From: "Medvedkin, Vladimir" <vladimir.medvedkin@intel.com>
To: Stephen Hemminger <stephen@networkplumber.org>,
"Ananyev, Konstantin" <konstantin.ananyev@intel.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>,
"Wang, Yipeng1" <yipeng1.wang@intel.com>,
"Gobriel, Sameh" <sameh.gobriel@intel.com>,
"Richardson, Bruce" <bruce.richardson@intel.com>
Subject: Re: [dpdk-dev] [PATCH v2 1/5] hash: add new toeplitz hash implementation
Date: Tue, 19 Oct 2021 17:42:41 +0200 [thread overview]
Message-ID: <fc25a654-00a9-63fc-fc5b-6653b7f8a8a0@intel.com> (raw)
In-Reply-To: <20211018181523.24f9657d@hermes.local>
Hi Stephen,
On 19/10/2021 03:15, Stephen Hemminger wrote:
> On Mon, 18 Oct 2021 10:40:00 +0000
> "Ananyev, Konstantin" <konstantin.ananyev@intel.com> wrote:
>
>>> On Fri, 15 Oct 2021 10:30:02 +0100
>>> Vladimir Medvedkin <vladimir.medvedkin@intel.com> wrote:
>>>
>>>> + m[i * 8 + j] = (rss_key[i] << j)|
>>>> + (uint8_t)((uint16_t)(rss_key[i + 1]) >>
>>>> + (8 - j));
>>>> + }
>>>
>>> This ends up being harder than necessary to read. Maybe split into
>>> multiple statements and/or use temporary variable.
>>>
>>>> +RTE_INIT(rte_thash_gfni_init)
>>>> +{
>>>> + rte_thash_gfni_supported = 0;
>>>
>>> Not necessary in C globals are initialized to zero by default.
>>>
>>> By removing that the constructor can be totally behind #ifdef
>>>
>>>> +__rte_internal
>>>> +static inline __m512i
>>>> +__rte_thash_gfni(const uint64_t *mtrx, const uint8_t *tuple,
>>>> + const uint8_t *secondary_tuple, int len)
>>>> +{
>>>> + __m512i permute_idx = _mm512_set_epi8(7, 6, 5, 4, 7, 6, 5, 4,
>>>> + 6, 5, 4, 3, 6, 5, 4, 3,
>>>> + 5, 4, 3, 2, 5, 4, 3, 2,
>>>> + 4, 3, 2, 1, 4, 3, 2, 1,
>>>> + 3, 2, 1, 0, 3, 2, 1, 0,
>>>> + 2, 1, 0, -1, 2, 1, 0, -1,
>>>> + 1, 0, -1, -2, 1, 0, -1, -2,
>>>> + 0, -1, -2, -3, 0, -1, -2, -3);
>>>
>>> NAK
>>>
>>> Please don't put the implementation in an inline. This makes it harder
>>> to support (API/ABI) and blocks other architectures from implementing
>>> same thing with different instructions.
>>
>> I don't really understand your reasoning here.
>> rte_thash_gfni.h is an arch-specific header, which provides
>> arch-specific optimizations for RSS hash calculation
>> (Vladimir pls correct me if I am wrong here).
>
> Ok, but rte_thash_gfni.h is included on all architectures.
>
Ok, I'll rework the patch to move x86 + avx512 related things into x86
arch specific header. Would that suit?
>> We do have dozens of inline functions that do use arch-specific instructions (both x86 and arm)
>> for different purposes:
>> sync primitives, memory-ordering, cache manipulations, LPM lookup, TSX, power-saving, etc.
>> That's a usual trade-off taken for performance reasons, when extra function call
>> costs too much comparing to the operation itself.
>> Why it suddenly became a problem for that particular case and how exactly it blocks other architectures?
>> Also I don't understand how it makes things harder in terms of API/ABI stability.
>> As I can see this patch doesn't introduce any public structs/unions.
>> All functions take as arguments just raw data buffers and length.
>> To summarize - in general, I don't see any good reason why this patch shouldn't be allowed.
>> Konstantin
>
> The comments about rte_thash_gfni_supported initialization still apply.
> Why not:
>
> #ifdef __GFNI__
> RTE_INIT(rte_thash_gfni_init)
> {
> if (rte_cpu_get_flag_enabled(RTE_CPUFLAG_GFNI))
> rte_thash_gfni_supported = 1;
> }
> #endif
>
Agree, I'll reflect this changes in v3.
