Re: Streaming, batching, concurrent and alternating

Hi John,

> Yes, I'm saying I disagree that non-deterministic
> streaming can yield higher performance than deterministic streaming
> in any scenarios
> [...]
> Yes, I'm saying I disagree that
> non-deterministic streaming is required to get *adequate* performance.
> I don't think it is required at all in fact and any SCE-MI 2.0
> proposal should not allow it.

Ok, thanks for clarifying.  BTW, I agree with the determinism
requirement especially if we are constraining it to the hardware
side.  And I don't think that either proposal violates this
either.  I'd still like to hear what Matt and Shabtay have to say re
determinism of the SceMiVarMessageInPort (the question you raised),
though.

> The Mentor pipes support both data shaping and eom.

Yes, I got that.  What I was trying to point out was that Shabtay's
batching proposal is equivalent to allowing multiple eom's per
pipe read.  Correct me if I'm wrong, but I don't think that is
supported in your proposal?  I'm just trying to understand the
difference between data shaping and message batching.

> You can use eom without data shaping, data shaping without eom,
> or both together. We tried to design this flexibly, and based
> on uses of it so far it seems to work in a fairly simple manner
> to serve a variety of different needs.

The optionality of eom is seems to be a distinguishing feature from
the Cadence VLM ports.  As far as I can tell, the equivalent of eom
in the VLM ports is not optional.  Of course, the application could
choose simply to ignore it.

> We're open to ideas for even further improvement. As I've said,
> this is a fairly compact facility that can be built over a
> base DPI capability.

Understood.  We need to decide on a base proposal and then work out
the details from there.

Does it seem fair to say that what distinguishes the two proposals
is not in the areas of determinism and streaming/batching.  Both
seem to have similar features in those two areas.  What really
distinguishes them is the function-based versus macro-based 
approach to the interface.

Per