[DOC] Developer documentation for InferBound pass. (apache#3126)

* Developer documentation for InferBound pass.

docs/dev/codebase_walkthrough.rst

@@ -136,7 +136,10 @@ Lowering is done by ``tvm.lower()`` function, defined in ``python/tvm/build_modu
       stmt = schedule.ScheduleOps(sch, bounds)
       ...
 
-Bound inference is the process where all loop bounds and sizes of intermediate buffers are inferred. If you target the CUDA backend and you use shared memory, its required minimum size is automatically determined here. Bound inference is implemented in ``src/schedule/bound.cc``, ``src/schedule/graph.cc`` and ``src/schedule/message_passing.cc``.
+Bound inference is the process where all loop bounds and sizes of intermediate buffers are inferred. If you target the CUDA backend and you use shared memory, its required minimum size is automatically determined here. Bound inference is implemented in ``src/schedule/bound.cc``, ``src/schedule/graph.cc`` and ``src/schedule/message_passing.cc``. For more information on how bound inference works, see `InferBound Pass`_.
+
+.. _InferBound Pass: http://docs.tvm.ai/dev/inferbound.html
+
 
 ``stmt``, which is the output of ``ScheduleOps()``, represents an initial loop nest structure. If you have applied ``reorder`` or ``split`` primitives to your schedule, then the initial loop nest already reflects that changes. ``ScheduleOps()`` is defined in ``src/schedule/schedule_ops.cc``.
 

docs/dev/index.rst

@@ -33,3 +33,4 @@ In this part of documentation, we share the rationale for the specific choices m
    relay_add_op
    relay_add_pass
    codebase_walkthrough
+   inferbound