• efc_J sparse efc_J #934
• ten_J
• actuator_moment sparse actuator_moment #946
• flexedge_J sparse flexedge_J #931

This is a pretty hard feature and may need to be broken down into subtasks.

We may need to study how to do efficient batched sparse matmul in Warp.

We might want to start off with a feasibility study to make sure Sparse is fast enough on GPU to be worth the extra complexity of supporting it.