On Wednesday, July 8, 2015 at 8:31:17 AM UTC-4, Don Y wrote:

Hi,

[on the off chance anyone actually reads this NG...]

I'm looking for "device-independent" means (units?)
by which to specify computational requirements of
tasks (operations per unit time?) to serve as
(rough) criteria for scheduling and schedulability.

Pointers to any work on this subject?

Thx,
--don

Hi Don,

Interesting question. I'd like to see the answer too.

ed

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 7/8/15 8:31 AM, Don Y wrote:

Hi,

[on the off chance anyone actually reads this NG...]

I'm looking for "device-independent" means (units?)
by which to specify computational requirements of
tasks (operations per unit time?) to serve as
(rough) criteria for scheduling and schedulability.

Pointers to any work on this subject?

Thx,
--don

I'm not sure such a thing even makes sense. The most fundamental thing
you need to know for scheduling is VERY device-DEPENDANT, is how much of
the device resources the task will need.

Knowing that Task A uses 5 Foos and Task B uses 3 Foos gives you no
information unless you know how many Foos you have. If you only have 6
Foos, you have a problem, if you have 100 you have no worries.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Wednesday, July 8, 2015 at 8:31:17 AM UTC-4, Don Y wrote:

I'm looking for "device-independent" means (units?)
by which to specify computational requirements of
tasks (operations per unit time?) to serve as
(rough) criteria for scheduling and schedulability.

Pointers to any work on this subject?

In article <829c8269-05f7-48b0-83dd-1eea3671ab8b@googlegroups.com>,
Ed Prochak <edprochak@gmail.com> wrote:

Interesting question. I'd like to see the answer too.
ed

I overlooked the question when it was posted, but noticed the recent
reactions.

The OP and others might be interested in the features that the Ada
programming language provides to specify and work with "budgets"
for tasks (although not really "device-independent", as related
to execution time, but more a pragmatic approach). Anyway, some
references and pointers follow.

In the 2005 Ada language standard, CPU clocks and timers were
introduced; an overview from the "Rationale for Ada 2005": <http://www.adaic.org/resources/add_content/standards/05rat/html/Rat-5-6.html>

In the latest (2012) Ada language standard, this was further extended;
see "Interrupt timers and budgets" in the "Rationale for Ada 2012": <http://www.ada-auth.org/standards/12rat/html/Rat12-5-4.html>

The relevant section in the Ada Reference Manual is "D.14 Execution
Time": <http://www.adaic.org/resources/add_content/standards/12rm/html/RM-D-14.html>

In the book "Concurrent and Real-Time Programming in Ada" by Alan
Burns and Andy Wellings, University of York, 2007, Chapter 15
entitled "Timing events and execution time control" might provide
some interesting reading. See: <http://www.cambridge.org/be/academic/subjects/computer-science/programming-languages-and-applied-logic/concurrent-and-real-time-programming-ada?format=HB&isbn=9780521866972>

Work in the Ada community on related issues is ongoing. See for
example the paper "An Execution Model for Fine-Grained Parallelism in
Ada" by Luis Miguel Pinho et al, presented at the latest Ada-Europe
conference June 2015 in Madrid. See pages 196-211 in the proceedings: <http://link.springer.com/book/10.1007%2F978-3-319-19584-1> <http://link.springer.com/chapter/10.1007/978-3-319-19584-1_13>

The relevant newsgroup for further questions on Ada related aspects
of this issue is comp.lang.ada.

I hope this helps,

Dirk
--
Dirk.Craeynest@cs.kuleuven.be (for Ada-Belgium/Ada-Europe/SIGAda/WG9)

*** 21st Intl.Conf.on Reliable Software Technologies - Ada-Europe'2016
*** June 13-17, 2016 *** Pisa, Italy *** http://www.ada-europe.org ***

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)