Archives


- Beowulf
- Beowulf Announce
- Scyld-users
- Beowulf on Debian

[Beowulf] MS HPC... Oh dear...

Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.

Search

Ashley Pittman ashley at quadrics.com
Mon Jun 12 08:46:57 PDT 2006


On Mon, 2006-06-12 at 11:18 -0400, Joe Landman wrote:
> 
> Ashley Pittman wrote:
> 
> >> More to the point, this dynamic binding allows you to write to the API,
> >> present a consistent ABI, and handle the hardware details elsewhere in a
> >> driver which can be linked in by the .so/.dll/.eieio method at runtime.
> > 
> > That's exactly what I thought.  I was thrown by your "This is a nice
> > idea" comment because using Windows is no different to Linux in this
> > respect.
> 
> The major difference will be that Microsoft enforces its worldview here,
> which in this particular point, will likely be more in line with what we
> should see on Linux, but don't for a variety of reasons (see Greg L's
> post on MPI ABI).  We need it, we just haven't collectively decided that
> there is enough pain to make this the general case yet.  Microsoft
> decided (correctly) that this was a good thing, and will just do it.
> This will have a number of side effects, including making "porting" to
> new hardware communication adapters MPI stacks a thing of the past, but
> only on windows.  Not on Linux.

I can almost see your argument here, currently there isn't a MPI ABI and
you think that on Windows this situation might be different?  The
technical problems of making a stable, cross ISV ABI are the same on
Linux and Windows (possibly even slightly worse under Windows), the only
difference being that under Linux there is a large amount of inertia in
keeping the status quo as moving to a consistent model would break at
least somebodies applications, under a brave new Windows world that
inertia isn't there so the ISV might as well mimic the ABI of somebody
(in this case Microsoft) from day one.

In practise I don't think this is going to be a step forward for two
reasons, firstly the devil is in the detail so having two different code
bases export the same ABI might work at the compile/link level but they
would still behave subtlety differently at run-time, secondly windows is
well known for it's DLL hell model, they can't even have maintain
consistency between release so tracking it externally would be a very
difficult (and error prone) thing to do.

Ashley,




More information about the Beowulf mailing list