end2end benches run with duplicate Store types

[samuelburnham]

Since #774 we no longer run benches with blstrs::Scalar. In end2end.rs and end2end_lem.rs we still run comparative benchmarks with a bls_store and pallas_store, but they are the same type.

Either we should restore the blstrs::Scalar as Fr import or dispense with the bls_store benchmark (I think the latter?).