Same as Figure 4 but with integrated measurements. The full/dense GP solution (green diamonds) uses the L26 method, which we implemented in tinygp. The SSM solution (blue circles) uses our augmented SSM approach (Section 3), as implemented in smolgp. Likewise, the parallel version (purple circles) implements Section 3.2.4 in smolgp. Takeaway: Because there is no quasi-separable framework compatible with integrated measurements, the SSM method provides a transformative speedup and memory savings over traditional GP methods which are forced to construct the full covariance matrix.