| [ATL] transformation efficiency? [message #86325] |
Thu, 10 July 2008 15:23  |
Eclipse User |
|
|
|
Hi,
I have a question, maybe is a little off-topic, but I thought it's worth a
try.
Are there papers/documents about how to evaluate the worst-case execution
time of an ATL transformation? My main concern is with the processing of
operation helpers.
Thanks,
- Leila
|
|
|
| Re: [ATL] transformation efficiency? [message #89320 is a reply to message #86325] |
Fri, 29 August 2008 04:40  |
| Eclipse User |
|
|
|
Leila wrote:
> Hi,
> I have a question, maybe is a little off-topic, but I thought it's worth
> a try.
>
> Are there papers/documents about how to evaluate the worst-case
> execution time of an ATL transformation? My main concern is with the
> processing of operation helpers.
>
>
> Thanks,
>
> - Leila
Hello Leila,
There are no papers on this as far as I know. Here's what I think, though:
There is no limit to the execution time of a (non-trivial) ATL
transformation. You can provide any size of input model, which requires
a proportional execution time. The execution time increases with input
model size according to the complexity order of your ATL "algorithm"
that is being executed over the input model.
Termination should at least be guaranteed by having finite input models
and no way to express infinite loops (infinite recursion will eventually
result in a stack overflow).
Regards,
--
Dennis
|
|
|
Powered by
FUDForum. Page generated in 0.23760 seconds
Home