> +
> +
> +Hash_multiwriter
> +================
> +This does the performance and function test of multi-threads case –
> +multiple writers.
> +
> +Introduce scalable multi-writer Cuckoo Hash insertion based on a split
> +cuckoo search and move operation using Intel TSX. It can do scalable
> +hash insertion with 22 cores with little performance loss and
> +negligible TSX abortion rate.
> +
> +
> +Hash_readwrite
> +==============
> +This does the performance and function test of multi-threads case –
> +multiple reader/writer.
> +
> +Read-write concurrency support in rte_hash. A new flag value is added
> +to indicate if read-write concurrency is needed during creation time.
> +The new concurrency model is based on rte_rwlock. When Intel TSX is
> +available and the users indicate to use it, the TM version of the
> +rte_rwlock will be called. Both multi-writer and read-write concurrency
> +are protected by the rte_rwlock instead of the x86 specific RTM
> +instructions, so the x86 specific header rte_cuckoo_hash_x86.h is
> +removed and the code is infused into the main .c file.
> +A new rte_hash_count API is proposed to count how many keys are
> +inserted into the hash table.
> +
> +
> +Hash_hash_readwrite_lf
> +======================
> +This does the unit tests to check for hash lookup and bulk-lookup perf
> +with lock-free enabled and with lock-free disabled. Unit tests
> +performed with readers running in parallel with writers.
> +Tests include:
> +
> +- Hash lookup on existing keys
> +
> + - Hash add causing NO key-shifts of existing keys in the table
> +
> +- Hash lookup on existing keys likely to be on shift-path
> +
> + - Hash add causing key-shifts of existing keys in the table
> +
> +- Hash lookup on existing keys NOT likely to be on shift-path
> +
> + - Hash add causing key-shifts of existing keys in the table
> +
> +- Hash lookup on non-existing keys
> +
> + - Hash add causing NO key-shifts of existing keys in the table
> + - Hash add causing key-shifts of existing keys in the table
> +
> +- Hash lookup on keys likely to be on shift-path
> +
> + - Multiple writers causing key-shifts of existing keys in the table
> --
> 2.5.5
More information about the dts
mailing list