feat: add driver API foundation and NVIDIA implementation

[fuyou4546]

Summary

• Add Driver abstraction DriverBase, DeviceDriver, CudaDriver in src/driver.h and src/native/cuda/driver_.h
• Implement NVIDIA CUDA driver specialization in src/native/cuda/nvidia/driver_.h
• Add DeviceGetAttribute in src/native/cuda/nvidia/runtime_.h

Motivation

Enable module loading, kernel function retrieval, and kernel launching through a Driver API.

Type of Change

• feat - new feature / new backend capability / new public API
• fix - bug fix
• perf - performance improvement without behavior change
• refactor - code restructuring without behavior change
• test - adding or fixing tests only
• docs - documentation only
• build / ci - build system or CI configuration
• chore - tooling, formatting, or other non-code changes
• Breaking change (requires a ! in the Conventional Commits prefix or a BREAKING CHANGE: footer)

Platforms Affected

• CPU (WITH_CPU)
• NVIDIA (WITH_NVIDIA)
• Iluvatar (WITH_ILUVATAR)
• Hygon (WITH_HYGON)
• MetaX (WITH_METAX)
• Moore (WITH_MOORE)
• Cambricon (WITH_CAMBRICON)
• Ascend (WITH_ASCEND)
• Build system / CMake / generated headers
• Public headers / installed consumer API
• Documentation only

Smoke Build and Test Result

cmake -S . -B build -DWITH_NVIDIA=ON -DINFINI_RT_BUILD_TESTING=ON
-- Configuring done
-- Generating done
-- Build files have been written to: /home/zhangshuo/projects/InfiniTensor/InfiniRT/build

cmake --build build -j
Consolidate compiler generated dependencies of target infinirt
···
[100%] Built target test_tensor_view_allocations

ctest --test-dir build --output-on-failure
···
100% tests passed, 0 tests failed out of 12
Total Test time (real) =  12.96 sec

Test Results on Supported Platforms

Platform	Affected	Build / Smoke Result	Full Result / Notes
CPU
NVIDIA	√
Iluvatar
Hygon
MetaX
Moore
Cambricon
Ascend

Full `ctest` output (optional)

ctest --test-dir build --output-on-failure
        Start  1: test_smoke
   1/12 Test  #1: test_smoke .......................   Passed    0.00 sec
        Start  2: test_core
   2/12 Test  #2: test_core ........................   Passed    0.00 sec
        Start  3: test_small_vector
   3/12 Test  #3: test_small_vector ................   Passed    0.00 sec
        Start  4: test_metadata_view
   4/12 Test  #4: test_metadata_view ...............   Passed    0.00 sec
        Start  5: test_shape_strides_storage
   5/12 Test  #5: test_shape_strides_storage .......   Passed    0.00 sec
        Start  6: test_tensor_view_allocations
   6/12 Test  #6: test_tensor_view_allocations .....   Passed    0.01 sec
        Start  7: test_cpu_runtime
   7/12 Test  #7: test_cpu_runtime .................   Passed    0.00 sec
        Start  8: test_nvidia_runtime
   8/12 Test  #8: test_nvidia_runtime ..............   Passed    3.05 sec
        Start  9: test_nvidia_graph
   9/12 Test  #9: test_nvidia_graph ................   Passed    2.84 sec
        Start 10: test_runtime_dispatch
  10/12 Test #10: test_runtime_dispatch ............   Passed    2.87 sec
        Start 11: test_install
  11/12 Test #11: test_install .....................   Passed    0.02 sec
        Start 12: test_install_consumer
  12/12 Test #12: test_install_consumer ............   Passed    2.98 sec

  100% tests passed, 0 tests failed out of 12

Benchmark / Performance Impact

N/A

Notes for Reviewers

• The Driver implementation mirrors the existing Runtime system