Apply remaining touch ups for patch release

- Issue #102: Rename root cuda to cuda_bindings - Issue #145: Explain layout change - Issue #108: Resolve how to run editable builds
NVIDIA · Oct 11, 2024 · 11a9013 · 11a9013
1 parent 797286b
commit 11a9013
Show file tree

Hide file tree

Showing 97 changed files with 137 additions and 70 deletions.
diff --git a/.gitignore b/.gitignore
@@ -11,32 +11,32 @@ __pycache__/
 .pytest_cache/
 .benchmarks/
 *.cpp
-!cuda/cuda/bindings/_lib/param_packer.cpp
-!cuda/cuda/bindings/_bindings/loader.cpp
+!cuda_bindings/cuda/bindings/_lib/param_packer.cpp
+!cuda_bindings/cuda/bindings/_bindings/loader.cpp
 
 # CUDA Python specific (auto-generated)
-cuda/cuda/bindings/_bindings/cydriver.pxd
-cuda/cuda/bindings/_bindings/cydriver.pyx
-cuda/cuda/bindings/_bindings/cynvrtc.pxd
-cuda/cuda/bindings/_bindings/cynvrtc.pyx
-cuda/cuda/bindings/_lib/cyruntime/cyruntime.pxd
-cuda/cuda/bindings/_lib/cyruntime/cyruntime.pyx
-cuda/cuda/bindings/_lib/cyruntime/utils.pxd
-cuda/cuda/bindings/_lib/cyruntime/utils.pyx
-cuda/cuda/bindings/_lib/utils.pxd
-cuda/cuda/bindings/_lib/utils.pyx
-cuda/cuda/bindings/cydriver.pxd
-cuda/cuda/bindings/cydriver.pyx
-cuda/cuda/bindings/cyruntime.pxd
-cuda/cuda/bindings/cyruntime.pyx
-cuda/cuda/bindings/cynvrtc.pxd
-cuda/cuda/bindings/cynvrtc.pyx
-cuda/cuda/bindings/driver.pxd
-cuda/cuda/bindings/driver.pyx
-cuda/cuda/bindings/runtime.pxd
-cuda/cuda/bindings/runtime.pyx
-cuda/cuda/bindings/nvrtc.pxd
-cuda/cuda/bindings/nvrtc.pyx
+cuda_bindings/cuda/bindings/_bindings/cydriver.pxd
+cuda_bindings/cuda/bindings/_bindings/cydriver.pyx
+cuda_bindings/cuda/bindings/_bindings/cynvrtc.pxd
+cuda_bindings/cuda/bindings/_bindings/cynvrtc.pyx
+cuda_bindings/cuda/bindings/_lib/cyruntime/cyruntime.pxd
+cuda_bindings/cuda/bindings/_lib/cyruntime/cyruntime.pyx
+cuda_bindings/cuda/bindings/_lib/cyruntime/utils.pxd
+cuda_bindings/cuda/bindings/_lib/cyruntime/utils.pyx
+cuda_bindings/cuda/bindings/_lib/utils.pxd
+cuda_bindings/cuda/bindings/_lib/utils.pyx
+cuda_bindings/cuda/bindings/cydriver.pxd
+cuda_bindings/cuda/bindings/cydriver.pyx
+cuda_bindings/cuda/bindings/cyruntime.pxd
+cuda_bindings/cuda/bindings/cyruntime.pyx
+cuda_bindings/cuda/bindings/cynvrtc.pxd
+cuda_bindings/cuda/bindings/cynvrtc.pyx
+cuda_bindings/cuda/bindings/driver.pxd
+cuda_bindings/cuda/bindings/driver.pyx
+cuda_bindings/cuda/bindings/runtime.pxd
+cuda_bindings/cuda/bindings/runtime.pyx
+cuda_bindings/cuda/bindings/nvrtc.pxd
+cuda_bindings/cuda/bindings/nvrtc.pyx
 
 # Distribution / packaging
 .Python

diff --git a/README.md b/README.md
@@ -42,23 +42,23 @@ Latest dependencies can be found in [requirements.txt](https://github.com/NVIDIA
 You can run the included tests with:
 
 ```
-python -m pytest
+python -m pytest tests/
 ```
 
