Version 1 of device CBC sweep chunk kernel

[eappen-nelluvelil]

In this PR, version 1 of the device CBC sweep chunk kernel has been implemented.

The following are the major changes:

• There is a new scheduling algorithm for device CBC sweeps called ASYNC_FIFO. This algorithm sweeps asynchronously over all anglesets in a given groupset via crb::Stream's. The main while-loop loops over all anglesets and checks for cells that can be swept. The ready cells for a given angleset are batched and given to the device CBC sweep chunk kernel, and is executed asynchronously. This is parallel over a given set of ready cells and for all groups in a given groupset. The CBC_AsynchronousCommunicator's still receives and sends upwind and downwind angular fluxes, respectively, as soon as possible, which mirrors the host CBC sweep chunk kernel.
• The CBCD_FLUDS class contains host and device buffers for boundary, local, and non-local angular flux data. Local cell angular flux data is kept on the device. CBCD_FLUDS provides methods for asynchronously copying incoming/outgoing boundary and non-local cell angular flux data between host and device.
• The CBCD_FLUDSCommonData class contains similar functionality as in the AAHD_FLUDSCommonData class for encoding/decoding the appropriate locations that cell face nodes need to be read from and written to in the local, boundary, and non-local host/device buffers.
• The device CBC sweep chunk kernel is templated on the number of cell spatial DOFs. The functionality of the device kernel is structured in a similar way as the device AAH sweep chunk kernel.

The device CBC sweep chunk kernel passes all of the GPU regression tests, save for the transport_3d_4_cycles_1_gpu.py. This is because the CBC sweep algorithm does not handle cyclic dependencies.

[eappen-nelluvelil]

@wdarylhawkins @quocdang1998 Requesting review for this PR.

[andrsd]

First rough pass through the PR.

[eappen-nelluvelil]

@quocdang1998 Made the following changes

• AAHD_NodeIndex and CBCD_NodeIndex classes inherit from DeviceNodeIndex class in device_structs.h
• AAHD_FLUDSPointerSet and CBCD_FLUDSPointerSet classes inherit from Device_FLUDSPointerSet in device_structs.h
• GPU solver kernels are templated on DeviceNodeIndex and Device_FLUDSPointerSet classes; CBC-specific functionality now contained in the cbc_gpu_kernel namespace
• Simplified CBCD_FLUDS class methods, and incoming/outgoing boundary and non-local node lookup maps are now built in CBCD_FLUDSCommonData

The device CBC sweep chunk kernel passes all GPU regression tests (confirmed by setting sweep_type="CBC" in these tests).

The code is still failing one clang-tidy check:

/opt/local/opensn/clang/21.1.0/dependencies/include/mpicpp-lite/impl/Request.h:37:58: error: Assigned value is uninitialized [clang-analyzer-core.uninitialized.Assign,-warnings-as-errors]
   37 | inline Request::Request(const MPI_Request & r) : request(r) {}

Using // NOLINT doesn't suppress this clang-tidy error.

Requesting re-review.

[eappen-nelluvelil]

@andrsd I haven't been able to resolve this clang-tidy error that pops up in cbc_async_comm.cc:

/opt/local/opensn/clang/21.1.0/dependencies/include/mpicpp-lite/impl/Request.h:37:58: error: Assigned value is uninitialized [clang-analyzer-core.uninitialized.Assign,-warnings-as-errors]
   37 | inline Request::Request(const MPI_Request & r) : request(r) {}
      |                                                          ^
/home/dhawkins/actions-runner/_work/opensn/opensn/modules/linear_boltzmann_solvers/discrete_ordinates_problem/sweep/communicators/cbc_async_comm.cc:38:7: note: Assuming the condition is false
   38 |   if (not outgoing_message_queue_.empty())

The only change I made to the CBC_AsynchronousCommunicator class is fixing the incorrect capitalization of the "s" in the name (CBC_ASynchronousCommunicator to CBC_AsynchronousCommunicator).

Do you have any ideas as to why clang-tidy is throwing this error?