From: "Ruifeng Wang (Arm Technology China)" <Ruifeng.Wang@arm.com>
To: "jerinj@marvell.com" <jerinj@marvell.com>,
"yipeng1.wang@intel.com" <yipeng1.wang@intel.com>
Cc: Honnappa Nagarahalli <Honnappa.Nagarahalli@arm.com>,
nd <nd@arm.com>, "dev@dpdk.org" <dev@dpdk.org>, nd <nd@arm.com>
Subject: Re: [dpdk-dev] [EXT] [PATCH v1] hash: optimize signature compare by using neon intrinsic
Date: Mon, 11 Feb 2019 09:47:10 +0000 [thread overview]
Message-ID: <AM6PR08MB37828611F424F19435EEBDBE9E640@AM6PR08MB3782.eurprd08.prod.outlook.com> (raw)
In-Reply-To: <8c27de0eb69bc17562e64f977044f59cb7169770.camel@marvell.com>
Hi Jerin,
> -----Original Message-----
> From: Jerin Jacob Kollanukkaran <jerinj@marvell.com>
> Sent: Monday, February 11, 2019 15:49
> To: Ruifeng Wang (Arm Technology China) <Ruifeng.Wang@arm.com>;
> yipeng1.wang@intel.com
> Cc: Honnappa Nagarahalli <Honnappa.Nagarahalli@arm.com>; nd
> <nd@arm.com>; dev@dpdk.org
> Subject: Re: [EXT] [PATCH v1] hash: optimize signature compare by using
> neon intrinsic
>
> On Mon, 2019-02-11 at 15:30 +0800, Ruifeng Wang wrote:
> >
> > -------------------------------------------------------------------
> > ---
> > Implemented signature compare function based on neon intrinsic.
> > Hash bulk lookup had 3% - 6% performance gain after optimization.
> >
> > Signed-off-by: Ruifeng Wang <ruifeng.wang@arm.com>
> > ---
> > lib/librte_hash/rte_cuckoo_hash.c | 32
> > ++++++++++++++++++++++++++++++-
> > lib/librte_hash/rte_cuckoo_hash.h | 1 +
> > 2 files changed, 32 insertions(+), 1 deletion(-)
> >
> > diff --git a/lib/librte_hash/rte_cuckoo_hash.c
> > b/lib/librte_hash/rte_cuckoo_hash.c
> > index c01489ba5..5745a254f 100644
> > --- a/lib/librte_hash/rte_cuckoo_hash.c
> > +++ b/lib/librte_hash/rte_cuckoo_hash.c
> > @@ -26,6 +26,9 @@
> > #include <rte_spinlock.h>
> > #include <rte_ring.h>
> > #include <rte_compat.h>
> > +#if defined(RTE_ARCH_ARM64)
> > +#include <arm_neon.h>
> > +#endif
>
>
> The use of rte_vector.h will remove the need for #if defined...
>
OK. Will change to rte_vect.h in next version.
>
> >
> > #include "rte_hash.h"
> > #include "rte_cuckoo_hash.h"
> > @@ -407,6 +410,10 @@ rte_hash_create(const struct
> rte_hash_parameters
> > *params)
> > if (rte_cpu_get_flag_enabled(RTE_CPUFLAG_SSE2))
> > h->sig_cmp_fn = RTE_HASH_COMPARE_SSE;
> > else
> > +#elif defined(RTE_ARCH_ARM64)
> > + if (rte_cpu_get_flag_enabled(RTE_CPUFLAG_NEON))
> > + h->sig_cmp_fn = RTE_HASH_COMPARE_NEON;
> > + else
> > #endif
> > h->sig_cmp_fn = RTE_HASH_COMPARE_SCALAR;
> >
> > @@ -1578,10 +1585,15 @@ compare_signatures(uint32_t
> > *prim_hash_matches, uint32_t *sec_hash_matches,
> > enum rte_hash_sig_compare_function sig_cmp_fn)
> {
> > unsigned int i;
> > +#ifdef RTE_MACHINE_CPUFLAG_NEON
> > + uint16x8_t vmat, vsig, x;
> > + uint64x2_t x64;
> > + int16x8_t shift = {-15, -13, -11, -9, -7, -5, -3, -1}; #endif
>
> Is it possible move down the variable declaration? to avoid the need for
> #ifdef here
>
Will move down the variable declaration into 'case' block.