 ### Benchmark
 
 You can run benchmark only tests with:
 
 ```
-python -m pytest --benchmark-only
+python -m pytest --benchmark-only benchmarks/
 ```
 
 ### Samples
 
 You can run the included tests with:
 
 ```
-python -m pytest examples
+python -m pytest examples/
 ```
 
 ## Examples

diff --git a/cuda/cuda/bindings/examples/__init__.py b/cuda/cuda/bindings/examples/__init__.py
diff --git a/cuda/cuda/bindings/examples/setup.cfg b/cuda/cuda/bindings/examples/setup.cfg
diff --git a/cuda/cuda/bindings/tests/__init__.py b/cuda/cuda/bindings/tests/__init__.py
diff --git a/cuda/LICENSE → cuda_bindings/LICENSE b/cuda/LICENSE → cuda_bindings/LICENSE
diff --git a/cuda/MANIFEST.in → cuda_bindings/MANIFEST.in b/cuda/MANIFEST.in → cuda_bindings/MANIFEST.in
diff --git a/cuda/README.md → cuda_bindings/README.md b/cuda/README.md → cuda_bindings/README.md
diff --git a/cuda/cuda/bindings/_bindings/__init__.py → cuda_bindings/benchmarks/__init__.py b/cuda/cuda/bindings/_bindings/__init__.py → cuda_bindings/benchmarks/__init__.py
diff --git a/cuda/cuda/bindings/benchmarks/kernels.py → cuda_bindings/benchmarks/kernels.py b/cuda/cuda/bindings/benchmarks/kernels.py → cuda_bindings/benchmarks/kernels.py
diff --git a/...da/bindings/benchmarks/perf_test_utils.py → cuda_bindings/benchmarks/perf_test_utils.py b/...da/bindings/benchmarks/perf_test_utils.py → cuda_bindings/benchmarks/perf_test_utils.py
diff --git a/cuda/pytest.ini → cuda_bindings/benchmarks/pytest.ini b/cuda/pytest.ini → cuda_bindings/benchmarks/pytest.ini
@@ -1,4 +1,3 @@
-# pytest.ini
 [pytest]
 required_plugins = pytest-benchmark
 addopts = --benchmark-skip
diff --git a/cuda/cuda/bindings/benchmarks/test_cupy.py → cuda_bindings/benchmarks/test_cupy.py b/cuda/cuda/bindings/benchmarks/test_cupy.py → cuda_bindings/benchmarks/test_cupy.py
diff --git a/...indings/benchmarks/test_launch_latency.py → ...indings/benchmarks/test_launch_latency.py b/...indings/benchmarks/test_launch_latency.py → ...indings/benchmarks/test_launch_latency.py
diff --git a/cuda/cuda/bindings/benchmarks/test_numba.py → cuda_bindings/benchmarks/test_numba.py b/cuda/cuda/bindings/benchmarks/test_numba.py → cuda_bindings/benchmarks/test_numba.py
diff --git a/...ngs/benchmarks/test_pointer_attributes.py → ...ngs/benchmarks/test_pointer_attributes.py b/...ngs/benchmarks/test_pointer_attributes.py → ...ngs/benchmarks/test_pointer_attributes.py
diff --git a/cuda/cuda/__init__.pxd → cuda_bindings/cuda/__init__.pxd b/cuda/cuda/__init__.pxd → cuda_bindings/cuda/__init__.pxd
diff --git a/cuda/cuda/__init__.py → cuda_bindings/cuda/__init__.py b/cuda/cuda/__init__.py → cuda_bindings/cuda/__init__.py
diff --git a/cuda/cuda/bindings/__init__.pxd → cuda_bindings/cuda/bindings/__init__.pxd b/cuda/cuda/bindings/__init__.pxd → cuda_bindings/cuda/bindings/__init__.pxd
diff --git a/cuda/cuda/bindings/__init__.py → cuda_bindings/cuda/bindings/__init__.py b/cuda/cuda/bindings/__init__.py → cuda_bindings/cuda/bindings/__init__.py
diff --git a/cuda/cuda/bindings/_lib/__init__.py → ...dings/cuda/bindings/_bindings/__init__.py b/cuda/cuda/bindings/_lib/__init__.py → ...dings/cuda/bindings/_bindings/__init__.py
diff --git a/cuda/cuda/bindings/_bindings/cydriver.pxd.in → ...s/cuda/bindings/_bindings/cydriver.pxd.in b/cuda/cuda/bindings/_bindings/cydriver.pxd.in → ...s/cuda/bindings/_bindings/cydriver.pxd.in
diff --git a/cuda/cuda/bindings/_bindings/cydriver.pyx.in → ...s/cuda/bindings/_bindings/cydriver.pyx.in b/cuda/cuda/bindings/_bindings/cydriver.pyx.in → ...s/cuda/bindings/_bindings/cydriver.pyx.in
