Open MPI logo

Hardware Locality Users' Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Hardware Locality Users mailing list

Subject: Re: [hwloc-users] [EXTERNAL] Re: hwloc_get_latency() failures and confusion
From: Brice Goglin (Brice.Goglin_at_[hidden])
Date: 2012-08-07 02:17:11


Le 07/08/2012 01:45, Kenneth A. Lloyd a écrit :
> But, isn't the returned matrix merely and adjacency matrix? Distance in
> steps? This doesn't encode a weight, such as latency does it?

There could be different types of matrices. Right now, there's a single
type called "latency" as defined by the ACPI SLIT specs (10 on the
diagonal, >10 everywhere else, but we normalize things to get 1.0 on the
diagonal).

Brice

>
> Ken
>
> -----Original Message-----
> From: hwloc-users-bounces_at_[hidden]
> [mailto:hwloc-users-bounces_at_[hidden]] On Behalf Of Brice Goglin
> Sent: Monday, August 06, 2012 4:47 PM
> To: Wheeler, Kyle Bruce
> Cc: <hwloc-users_at_[hidden]>
> Subject: Re: [hwloc-users] [EXTERNAL] Re: hwloc_get_latency() failures and
> confusion
>
> Le 07/08/2012 00:36, Wheeler, Kyle Bruce a écrit :
>> Ahhhh, that's key! The documentation currently says "Look at ancestor
>> objects from the bottom to the top until one of them contains a
>> distance matrix that matches the objects exactly", which suggests to
>> me that it will traverse the object hierarchy looking for the
>> NUMANodes *for* me.
> Ahh, this one is exactly what's really confusing. There are two things here:
> 1) the object that contains the distance matrix
> 2) the object that are covered by the matrix
>
> When the matrix covers the entire machine (usual case), (1) is the root
> object and (2) are NUMA nodes.
> If you ever have a distance matrix between all cores of the first socket
> (and not any other core in the machine), the first socket object would
> contain a matrix with distance->relative_depth = depth(socket)-depth(core)
>
> So when you're looking for Core latencies, you check whether the ancestor
> immediately above Core contain a matrix for Core distances, then its parent,
> ... up to the root object. So it's about (1) moving up but (2) remains the
> same type (but (2) gets wider when (1) goes up).
>
> Brice
>
> _______________________________________________
> hwloc-users mailing list
> hwloc-users_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/hwloc-users
>
>
> _______________________________________________
> hwloc-users mailing list
> hwloc-users_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/hwloc-users