By: none (none.delete@this.none.com), November 14, 2012 3:40 am
Room: Moderated Discussions
anonymou5 (no.delete@this.spam.com) on November 13, 2012 5:44 pm wrote:
> > I seem to recall that a scatter instruction was included in LRBni (and
> > this article that I found from Michael Abrash makes mention of one); is
> > there any particular reason why it could not be included in AVX2 or
> > could not foreseeably be added in future iterations?
>
> KNF/L1OM has VSCATTERD.
> KNC/K1OM has VSCATTER{PS,PD} and VPSCATTERD{D,Q}.
Knight Ferry and Knight Corner I guess?
In Xeon Phi manual I see this:
- VPSCATTERD{D,Q} -> Scatter Int{32,64} Vector With Signed Dword Indices
- VSCATTERDP{D,S} -> Scatter Float{64,32} Vector With Signed Dword Indices
- and various Float{32,64} with prefetch and hint
> AVX2 doesn't have it.
>
> AVX3 might.
> > I seem to recall that a scatter instruction was included in LRBni (and
> > this article that I found from Michael Abrash makes mention of one); is
> > there any particular reason why it could not be included in AVX2 or
> > could not foreseeably be added in future iterations?
>
> KNF/L1OM has VSCATTERD.
> KNC/K1OM has VSCATTER{PS,PD} and VPSCATTERD{D,Q}.
Knight Ferry and Knight Corner I guess?
In Xeon Phi manual I see this:
- VPSCATTERD{D,Q} -> Scatter Int{32,64} Vector With Signed Dword Indices
- VSCATTERDP{D,S} -> Scatter Float{64,32} Vector With Signed Dword Indices
- and various Float{32,64} with prefetch and hint
> AVX2 doesn't have it.
>
> AVX3 might.
