This is fairly easy to do with @parallel:
def fast(x):
return x
def slow(x):
sleep(x)
return x
def slower(x):
sleep(x*x)
return x
algorithms = [slower, slow, fast]
@parallel(len(algorithms))
def fastest(i,x):
global algorithms
return algorithms[i](x)
def compute(x):
return fastest([(i,x) for i in range(len(algorithms))]).next()[1]
On Thu, Nov 3, 2011 at 11:37 AM, mmarco <mma...@unizar.es> wrote:
> There are some computations for which there are several algorithms
> available. Think, for example, of GTZ and SY algorithms for primary
> decomposition. There is no useful criterion to determine which one
> will be faster for a given case in advance. With modern computers
> having multiple cores, i think it could be a good idea to run both of
> them in parallel, until one of them finishes, and then stop the other.
>
> I still haven't thought how, but it seems reasonable that this aproach
> would be doable in Sage. So, do you think we should follow this
> aproach in sage? Namely, include an option algorithm='faster' (or
> something like that) for functions that admit several algorithms.
>
> --
> To post to this group, send an email to sage-devel@googlegroups.com
> To unsubscribe from this group, send an email to
> sage-devel+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/sage-devel
> URL: http://www.sagemath.org
>
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