--
Regards,
Vladimir
next prev parent reply other threads:[~2021-10-19 15:50 UTC|newest]
Thread overview: 72+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-09-06 16:03 [dpdk-dev] [PATCH 0/5] optimized Toeplitz " Vladimir Medvedkin
2021-09-06 16:03 ` [dpdk-dev] [PATCH 1/5] hash: add new toeplitz " Vladimir Medvedkin
2021-10-07 18:23 ` Ananyev, Konstantin
2021-10-08 11:19 ` Ananyev, Konstantin
2021-10-15 9:11 ` Medvedkin, Vladimir
2021-10-15 10:55 ` Ananyev, Konstantin
2021-10-15 13:09 ` Medvedkin, Vladimir
2021-09-06 16:03 ` [dpdk-dev] [PATCH 2/5] hash: enable gfni thash implementation Vladimir Medvedkin
2021-10-08 11:31 ` Ananyev, Konstantin
2021-10-15 9:13 ` Medvedkin, Vladimir
2021-09-06 16:03 ` [dpdk-dev] [PATCH 3/5] doc/hash: update documentation for the thash library Vladimir Medvedkin
2021-09-06 16:03 ` [dpdk-dev] [PATCH 4/5] test/thash: add tests for a new Toeplitz hash function Vladimir Medvedkin
2021-09-07 0:35 ` Stephen Hemminger
2021-09-08 13:59 ` Medvedkin, Vladimir
2021-09-06 16:03 ` [dpdk-dev] [PATCH 5/5] test/thash: add performance tests for the Toeplitz hash Vladimir Medvedkin
2021-10-15 9:30 ` [dpdk-dev] [PATCH v2 0/5] optimized Toeplitz hash implementation Vladimir Medvedkin
2021-10-20 18:20 ` [dpdk-dev] [PATCH v3 " Vladimir Medvedkin
2021-10-21 17:18 ` [dpdk-dev] [PATCH v4 " Vladimir Medvedkin
2021-10-21 18:54 ` [dpdk-dev] [PATCH v5 " Vladimir Medvedkin
2021-10-21 18:54 ` [dpdk-dev] [PATCH v5 1/5] hash: add new toeplitz " Vladimir Medvedkin
2021-10-25 17:05 ` Thomas Monjalon
2021-10-21 18:54 ` [dpdk-dev] [PATCH v5 2/5] hash: enable gfni thash implementation Vladimir Medvedkin
2021-10-21 18:54 ` [dpdk-dev] [PATCH v5 3/5] doc/hash: update documentation for the thash library Vladimir Medvedkin
2021-10-25 17:04 ` Thomas Monjalon
2021-10-26 20:30 ` Medvedkin, Vladimir
2021-10-21 18:54 ` [dpdk-dev] [PATCH v5 4/5] test/thash: add tests for a new Toeplitz hash function Vladimir Medvedkin
2021-10-21 18:54 ` [dpdk-dev] [PATCH v5 5/5] test/thash: add performance tests for the Toeplitz hash Vladimir Medvedkin
2021-10-25 17:02 ` Thomas Monjalon
2021-10-26 20:29 ` Medvedkin, Vladimir
2021-10-27 8:29 ` Thomas Monjalon
2021-10-27 15:48 ` Medvedkin, Vladimir
2021-10-25 17:27 ` Stephen Hemminger
2021-10-26 20:31 ` Medvedkin, Vladimir
2021-10-21 17:18 ` [dpdk-dev] [PATCH v4 1/5] hash: add new toeplitz hash implementation Vladimir Medvedkin
2021-10-21 17:18 ` [dpdk-dev] [PATCH v4 2/5] hash: enable gfni thash implementation Vladimir Medvedkin
2021-10-21 17:18 ` [dpdk-dev] [PATCH v4 3/5] doc/hash: update documentation for the thash library Vladimir Medvedkin
2021-10-21 17:18 ` [dpdk-dev] [PATCH v4 4/5] test/thash: add tests for a new Toeplitz hash function Vladimir Medvedkin
2021-10-21 17:18 ` [dpdk-dev] [PATCH v4 5/5] test/thash: add performance tests for the Toeplitz hash Vladimir Medvedkin
2021-10-20 18:20 ` [dpdk-dev] [PATCH v3 1/5] hash: add new toeplitz hash implementation Vladimir Medvedkin
2021-10-21 9:42 ` Ananyev, Konstantin
2021-10-21 17:17 ` Medvedkin, Vladimir
2021-10-20 18:20 ` [dpdk-dev] [PATCH v3 2/5] hash: enable gfni thash implementation Vladimir Medvedkin