diff --git a/cuda/cuda/bindings/_bindings/cynvrtc.pxd.in → ...gs/cuda/bindings/_bindings/cynvrtc.pxd.in b/cuda/cuda/bindings/_bindings/cynvrtc.pxd.in → ...gs/cuda/bindings/_bindings/cynvrtc.pxd.in
diff --git a/cuda/cuda/bindings/_bindings/cynvrtc.pyx.in → ...gs/cuda/bindings/_bindings/cynvrtc.pyx.in b/cuda/cuda/bindings/_bindings/cynvrtc.pyx.in → ...gs/cuda/bindings/_bindings/cynvrtc.pyx.in
diff --git a/cuda/cuda/bindings/_bindings/loader.cpp → ...ndings/cuda/bindings/_bindings/loader.cpp b/cuda/cuda/bindings/_bindings/loader.cpp → ...ndings/cuda/bindings/_bindings/loader.cpp
diff --git a/cuda/cuda/bindings/_bindings/loader.h → ...bindings/cuda/bindings/_bindings/loader.h b/cuda/cuda/bindings/_bindings/loader.h → ...bindings/cuda/bindings/_bindings/loader.h
diff --git a/cuda/cuda/bindings/_bindings/loader.pxd → ...ndings/cuda/bindings/_bindings/loader.pxd b/cuda/cuda/bindings/_bindings/loader.pxd → ...ndings/cuda/bindings/_bindings/loader.pxd
diff --git a/.../cuda/bindings/_lib/cyruntime/__init__.py → cuda_bindings/cuda/bindings/_lib/__init__.py b/.../cuda/bindings/_lib/cyruntime/__init__.py → cuda_bindings/cuda/bindings/_lib/__init__.py
diff --git a/cuda/cuda/bindings/benchmarks/__init__.py → .../cuda/bindings/_lib/cyruntime/__init__.py b/cuda/cuda/bindings/benchmarks/__init__.py → .../cuda/bindings/_lib/cyruntime/__init__.py
diff --git a/.../bindings/_lib/cyruntime/cyruntime.pxd.in → .../bindings/_lib/cyruntime/cyruntime.pxd.in b/.../bindings/_lib/cyruntime/cyruntime.pxd.in → .../bindings/_lib/cyruntime/cyruntime.pxd.in
diff --git a/.../bindings/_lib/cyruntime/cyruntime.pyx.in → .../bindings/_lib/cyruntime/cyruntime.pyx.in b/.../bindings/_lib/cyruntime/cyruntime.pyx.in → .../bindings/_lib/cyruntime/cyruntime.pyx.in
diff --git a/...cuda/bindings/_lib/cyruntime/utils.pxd.in → ...cuda/bindings/_lib/cyruntime/utils.pxd.in b/...cuda/bindings/_lib/cyruntime/utils.pxd.in → ...cuda/bindings/_lib/cyruntime/utils.pxd.in
diff --git a/...cuda/bindings/_lib/cyruntime/utils.pyx.in → ...cuda/bindings/_lib/cyruntime/utils.pyx.in b/...cuda/bindings/_lib/cyruntime/utils.pyx.in → ...cuda/bindings/_lib/cyruntime/utils.pyx.in
diff --git a/cuda/cuda/bindings/_lib/dlfcn.pxd → cuda_bindings/cuda/bindings/_lib/dlfcn.pxd b/cuda/cuda/bindings/_lib/dlfcn.pxd → cuda_bindings/cuda/bindings/_lib/dlfcn.pxd
diff --git a/cuda/cuda/bindings/_lib/param_packer.cpp → ...dings/cuda/bindings/_lib/param_packer.cpp b/cuda/cuda/bindings/_lib/param_packer.cpp → ...dings/cuda/bindings/_lib/param_packer.cpp
diff --git a/cuda/cuda/bindings/_lib/param_packer.h → ...indings/cuda/bindings/_lib/param_packer.h b/cuda/cuda/bindings/_lib/param_packer.h → ...indings/cuda/bindings/_lib/param_packer.h
diff --git a/cuda/cuda/bindings/_lib/param_packer.pxd → ...dings/cuda/bindings/_lib/param_packer.pxd b/cuda/cuda/bindings/_lib/param_packer.pxd → ...dings/cuda/bindings/_lib/param_packer.pxd
diff --git a/cuda/cuda/bindings/_lib/utils.pxd.in → ..._bindings/cuda/bindings/_lib/utils.pxd.in b/cuda/cuda/bindings/_lib/utils.pxd.in → ..._bindings/cuda/bindings/_lib/utils.pxd.in
diff --git a/cuda/cuda/bindings/_lib/utils.pyx.in → ..._bindings/cuda/bindings/_lib/utils.pyx.in b/cuda/cuda/bindings/_lib/utils.pyx.in → ..._bindings/cuda/bindings/_lib/utils.pyx.in
diff --git a/cuda/cuda/bindings/_version.py → cuda_bindings/cuda/bindings/_version.py b/cuda/cuda/bindings/_version.py → cuda_bindings/cuda/bindings/_version.py
diff --git a/cuda/cuda/bindings/cydriver.pxd.in → cuda_bindings/cuda/bindings/cydriver.pxd.in b/cuda/cuda/bindings/cydriver.pxd.in → cuda_bindings/cuda/bindings/cydriver.pxd.in
