RE: MOSIX and kernel mods.

Andrej Presern (andrejp@luz.fe.uni-lj.si)
Tue, 9 Mar 1999 13:29:12 +0100


On Mon, 08 Mar 1999, Kamran Karimi wrote:
>On Sun, 7 Mar 1999, Andrej Presern wrote:
>
>> > You are confising distinct concepts. There is a difference between
>> >"distributed" and "parallel", even though for you and many other people
>> >the _only_ reason for gooing distributed is to make an application
>> >execute in parallel. This should not always be the case.
>>
>> Would you mind giving a few examples what other benefits of going distributed
>> could be (and that are present when using DSM to distribute an application)?
>
> I leave this to you as an exercise (hint: When you are about to give
>up, look at some of the simple example programs in DIPC's package)

Well, actually, I already did my homework on distributed parallel processing
and I don't think there are any real benefits of going distributed via DSM.
Not so many or of such nature that would outweight the numerous disadvantages
anyway. In this respect I tend to agree with Chris: [quote] "Why code a
parallel application to be slow? If you are diving into parallelism of any
kind, the assumption is that you are looking for some speed. For instance, you
have decided that you have some tasks that can take place at the same time; ie
a calculation that can be distributed. Why put all the effort into
parallelization, which implicitly means looking at classical critical section
problems, deadlocks, etc for a nice big slowdown? "Parallel for free" isn't
good if it runs worse than a non-parallel program. It would also stink to have
someone have to start over because the way the used an abstraction is horrible,
but seemed sensible at the time." [end quote]

I would still like to read your arguments though as I would not like to
discredit any valid technical reasoning. You will have to present some though.
This said, I admit that I have not taken a single look at the documentation,
nor the code itself as I don't find them to be relevant in the discussion - I
don't oppose 'DSM implementation by Kamran Karimi', I question DSM's fitness
as a concept and solution to distributed parallel processing itself. And I do
that based on technical merits that I have learned while researching the topic.
If you would like, I can present what I have concluded, even though it would
probably be appropriate to hear your reasoning first.

So, would you mind giving a few examples of what other benefits (besides
executing in parallel) of going distributed could be (AND that are present when
using DSM to distribute an application)?

Andrej

--
Andrej Presern, andrejp@luz.fe.uni-lj.si

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/