From: Yoan Picchi <yoan.picchi@foss.arm.com>
To: Pavan Nikhilesh Bhagavatula <pbhagavatula@marvell.com>,
Yoan Picchi <yoan.picchi@arm.com>,
Yipeng Wang <yipeng1.wang@intel.com>,
Sameh Gobriel <sameh.gobriel@intel.com>,
Bruce Richardson <bruce.richardson@intel.com>,
Vladimir Medvedkin <vladimir.medvedkin@intel.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>, "nd@arm.com" <nd@arm.com>,
Ruifeng Wang <ruifeng.wang@arm.com>,
Nathan Brown <nathan.brown@arm.com>,
Jerin Jacob <jerinj@marvell.com>
Subject: Re: [EXTERNAL] [PATCH v7 2/4] hash: optimize compare signature for NEON
Date: Thu, 11 Apr 2024 14:32:53 +0100 [thread overview]
Message-ID: <cbcca228-0b4f-4b95-8841-e30ef09143e6@foss.arm.com> (raw)
In-Reply-To: <PH0PR18MB408670D0ECB88BEFE5A750EADE332@PH0PR18MB4086.namprd18.prod.outlook.com>
On 3/20/24 07:37, Pavan Nikhilesh Bhagavatula wrote:
>> Upon a successful comparison, NEON sets all the bits in the lane to 1
>> We can skip shifting by simply masking with specific masks.
>>
>> Signed-off-by: Yoan Picchi <yoan.picchi@arm.com>
>> Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com>
>> Reviewed-by: Nathan Brown <nathan.brown@arm.com>
>> ---
>> lib/hash/arch/arm/compare_signatures.h | 24 +++++++++++-------------
>> 1 file changed, 11 insertions(+), 13 deletions(-)
>>
>> diff --git a/lib/hash/arch/arm/compare_signatures.h
>> b/lib/hash/arch/arm/compare_signatures.h
>> index 1af6ba8190..b5a457f936 100644
>> --- a/lib/hash/arch/arm/compare_signatures.h
>> +++ b/lib/hash/arch/arm/compare_signatures.h
>> @@ -30,23 +30,21 @@ compare_signatures_dense(uint16_t
>> *hitmask_buffer,
>> switch (sig_cmp_fn) {
>> #if RTE_HASH_BUCKET_ENTRIES <= 8
>> case RTE_HASH_COMPARE_NEON: {
>> - uint16x8_t vmat, vsig, x;
>> - int16x8_t shift = {0, 1, 2, 3, 4, 5, 6, 7};
>> - uint16_t low, high;
>> + uint16x8_t vmat, hit1, hit2;
>> + const uint16x8_t mask = {0x1, 0x2, 0x4, 0x8, 0x10, 0x20,
>> 0x40, 0x80};
>> + const uint16x8_t vsig = vld1q_dup_u16((uint16_t const
>> *)&sig);
>>
>> - vsig = vld1q_dup_u16((uint16_t const *)&sig);
>> /* Compare all signatures in the primary bucket */
>> - vmat = vceqq_u16(vsig,
>> - vld1q_u16((uint16_t const *)prim_bucket_sigs));
>> - x = vshlq_u16(vandq_u16(vmat, vdupq_n_u16(0x0001)),
>> shift);
>> - low = (uint16_t)(vaddvq_u16(x));
>> + vmat = vceqq_u16(vsig, vld1q_u16(prim_bucket_sigs));
>> + hit1 = vandq_u16(vmat, mask);
>> +
>> /* Compare all signatures in the secondary bucket */
>> - vmat = vceqq_u16(vsig,
>> - vld1q_u16((uint16_t const *)sec_bucket_sigs));
>> - x = vshlq_u16(vandq_u16(vmat, vdupq_n_u16(0x0001)),
>> shift);
>> - high = (uint16_t)(vaddvq_u16(x));
>> - *hitmask_buffer = low | high << RTE_HASH_BUCKET_ENTRIES;
>> + vmat = vceqq_u16(vsig, vld1q_u16(sec_bucket_sigs));
>> + hit2 = vandq_u16(vmat, mask);
>>
>> + hit2 = vshlq_n_u16(hit2, RTE_HASH_BUCKET_ENTRIES);
>> + hit2 = vorrq_u16(hit1, hit2);
>> + *hitmask_buffer = vaddvq_u16(hit2);
>
> Since vaddv is expensive could you convert it to vshrn?
>
> https://community.arm.com/arm-community-blogs/b/infrastructure-solutions-blog/posts/porting-x86-vector-bitmask-optimizations-to-arm-neon
>
> https://github.com/DPDK/dpdk/blob/main/examples/l3fwd/l3fwd_neon.h#L226
Thank you for those links, it was a good read.
Unfortunatly I don't think it is a good use case here. A decent part of
the speedup I get is by using a dense hitmask: ie every bit count with
no padding. Using the vshrn would have 4 bits of padding, and stripping
them would be more expensive than using a regular reduce.