2021-10-21 9:46 ` Ananyev, Konstantin
2021-10-20 18:20 ` [dpdk-dev] [PATCH v3 3/5] doc/hash: update documentation for the thash library Vladimir Medvedkin
2021-10-20 18:20 ` [dpdk-dev] [PATCH v3 4/5] test/thash: add tests for a new Toeplitz hash function Vladimir Medvedkin
2021-10-20 18:20 ` [dpdk-dev] [PATCH v3 5/5] test/thash: add performance tests for the Toeplitz hash Vladimir Medvedkin
2021-10-15 9:30 ` [dpdk-dev] [PATCH v2 1/5] hash: add new toeplitz hash implementation Vladimir Medvedkin
2021-10-15 16:58 ` Stephen Hemminger
2021-10-18 10:40 ` Ananyev, Konstantin
2021-10-19 1:15 ` Stephen Hemminger
2021-10-19 15:42 ` Medvedkin, Vladimir [this message]
2021-10-18 11:08 ` Medvedkin, Vladimir
2021-10-15 9:30 ` [dpdk-dev] [PATCH v2 2/5] hash: enable gfni thash implementation Vladimir Medvedkin
2021-10-15 9:30 ` [dpdk-dev] [PATCH v2 3/5] doc/hash: update documentation for the thash library Vladimir Medvedkin
2021-10-15 9:30 ` [dpdk-dev] [PATCH v2 4/5] test/thash: add tests for a new Toeplitz hash function Vladimir Medvedkin
2021-10-15 9:30 ` [dpdk-dev] [PATCH v2 5/5] test/thash: add performance tests for the Toeplitz hash Vladimir Medvedkin
2021-10-26 20:32 ` [dpdk-dev] [PATCH v6 0/4] optimized Toeplitz hash implementation Vladimir Medvedkin
2021-10-26 20:32 ` [dpdk-dev] [PATCH v6 1/4] hash: add new toeplitz " Vladimir Medvedkin
2021-10-26 20:32 ` [dpdk-dev] [PATCH v6 2/4] hash: add bulk " Vladimir Medvedkin
2021-10-26 20:32 ` [dpdk-dev] [PATCH v6 3/4] hash: enable gfni thash implementation Vladimir Medvedkin
2021-10-26 20:32 ` [dpdk-dev] [PATCH v6 4/4] test/thash: add performance tests for the Toeplitz hash Vladimir Medvedkin
2021-10-27 16:16 ` [dpdk-dev] [PATCH v7 0/4] optimized Toeplitz hash implementation Vladimir Medvedkin
2021-10-27 16:16 ` [dpdk-dev] [PATCH v7 1/4] hash: add new toeplitz " Vladimir Medvedkin
2021-10-27 16:16 ` [dpdk-dev] [PATCH v7 2/4] hash: add bulk " Vladimir Medvedkin
2021-10-27 16:16 ` [dpdk-dev] [PATCH v7 3/4] hash: enable gfni thash implementation Vladimir Medvedkin
2021-10-27 16:16 ` [dpdk-dev] [PATCH v7 4/4] test/thash: add performance tests for the Toeplitz hash Vladimir Medvedkin
2021-11-02 18:38 ` [dpdk-dev] [PATCH v8 0/4] optimized Toeplitz hash implementation Vladimir Medvedkin
2021-11-04 10:20 ` Thomas Monjalon
2021-11-02 18:38 ` [dpdk-dev] [PATCH v8 1/4] hash: add new toeplitz " Vladimir Medvedkin
2021-11-02 18:38 ` [dpdk-dev] [PATCH v8 2/4] hash: add bulk " Vladimir Medvedkin
2021-11-02 18:38 ` [dpdk-dev] [PATCH v8 3/4] hash: enable gfni thash implementation Vladimir Medvedkin
2021-11-02 18:38 ` [dpdk-dev] [PATCH v8 4/4] test/thash: add performance tests for the Toeplitz hash Vladimir Medvedkin
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=fc25a654-00a9-63fc-fc5b-6653b7f8a8a0@intel.com \
--to=vladimir.medvedkin@intel.com \
--cc=bruce.richardson@intel.com \
--cc=dev@dpdk.org \
--cc=konstantin.ananyev@intel.com \
--cc=sameh.gobriel@intel.com \
--cc=stephen@networkplumber.org \
--cc=yipeng1.wang@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).