[mercury-users] Mercury and NLP on smartphones

[Chris King]

On Fri, Apr 8, 2011 at 2:41 PM, Alexsandro Soares
<prof.asoares at gmail.com> wrote:
> 1 - Is it possible to generate compact code to fit into available
> memory on smartphones using different platforms such as Android,
> iPhone, Windows Mobile and others?

I don't think the Apple Store's TOS will let you publish an app
written in Mercury (ridiculous I know).  If you're just doing research
it might not be a problem.

> 3- Is there much difference in efficiency and code size between the
> different back-ends available for Mercury?

I've been working with various benchmarks lately and I can say that
between the low-level C (asm_fast), high-level C (hlc), and Java
backends, which backend is fastest depends on the application.  In
particular I've had surprisingly good results with floating-point
operations in the Java backend, and programs which make many small
allocations seem to do well in hlc.

If you used the Java backend to write code targeting the Android, I
wouldn't be surprised to see decent performance, especially if you use
a custom (minimal) standard library.

- Chris
--------------------------------------------------------------------------
mercury-users mailing list
Post messages to:       mercury-users at csse.unimelb.edu.au
Administrative Queries: owner-mercury-users at csse.unimelb.edu.au
Subscriptions:          mercury-users-request at csse.unimelb.edu.au
--------------------------------------------------------------------------