>
>> }
>> break;
>> #endif
>> --
>> 2.25.1
>
next prev parent reply other threads:[~2024-04-11 13:33 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-10-20 16:51 [PATCH v2 0/4] hash: add SVE support for bulk key lookup Yoan Picchi
2023-10-20 16:51 ` [PATCH v2 1/4] hash: pack the hitmask for hash in bulk lookup Yoan Picchi
2023-10-20 16:51 ` [PATCH v2 2/4] hash: optimize compare signature for NEON Yoan Picchi
2023-10-20 16:51 ` [PATCH v2 3/4] test/hash: check bulk lookup of keys after collision Yoan Picchi
2023-10-20 16:51 ` [PATCH v2 4/4] hash: add SVE support for bulk key lookup Yoan Picchi
2024-02-27 17:41 ` [PATCH v5 0/4] " Yoan Picchi
2024-02-27 17:42 ` [PATCH v5 1/4] hash: pack the hitmask for hash in bulk lookup Yoan Picchi
2024-02-27 17:42 ` [PATCH v5 2/4] hash: optimize compare signature for NEON Yoan Picchi
2024-02-27 17:42 ` [PATCH v5 3/4] test/hash: check bulk lookup of keys after collision Yoan Picchi
2024-02-27 17:42 ` [PATCH v5 4/4] hash: add SVE support for bulk key lookup Yoan Picchi
2024-02-28 10:56 ` Konstantin Ananyev
2024-02-28 14:48 ` Yoan Picchi
2024-03-04 13:35 ` Konstantin Ananyev
2024-03-05 15:36 ` Yoan Picchi
2024-03-11 23:21 ` [PATCH v6 0/4] " Yoan Picchi
2024-03-11 23:21 ` [PATCH v6 1/4] hash: pack the hitmask for hash in bulk lookup Yoan Picchi
2024-03-11 23:21 ` [PATCH v6 2/4] hash: optimize compare signature for NEON Yoan Picchi
2024-03-11 23:21 ` [PATCH v6 3/4] test/hash: check bulk lookup of keys after collision Yoan Picchi
2024-03-11 23:21 ` [PATCH v6 4/4] hash: add SVE support for bulk key lookup Yoan Picchi
2024-03-12 3:57 ` fengchengwen
2024-03-12 15:08 ` Yoan Picchi
2024-03-12 15:42 ` [PATCH v7 0/4] " Yoan Picchi
2024-03-12 15:42 ` [PATCH v7 1/4] hash: pack the hitmask for hash in bulk lookup Yoan Picchi
2024-03-19 10:41 ` Konstantin Ananyev
2024-03-19 13:09 ` Yoan Picchi
2024-03-19 13:25 ` Konstantin Ananyev
2024-03-19 16:09 ` Stephen Hemminger
2024-03-12 15:42 ` [PATCH v7 2/4] hash: optimize compare signature for NEON Yoan Picchi
2024-03-20 7:37 ` [EXTERNAL] " Pavan Nikhilesh Bhagavatula
2024-04-11 13:32 ` Yoan Picchi [this message]
2024-03-12 15:42 ` [PATCH v7 3/4] test/hash: check bulk lookup of keys after collision Yoan Picchi
2024-03-12 15:42 ` [PATCH v7 4/4] hash: add SVE support for bulk key lookup Yoan Picchi
2024-04-17 16:08 ` [PATCH v8 0/4] " Yoan Picchi
2024-04-17 16:08 ` [PATCH v8 1/4] hash: pack the hitmask for hash in bulk lookup Yoan Picchi
2024-04-17 18:12 ` Stephen Hemminger
2024-04-17 16:08 ` [PATCH v8 2/4] hash: optimize compare signature for NEON Yoan Picchi
2024-04-17 16:08 ` [PATCH v8 3/4] test/hash: check bulk lookup of keys after collision Yoan Picchi
2024-04-17 16:08 ` [PATCH v8 4/4] hash: add SVE support for bulk key lookup Yoan Picchi
2024-04-30 16:27 ` [PATCH v9 0/4] " Yoan Picchi
2024-04-30 16:27 ` [PATCH v9 1/4] hash: pack the hitmask for hash in bulk lookup Yoan Picchi
2024-04-30 16:27 ` [PATCH v9 2/4] hash: optimize compare signature for NEON Yoan Picchi
2024-04-30 16:27 ` [PATCH v9 3/4] test/hash: check bulk lookup of keys after collision Yoan Picchi
2024-04-30 16:27 ` [PATCH v9 4/4] hash: add SVE support for bulk key lookup Yoan Picchi
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=cbcca228-0b4f-4b95-8841-e30ef09143e6@foss.arm.com \
--to=yoan.picchi@foss.arm.com \
--cc=bruce.richardson@intel.com \
--cc=dev@dpdk.org \
--cc=jerinj@marvell.com \
--cc=nathan.brown@arm.com \
--cc=nd@arm.com \
--cc=pbhagavatula@marvell.com \
--cc=ruifeng.wang@arm.com \
--cc=sameh.gobriel@intel.com \
--cc=vladimir.medvedkin@intel.com \
--cc=yipeng1.wang@intel.com \
--cc=yoan.picchi@arm.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).