diff --git a/cuda/cuda/bindings/cydriver.pyx.in → cuda_bindings/cuda/bindings/cydriver.pyx.in b/cuda/cuda/bindings/cydriver.pyx.in → cuda_bindings/cuda/bindings/cydriver.pyx.in
diff --git a/cuda/cuda/bindings/cynvrtc.pxd.in → cuda_bindings/cuda/bindings/cynvrtc.pxd.in b/cuda/cuda/bindings/cynvrtc.pxd.in → cuda_bindings/cuda/bindings/cynvrtc.pxd.in
diff --git a/cuda/cuda/bindings/cynvrtc.pyx.in → cuda_bindings/cuda/bindings/cynvrtc.pyx.in b/cuda/cuda/bindings/cynvrtc.pyx.in → cuda_bindings/cuda/bindings/cynvrtc.pyx.in
diff --git a/cuda/cuda/bindings/cyruntime.pxd.in → cuda_bindings/cuda/bindings/cyruntime.pxd.in b/cuda/cuda/bindings/cyruntime.pxd.in → cuda_bindings/cuda/bindings/cyruntime.pxd.in
diff --git a/cuda/cuda/bindings/cyruntime.pyx.in → cuda_bindings/cuda/bindings/cyruntime.pyx.in b/cuda/cuda/bindings/cyruntime.pyx.in → cuda_bindings/cuda/bindings/cyruntime.pyx.in
diff --git a/cuda/cuda/bindings/driver.pxd.in → cuda_bindings/cuda/bindings/driver.pxd.in b/cuda/cuda/bindings/driver.pxd.in → cuda_bindings/cuda/bindings/driver.pxd.in
diff --git a/cuda/cuda/bindings/driver.pyx.in → cuda_bindings/cuda/bindings/driver.pyx.in b/cuda/cuda/bindings/driver.pyx.in → cuda_bindings/cuda/bindings/driver.pyx.in
diff --git a/cuda/cuda/bindings/nvrtc.pxd.in → cuda_bindings/cuda/bindings/nvrtc.pxd.in b/cuda/cuda/bindings/nvrtc.pxd.in → cuda_bindings/cuda/bindings/nvrtc.pxd.in
diff --git a/cuda/cuda/bindings/nvrtc.pyx.in → cuda_bindings/cuda/bindings/nvrtc.pyx.in b/cuda/cuda/bindings/nvrtc.pyx.in → cuda_bindings/cuda/bindings/nvrtc.pyx.in
diff --git a/cuda/cuda/bindings/runtime.pxd.in → cuda_bindings/cuda/bindings/runtime.pxd.in b/cuda/cuda/bindings/runtime.pxd.in → cuda_bindings/cuda/bindings/runtime.pxd.in
diff --git a/cuda/cuda/bindings/runtime.pyx.in → cuda_bindings/cuda/bindings/runtime.pyx.in b/cuda/cuda/bindings/runtime.pyx.in → cuda_bindings/cuda/bindings/runtime.pyx.in
diff --git a/cuda/cuda/ccuda.pxd → cuda_bindings/cuda/ccuda.pxd b/cuda/cuda/ccuda.pxd → cuda_bindings/cuda/ccuda.pxd
diff --git a/cuda/cuda/ccuda.pyx → cuda_bindings/cuda/ccuda.pyx b/cuda/cuda/ccuda.pyx → cuda_bindings/cuda/ccuda.pyx
diff --git a/cuda/cuda/ccudart.pxd → cuda_bindings/cuda/ccudart.pxd b/cuda/cuda/ccudart.pxd → cuda_bindings/cuda/ccudart.pxd
diff --git a/cuda/cuda/ccudart.pyx → cuda_bindings/cuda/ccudart.pyx b/cuda/cuda/ccudart.pyx → cuda_bindings/cuda/ccudart.pyx
diff --git a/cuda/cuda/cnvrtc.pxd → cuda_bindings/cuda/cnvrtc.pxd b/cuda/cuda/cnvrtc.pxd → cuda_bindings/cuda/cnvrtc.pxd
diff --git a/cuda/cuda/cnvrtc.pyx → cuda_bindings/cuda/cnvrtc.pyx b/cuda/cuda/cnvrtc.pyx → cuda_bindings/cuda/cnvrtc.pyx
diff --git a/cuda/cuda/cuda.pyx → cuda_bindings/cuda/cuda.pyx b/cuda/cuda/cuda.pyx → cuda_bindings/cuda/cuda.pyx
diff --git a/cuda/cuda/cudart.pyx → cuda_bindings/cuda/cudart.pyx b/cuda/cuda/cudart.pyx → cuda_bindings/cuda/cudart.pyx
diff --git a/cuda/cuda/nvrtc.pyx → cuda_bindings/cuda/nvrtc.pyx b/cuda/cuda/nvrtc.pyx → cuda_bindings/cuda/nvrtc.pyx
diff --git a/...amples/0_Introduction/clock_nvrtc_test.py → ...amples/0_Introduction/clock_nvrtc_test.py b/...amples/0_Introduction/clock_nvrtc_test.py → ...amples/0_Introduction/clock_nvrtc_test.py
@@ -7,8 +7,8 @@
 # is strictly prohibited.
 import numpy as np
 from cuda import cuda
-from cuda.bindings.examples.common import common
-from cuda.bindings.examples.common.helper_cuda import checkCudaErrors, findCudaDevice
+from common import common
+from common.helper_cuda import checkCudaErrors, findCudaDevice
 
