Feature #1417
RadicalZeroDim with extra parameter for GBasis timeout
Description
We have seen how changing the timeout for GBasis can affect the time dramatically.
Should we make a function RadicalZeroDim(I, GBTimeout)
so that the user can choose?
Related issues
History
#1 Updated by Anna Maria Bigatti about 4 years ago
- Related to Slug #1114: Some other examples for 0-dim radical added
#2 Updated by Anna Maria Bigatti about 4 years ago
- Related to Slug #1375: Radical 0-dim: varied timings added
#3 Updated by Anna Maria Bigatti about 4 years ago
- Related to Design #1378: Create two separate radical fns (for 0-dim ideals) added
#4 Updated by Anna Maria Bigatti about 4 years ago
This is related but not quite the same as #1378.
The point is this: if we let more time to compute the partial GBases, we'll also have a better chance to compute the GBasis of the radical ideal.
If we just use "Seidenberg" (with a very fast radical computation) then its GBasis might be much slower!
Compare, think, decide.
#5 Updated by John Abbott about 3 years ago
What is the status of this issue?
As I recall radical
now works fairly well for 0-dim ideals (see issue #948). But see also issue #1375.
A problem with the "timeout" parameter is that it could be hard to describe what exactly the parameter does/means.
Also what would the "timeout" parameter mean if the given ideal is not 0-dim?
#6 Updated by John Abbott about 3 years ago
- % Done changed from 10 to 20
If no timeout is specified (i.e. the current situation) then some "heuristic" timeout is used.
Currently the heuristic seems to be 30s plus epsilon -- I did not investigate to see where epsilon comes from.
It might be nice to use a better heuristic? Or just KISS?
#7 Updated by John Abbott over 2 years ago
- Target version changed from CoCoALib-0.99800 to CoCoALib-0.99850
#8 Updated by Anna Maria Bigatti 3 months ago
- Related to Feature #1780: radical for ideals in SparsePolyRing: code in C++ added
#9 Updated by John Abbott about 1 month ago
- Target version changed from CoCoALib-0.99850 to CoCoALib-0.99880
I am currently unsure how useful it would be to let the caller choose a time-out in seconds.
It may be simpler to offer just vaguer options e.g. "low, medium, high" (or even just "low, high"). What do you think?