> 1.    Persistent memory over fabrics – Chet Douglas to discuss NVML
> (and the applications that consume it) as a potential source for API
> requirements
> 2.    Update on Rel 1.5
> 3.    Others?

An issue was raised regarding the semantics of FI_FENCE.  It seems that the 
current definition is not usable by apps, and that a strong semantic is needed. 
 I'd like to touch on this as well, particularly since the FENCE command may be 
applicable to persistent memory.

- Sean 
_______________________________________________
ofiwg mailing list
[email protected]
http://lists.openfabrics.org/mailman/listinfo/ofiwg

Reply via email to