 clock_nvrtc = '''\
 extern "C" __global__  void timedReduction(const float *hinput, float *output, clock_t *timer)

diff --git a/...Introduction/simpleCubemapTexture_test.py → ...Introduction/simpleCubemapTexture_test.py b/...Introduction/simpleCubemapTexture_test.py → ...Introduction/simpleCubemapTexture_test.py
@@ -11,8 +11,8 @@
 import sys
 import time
 from cuda import cuda, cudart
-from cuda.bindings.examples.common import common
-from cuda.bindings.examples.common.helper_cuda import checkCudaErrors, findCudaDevice
+from common import common
+from common.helper_cuda import checkCudaErrors, findCudaDevice
 
 simpleCubemapTexture = '''\
 extern "C"

diff --git a/...examples/0_Introduction/simpleP2P_test.py → ...examples/0_Introduction/simpleP2P_test.py b/...examples/0_Introduction/simpleP2P_test.py → ...examples/0_Introduction/simpleP2P_test.py
@@ -9,8 +9,8 @@
 import numpy as np
 import sys
 from cuda import cuda, cudart
-from cuda.bindings.examples.common import common
-from cuda.bindings.examples.common.helper_cuda import checkCudaErrors
+from common import common
+from common.helper_cuda import checkCudaErrors
 
 simplep2p = '''\
 extern "C"