> >
> > /* For match mask the first bit of every two bits indicates the
> > match */
> > switch (sig_cmp_fn) {
> > -#ifdef RTE_MACHINE_CPUFLAG_SSE2
> > +#if defined(RTE_MACHINE_CPUFLAG_SSE2)
> > case RTE_HASH_COMPARE_SSE:
> > /* Compare all signatures in the bucket */
> > *prim_hash_matches =
> _mm_movemask_epi8(_mm_cmpeq_epi16(
> > @@ -1594,6 +1606,24 @@ compare_signatures(uint32_t
> *prim_hash_matches,
> > uint32_t *sec_hash_matches,
> > (__m128i const *)sec_bkt-
> > >sig_current),
> > _mm_set1_epi16(sig)));
> > break;
> > +#elif defined(RTE_MACHINE_CPUFLAG_NEON)
> > + case RTE_HASH_COMPARE_NEON:
> > + 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_bkt-
> > >sig_current));
> > + x = vshlq_u16(vandq_u16(vmat, vdupq_n_u16(0x8000)),
> > shift);
> > + x64 = vpaddlq_u32(vpaddlq_u16(x));
> > + *prim_hash_matches = (uint32_t)(vgetq_lane_u64(x64, 0)
> > +
> > + vgetq_lane_u64(x64, 1));
> > + /* Compare all signatures in the secondary bucket */
> > + vmat = vceqq_u16(vsig,
> > + vld1q_u16((uint16_t const *)sec_bkt-
> > >sig_current));
> > + x = vshlq_u16(vandq_u16(vmat, vdupq_n_u16(0x8000)),
> > shift);
> > + x64 = vpaddlq_u32(vpaddlq_u16(x));
> > + *sec_hash_matches = (uint32_t)(vgetq_lane_u64(x64, 0) +
> > + vgetq_lane_u64(x64, 1));
> > + break;
> > #endif
> > default:
> > for (i = 0; i < RTE_HASH_BUCKET_ENTRIES; i++) { diff --git
> > a/lib/librte_hash/rte_cuckoo_hash.h
> > b/lib/librte_hash/rte_cuckoo_hash.h
> > index eacdaa8d4..0548c97f0 100644
> > --- a/lib/librte_hash/rte_cuckoo_hash.h
> > +++ b/lib/librte_hash/rte_cuckoo_hash.h
> > @@ -141,6 +141,7 @@ struct rte_hash_key { enum
> > rte_hash_sig_compare_function {
> > RTE_HASH_COMPARE_SCALAR = 0,
> > RTE_HASH_COMPARE_SSE,
> > + RTE_HASH_COMPARE_NEON,
> > RTE_HASH_COMPARE_NUM
> > };
> >
prev parent reply other threads:[~2019-02-11 9:47 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-02-11 7:30 [dpdk-dev] " Ruifeng Wang
2019-02-11 7:49 ` [dpdk-dev] [EXT] " Jerin Jacob Kollanukkaran
2019-02-11 9:47 ` Ruifeng Wang (Arm Technology China) [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=AM6PR08MB37828611F424F19435EEBDBE9E640@AM6PR08MB3782.eurprd08.prod.outlook.com \
--to=ruifeng.wang@arm.com \
--cc=Honnappa.Nagarahalli@arm.com \
--cc=dev@dpdk.org \
--cc=jerinj@marvell.com \
--cc=nd@arm.com \
--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).