options to help lift or liftstd complete

I have a system of polynomials that slimgb can calculate the groebner basis in about 10 min and total memory usage under 2GB.

I'd like to write some of the basis elements in terms of the original ideal polynomials.
I tried to use the version of liftstd that lets me specify slimgb, but in about an hour and a half it uses 256GB and dies from out of memory.

I tried using lift instead, and that also runs out of memory.

Since calculating the basis appears easy for Singular, I really hoped I could get this information.
Are there any options you recommend I set?
Or possibly some other way to get this information which is easier once I have the basis?

I am using this version:
Singular for x86_64-Linux version 4.2.0 (4200, 64 bit) Dec 24 2020 14:40:58

- check that the attribute "isSB" is set for the Groebner basis
- try lift, and apply it to one polynomial at a time

The original set of generators for the ideal is not a basis. Let's call this ideal idOrg.
I can use Singular to calculate a groebner basis: "ideal idBasis = slimgb(idOrg);"
I would like to be able to write the polynomials in idBasis as a linear combination of the polynomials in idOrg.

So if I select a single polynomial from idBasis, and make a new ideal idSingle with just that polynomial, I can call:
matrix T = lift(idOrg, idSingle)
Neither idOrg nor idSingle are a basis, and the ideal which is a basis (idBasis) is not an argument.
Similarly if I call liftstd to calculate the full matrix in one go
matrix T;
ideal idBasis2 = liftstd(idOrg, T);
the basis is not an argument to liftstd.
So I'm not sure what you mean by: "check that the attribute "isSB" is set for the Groebner basis".
Are you saying there is some way I can pass an already calculated basis to lift or liftstd to help it out?
I don't see that in the documentation. How do I do that?