DPDK patches and discussions
 help / color / mirror / Atom feed
* Vector optimized IP checksum
@ 2022-06-16  7:07 Morten Brørup
  2022-06-16 13:46 ` Mattias Rönnblom
  0 siblings, 1 reply; 3+ messages in thread
From: Morten Brørup @ 2022-06-16  7:07 UTC (permalink / raw)
  To: Jasvinder Singh; +Cc: dev, Olivier Matz, Maxime Coquelin, Chenbo Xia

[-- Attachment #1: Type: text/plain, Size: 287 bytes --]

Hi Jasvinder,

 

As the Packet CRC maintainer, have you considered adding vector optimized variants of the rte_raw_cksum() function?

 

It might benefit some of the software PMDs, such as tap and virtio.

 

 

Med venlig hilsen / Kind regards,

-Morten Brørup

 


[-- Attachment #2: Type: text/html, Size: 2266 bytes --]

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Vector optimized IP checksum
  2022-06-16  7:07 Vector optimized IP checksum Morten Brørup
@ 2022-06-16 13:46 ` Mattias Rönnblom
  2022-06-16 14:11   ` Morten Brørup
  0 siblings, 1 reply; 3+ messages in thread
From: Mattias Rönnblom @ 2022-06-16 13:46 UTC (permalink / raw)
  To: Morten Brørup, Jasvinder Singh
  Cc: dev, Olivier Matz, Maxime Coquelin, Chenbo Xia

On 2022-06-16 09:07, Morten Brørup wrote:
> Hi Jasvinder,
> 
> As the Packet CRC maintainer, have you considered adding vector 
> optimized variants of the rte_raw_cksum() function?
> 

The current implementation is vectorized by the compiler.

In my experience, hand-crafting it using SIMD intrinsics doesn't improve 
the performance enough to warrant the trouble.

> It might benefit some of the software PMDs, such as tap and virtio.
> 
> Med venlig hilsen / Kind regards,
> 
> -Morten Brørup
> 

^ permalink raw reply	[flat|nested] 3+ messages in thread

* RE: Vector optimized IP checksum
  2022-06-16 13:46 ` Mattias Rönnblom
@ 2022-06-16 14:11   ` Morten Brørup
  0 siblings, 0 replies; 3+ messages in thread
From: Morten Brørup @ 2022-06-16 14:11 UTC (permalink / raw)
  To: Mattias Rönnblom, Jasvinder Singh
  Cc: dev, Olivier Matz, Maxime Coquelin, Chenbo Xia

> From: Mattias Rönnblom [mailto:hofors@lysator.liu.se]
> Sent: Thursday, 16 June 2022 15.46
> 
> On 2022-06-16 09:07, Morten Brørup wrote:
> > Hi Jasvinder,
> >
> > As the Packet CRC maintainer, have you considered adding vector
> > optimized variants of the rte_raw_cksum() function?
> >
> 
> The current implementation is vectorized by the compiler.
> 
> In my experience, hand-crafting it using SIMD intrinsics doesn't
> improve
> the performance enough to warrant the trouble.
> 

That is great, Mattias!

That definitely puts my question to rest. :-)


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2022-06-16 14:17 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-06-16  7:07 Vector optimized IP checksum Morten Brørup
2022-06-16 13:46 ` Mattias Rönnblom
2022-06-16 14:11   ` Morten Brørup

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).