[EXT] Re: [PATCH 2/2] config/arm: disable SVE for cn10k

[Ruifeng Wang]

> -----Original Message-----
> From: fengchengwen <fengchengwen at huawei.com>
> Sent: Friday, May 6, 2022 2:36 PM
> To: Rahul Bhansali <rbhansali at marvell.com>; dev at dpdk.org; Ruifeng Wang
> <Ruifeng.Wang at arm.com>; Jan Viktorin <viktorin at rehivetech.com>; Bruce
> Richardson <bruce.richardson at intel.com>
> Cc: jerinj at marvell.com
> Subject: Re: [EXT] Re: [PATCH 2/2] config/arm: disable SVE for cn10k
> 
> On 2022/5/6 12:54, Rahul Bhansali wrote:
> >
> >
> >> -----Original Message-----
> >> From: fengchengwen <fengchengwen at huawei.com>
> >> Sent: Friday, May 6, 2022 8:00 AM
> >> To: Rahul Bhansali <rbhansali at marvell.com>; dev at dpdk.org; Ruifeng
> >> Wang <ruifeng.wang at arm.com>; Jan Viktorin
> <viktorin at rehivetech.com>;
> >> Bruce Richardson <bruce.richardson at intel.com>
> >> Cc: Jerin Jacob Kollanukkaran <jerinj at marvell.com>
> >> Subject: [EXT] Re: [PATCH 2/2] config/arm: disable SVE for cn10k
> >>
> >> External Email
> >>
> >> ---------------------------------------------------------------------
> >> - On 2022/5/5 22:27, Rahul Bhansali wrote:
> >>> This disable the SVE flag for cn10k.
> >>>
> >>> Performance impact:-
> >>> With l3fwd example, lpm lookup performance increased by ~21% if
> Neon
> >>> is used instead of SVE.
> >>>
> >>> Signed-off-by: Rahul Bhansali <rbhansali at marvell.com>
> >>> ---
> >>>  config/arm/meson.build | 3 ++-
> >>>  1 file changed, 2 insertions(+), 1 deletion(-)
> >>>
> >>> diff --git a/config/arm/meson.build b/config/arm/meson.build index
> >>> dafb342cc6..39b7a1270c 100644
> >>> --- a/config/arm/meson.build
> >>> +++ b/config/arm/meson.build
> >>> @@ -281,7 +281,8 @@ soc_cn10k = {
> >>>      ],
> >>>      'part_number': '0xd49',
> >>>      'extra_march_features': ['crypto'],
> >>> -    'numa': false
> >>> +    'numa': false,
> >>> +    'sve': false
> >>
> >> Suggest remove sve2 flag:
> >>     '0xd49': {
> >>         'march': 'armv8.5-a',
> >>         'march_features': ['sve2'],          ---remove 'sve2'
> >>         'flags': [
> >>             ['RTE_MACHINE', '"neoverse-n2"'],
> >>             ['RTE_ARM_FEATURE_ATOMICS', true],
> >>             ['RTE_MAX_LCORE', 64],
> >>             ['RTE_MAX_NUMA_NODES', 1]
> >>         ]
> >>     }
> >>
> > If I remove here, then this will also change for " Arm Neoverse N2 soc_n2",
> because part_number is same, Right ?
> > Because of this reason, I thought to have separate flag instead of updating
> march_features.
> 
> This new add flag only impact hand-writen sve code, but auto-vectorization is
> also enabled when sve is enabled at march_features.
Agree.

> Maybe NEON-based automated vector code performs better than SVE-
> based.
> 
> I think it's OK to add separate flag in soc_xxx struct, but suggest it also impact
> auto-vectorization.
I would suggest the flag to control only RTE_HAS_SVE_ACLE, i.e. hand written code using SVE C language intrinsics.
For auto-vectorization, I think it is compilers duty to vectorize in the most performant way, use whatever resource hardware provided.

> 
> So for one soc which test or optimize well on sve, it could turn the flag to true.
> 
> >
> >>>  }
> >>>
> >>>  soc_dpaa = {
> >>>
> >