Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.arch > #6314

Re: Ne(ish) IBM z196 synchronization instructions

Show all headers | View raw

On Thursday, March 8, 2012 11:39:13 AM UTC-6, Terje Mathisen wrote:
> Andy (Super) Glew wrote:
> > On 3/7/2012 1:43 PM, MitchAlsup wrote:
> >> On Wednesday, March 7, 2012 1:00:20 PM UTC-6, Terje Mathisen wrote:
> >>> I have some problem figuring out the intended usage for the AND/OR/XOR
> >>> versions of this operation:
> >>
> >> XOR is useful for setting bits and resetting bits (from the same
> >> constant).
> >>
> >> So one can "check in their request" and "checkout their request" using
> >> the same value. Otherwise one has to ADD a bit to check in a request
> >> and then ADD in the 2's complement to check it back out (or OR on a
> >> bit and AND-NOT it back off).
> >>
> >> Most of the statemachine logic in my simulators uses XOR as in:
> >>
> >> if( p->state& STATE1 )
> >> p->state ^= STATE1 | STATE2;
> >> else
> >>
> >> // where it is known STATE1 and STATE2 have no common bits
> >> // but may have more than one bit set.
> >>
> >> Mitch
> >
> >
> > Right. But do you need that to be atomic?
> >
> > Can you have multiple threads using the same p->state at the same time?
> > Setting the same bits?
> >
> > Or possibly setting disjoint bits in the same word?
> >
> > To get XOR the safely set and clear different bits, you have to know the
> > prior state. Perhaps and AND-OR instruction?
> >
> > AND-OR
> > tmp := mem
> > tmp &= and_mask (or ~and_mask)
> > tmp |= or_mask
> > mem := tmp
> >
> > A nice instruction, but it really wants two immediates.
> >
> 
> Exactly, it is hard to figure out a nice usage scenario for the existing 
> logical versions.

I have seen a parallel processing implementation that used the bit to denote which processor was requesting/waiting-for the lock (kernel only). As long as the number of processors is smaller than wordsize this works well.

Mitch

Back to comp.arch | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

Ne(ish) IBM z196 synchronization instructions "Andy (Super) Glew" <andy@SPAM.comp-arch.net> - 2012-03-06 20:20 -0800
  Re: Ne(ish) IBM z196 synchronization instructions "Paul A. Clayton" <paaronclayton@gmail.com> - 2012-03-07 05:43 -0800
    Re: Ne(ish) IBM z196 synchronization instructions "Andy (Super) Glew" <andy@SPAM.comp-arch.net> - 2012-03-07 07:00 -0800
      Re: Ne(ish) IBM z196 synchronization instructions nmm1@cam.ac.uk - 2012-03-07 15:25 +0000
      Re: Ne(ish) IBM z196 synchronization instructions "Paul A. Clayton" <paaronclayton@gmail.com> - 2012-03-09 11:05 -0800
  Re: Ne(ish) IBM z196 synchronization instructions Terje Mathisen <"terje.mathisen at tmsw.no"> - 2012-03-07 20:00 +0100
    Re: Ne(ish) IBM z196 synchronization instructions MitchAlsup <MitchAlsup@aol.com> - 2012-03-07 13:43 -0800
      Re: Ne(ish) IBM z196 synchronization instructions "Andy (Super) Glew" <andy@SPAM.comp-arch.net> - 2012-03-08 09:08 -0800
        Re: Ne(ish) IBM z196 synchronization instructions Terje Mathisen <"terje.mathisen at tmsw.no"> - 2012-03-08 18:39 +0100
          Re: Ne(ish) IBM z196 synchronization instructions Nomen Nescio <nobody@dizum.com> - 2012-03-08 22:33 +0100
          Re: Ne(ish) IBM z196 synchronization instructions "Andy (Super) Glew" <andy@SPAM.comp-arch.net> - 2012-03-08 17:17 -0800
          Re: Ne(ish) IBM z196 synchronization instructions MitchAlsup <MitchAlsup@aol.com> - 2012-03-09 08:19 -0800
            Re: Ne(ish) IBM z196 synchronization instructions Terje Mathisen <"terje.mathisen at tmsw.no"> - 2012-03-09 17:45 +0100
    Re: Ne(ish) IBM z196 synchronization instructions jgk@panix.com (Joe keane) - 2012-03-09 20:30 +0000
      Re: Ne(ish) IBM z196 synchronization instructions MitchAlsup <MitchAlsup@aol.com> - 2012-03-09 13:10 -0800

csiph-web