Fix logic mismatch in newR1C for better Groth16 optimization

[tomasandroil]

Previously, the comment suggested that the goal was to minimize the number of variables in the B matrix, but the implementation did the opposite — swapping L and R when L was longer. This PR corrects the condition to match the intended optimization logic, aligning with Groth16 performance best practices.

Fixes # (no issue number yet)

---

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

---

How has this been tested?

• Manual inspection of Groth16 output matrices
• Unit test (TBD - logic is straightforward but may benefit from regression test)

[gbotrel]

@ThomasPiellard can you check that one? I recall empirically, the code as is (before this PR) actually produce better perf