[llvm] [NVPTX] Add idp2a, idp4a intrinsics (PR #102763)

[Artem Belevich via llvm-commits]

================
@@ -287,6 +287,62 @@ The ``@llvm.nvvm.fence.proxy.tensormap_generic.*`` is a uni-directional fence us
 
 The address operand ``addr`` and the operand ``size`` together specify the memory range ``[addr, addr+size)`` on which the ordering guarantees on the memory accesses across the proxies is to be provided. The only supported value for the ``size`` operand is ``128`` and must be an immediate. Generic Addressing is used unconditionally, and the address specified by the operand addr must fall within the ``.global`` state space. Otherwise, the behavior is undefined. For more information, see `PTX ISA <https://docs.nvidia.com/cuda/parallel-thread-execution/#parallel-synchronization-and-communication-instructions-membar>`_.
 
+Arithmetic Intrinsics
+---------------------
+
+'``llvm.nvvm.idp2a``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+.. code-block:: llvm
+
+    declare i32 @llvm.nvvm.idp2a(i32 %a, i1 immarg %a.unsigned, i32 %b, i1 immarg %b.unsigned, i1 immarg %is.hi, i32 %c)
+
+Overview:
+"""""""""
+
+The '``llvm.nvvm.idp2a``' intrinsic performs a 2-element vector dot product
+followed by addition. It corresponds directly to the ``dp2a`` PTX instruction.
+
+Semantics:
+""""""""""
+
+The 32-bit value in ``%a`` is broken into 2 16-bit values which are either sign
+or zero extended, depending on the value of ``%a.unsigned``, to 32 bits. Two
+bytes are selected from ``%b``, if ``%is.hi`` is true, the most significant
+bytes are selected, otherwise the least significant bytes are selected. These
+bytes are each sign or zero extended, depending on ``%b.unsigned``. The dot
+product of these 2-element vectors is added to ``%c`` to produce the return.
----------------
Artem-B wrote:

> for N extra immediate-only arguments, we will instead need N^2 intrinsic names

Yes, though `N^2` scaling applies only if the each argument is boolean. In general case, each argument may have more than 2 values, and we're looking at a full cartesian product of all such arguments. That obviously would not scale if we have too many combinations, but in practice it works well enough in most cases. We'll discuss/address exceptions on a case-by-case basis.

https://github.com/llvm/llvm-project/pull/102763