RE: [PATCH v3 net-next] net: Implement fast csum_partial for x86_64

From	David Laight <David.Laight@ACULAB.COM>
Newsgroups	linux.kernel
Subject	RE: [PATCH v3 net-next] net: Implement fast csum_partial for x86_64
Date	2016-02-05 11:20 +0100
Message-ID	<qYLEC-7ji-7@gated-at.bofh.it> (permalink)
References	<qYoyn-6KS-25@gated-at.bofh.it> <qYoyn-6KS-23@gated-at.bofh.it> <qYzWN-7GJ-1@gated-at.bofh.it> <qYAST-8lI-27@gated-at.bofh.it> <qYJCP-61a-33@gated-at.bofh.it>
Organization	linux.* mail to news gateway

Show all headers | View raw

From: Ingo Molnar
...
> As Linus noticed, data lookup tables are the intelligent solution: if you manage
> to offload the logic into arithmetics and not affect the control flow then that's
> a big win. The inherent branching will be hidden by executing on massively
> parallel arithmetics units which effectively execute everything fed to them in a
> single cycle.

Not necessarily, in real-life they are likely to be a cache miss.

With the parallel execution units you just need to worry about the
register dependency chains.
In this case the 'adc' have dependencies on both the result register
and the flags register.

The best loop might even be:
10:	adc	%rax,0(%rdi,%rcx)
	lea	%rcx,8(%rcx)
	jcxz	20f
	jmp	10b
20:

You then need to insert just enough copies of the adc so they take as long
as the other instructions.

	David

Back to linux.kernel | Previous | Next — Previous in thread | Find similar | Unroll thread

Thread

Re: [PATCH v3 net-next] net: Implement fast csum_partial for x86_64 Ingo Molnar <mingo@kernel.org> - 2016-02-04 10:40 +0100
  Re: [PATCH v3 net-next] net: Implement fast csum_partial for x86_64 Ingo Molnar <mingo@kernel.org> - 2016-02-04 12:00 +0100
    Re: [PATCH v3 net-next] net: Implement fast csum_partial for x86_64 Tom Herbert <tom@herbertland.com> - 2016-02-04 20:30 +0100
      Re: [PATCH v3 net-next] net: Implement fast csum_partial for x86_64 Ingo Molnar <mingo@kernel.org> - 2016-02-05 10:30 +0100
  Re: [PATCH v3 net-next] net: Implement fast csum_partial for x86_64 Linus Torvalds <torvalds@linux-foundation.org> - 2016-02-04 22:50 +0100
    Re: [PATCH v3 net-next] net: Implement fast csum_partial for x86_64 Linus Torvalds <torvalds@linux-foundation.org> - 2016-02-04 23:10 +0100
      Re: [PATCH v3 net-next] net: Implement fast csum_partial for x86_64 Linus Torvalds <torvalds@linux-foundation.org> - 2016-02-05 02:30 +0100
        Re: [PATCH v3 net-next] net: Implement fast csum_partial for x86_64 Linus Torvalds <torvalds@linux-foundation.org> - 2016-02-05 02:40 +0100
    Re: [PATCH v3 net-next] net: Implement fast csum_partial for x86_64 Tom Herbert <tom@herbertland.com> - 2016-02-04 23:50 +0100
      Re: [PATCH v3 net-next] net: Implement fast csum_partial for x86_64 Linus Torvalds <torvalds@linux-foundation.org> - 2016-02-05 00:00 +0100
      Re: [PATCH v3 net-next] net: Implement fast csum_partial for x86_64 Ingo Molnar <mingo@kernel.org> - 2016-02-05 09:10 +0100
        RE: [PATCH v3 net-next] net: Implement fast csum_partial for x86_64 David Laight <David.Laight@ACULAB.COM> - 2016-02-05 11:20 +0100

csiph-web