On Saturday, December 6, 2014 11:04:31 AM UTC-8, Ralf Hemmecke wrote:
>
> I'd say yes. But it's probably Waldek who has more knowledge of ecl vs. 
> sbcl. I only remember that compilation (at least some years ago) with 
> ecl took quite a bit longer than with sbcl. 
>

Yes, ecl tends to be quite slow in compiling relative to other lisps, 
partially because gcc is invoked in the process. That's not necessarily 
indicative of runtime performance. SBCL tends to be faster in general, 
though.

In fact, I don't see any "embeddable" advantage of ecl for Sage. 
> It's usually the computation that costs time (since I only expect 
> heavier or complicated computations exported to FriCAS) not the 
> conversion from one Sage to FriCAS and back. 
>

Using a binary transfer interface rather than a strings-based one doesn't 
just have performance advantages, it also solves reliability issues. You 
wouldn't need to run in the same process, though. With some programming 
work on both sides, it should be possible to set up a communication 
interface that does not rely on "typed mathematics".
 

> Especially the translation of FriCAS types into Sage types is nearly 
> impossible. FriCAS has quite a sophisticated type hierarchy. I would 
> expect that there isn't even always a corresponding Sage type that one 
> could convert the FriCAS result to in order to have those values 
> natively available for further computations in Sage. 
>

Types that can't be represented in sage would obviously not be very 
interesting for sage (but this can change in time). The ones that can would 
be very useful, though. Plus, we can just extend what is available in sage. 
Note that FriCAS uses "types" for purposes that in sage are modeled using 
its "parent/category/type" framework, so for translation you shouldn't try 
and stuff it all into just the python type hierarchy that sage uses.


-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Reply via email to