diff --git a/...les/0_Introduction/simpleZeroCopy_test.py → ...les/0_Introduction/simpleZeroCopy_test.py b/...les/0_Introduction/simpleZeroCopy_test.py → ...les/0_Introduction/simpleZeroCopy_test.py
@@ -11,9 +11,9 @@
 import random as rnd
 import sys
 from cuda import cuda, cudart
-from cuda.bindings.examples.common import common
-from cuda.bindings.examples.common.helper_cuda import checkCudaErrors
-from cuda.bindings.examples.common.helper_string import checkCmdLineFlag
+from common import common
+from common.helper_cuda import checkCudaErrors
+from common.helper_string import checkCmdLineFlag
 
 simpleZeroCopy = '''\
 extern "C"

diff --git a/.../0_Introduction/systemWideAtomics_test.py → .../0_Introduction/systemWideAtomics_test.py b/.../0_Introduction/systemWideAtomics_test.py → .../0_Introduction/systemWideAtomics_test.py
@@ -10,8 +10,8 @@
 import sys
 import os
 from cuda import cuda, cudart
-from cuda.bindings.examples.common import common
-from cuda.bindings.examples.common.helper_cuda import checkCudaErrors, findCudaDevice
+from common import common
+from common.helper_cuda import checkCudaErrors, findCudaDevice
 
 systemWideAtomics =  '''\
 #define LOOP_NUM 50

diff --git a/...mples/0_Introduction/vectorAddDrv_test.py → ...mples/0_Introduction/vectorAddDrv_test.py b/...mples/0_Introduction/vectorAddDrv_test.py → ...mples/0_Introduction/vectorAddDrv_test.py
@@ -9,8 +9,8 @@
 import math
 import numpy as np
 from cuda import cuda
-from cuda.bindings.examples.common import common
-from cuda.bindings.examples.common.helper_cuda import checkCudaErrors, findCudaDeviceDRV
+from common import common
+from common.helper_cuda import checkCudaErrors, findCudaDeviceDRV
 
 vectorAddDrv = '''\
 /* Vector addition: C = A + B.

diff --git a/...ples/0_Introduction/vectorAddMMAP_test.py → ...ples/0_Introduction/vectorAddMMAP_test.py b/...ples/0_Introduction/vectorAddMMAP_test.py → ...ples/0_Introduction/vectorAddMMAP_test.py
@@ -10,8 +10,8 @@
 import numpy as np
 import sys
 from cuda import cuda
-from cuda.bindings.examples.common import common
-from cuda.bindings.examples.common.helper_cuda import checkCudaErrors, findCudaDeviceDRV
+from common import common
+from common.helper_cuda import checkCudaErrors, findCudaDeviceDRV
 
 vectorAddMMAP = '''\
 /* Vector addition: C = A + B.

diff --git a/...echniques/streamOrderedAllocation_test.py → ...echniques/streamOrderedAllocation_test.py b/...echniques/streamOrderedAllocation_test.py → ...echniques/streamOrderedAllocation_test.py
@@ -11,9 +11,9 @@
 import random as rnd
 import sys
 from cuda import cuda, cudart
-from cuda.bindings.examples.common import common
-from cuda.bindings.examples.common.helper_cuda import checkCudaErrors, findCudaDevice
-from cuda.bindings.examples.common.helper_string import checkCmdLineFlag
+from common import common
+from common.helper_cuda import checkCudaErrors, findCudaDevice
+from common.helper_string import checkCmdLineFlag
 
 streamOrderedAllocation = '''\
 /* Add two vectors on the GPU */
@@ -166,6 +166,7 @@ def streamOrderedAllocationPostSync(dev, nelem, a, b, c) :
     return errorNorm/refNorm < 1.e-6
 
 def main():
+    cuda.cuInit(0)
     if checkCmdLineFlag("help"):
         print("Usage:  streamOrderedAllocation [OPTION]\n");
         print("Options:");

diff --git a/...A_Features/globalToShmemAsyncCopy_test.py → ...A_Features/globalToShmemAsyncCopy_test.py b/...A_Features/globalToShmemAsyncCopy_test.py → ...A_Features/globalToShmemAsyncCopy_test.py
@@ -12,9 +12,9 @@
 import pytest
 from cuda import cuda, cudart
 from enum import Enum
-from cuda.bindings.examples.common import common
-from cuda.bindings.examples.common.helper_cuda import checkCudaErrors, findCudaDevice
-from cuda.bindings.examples.common.helper_string import checkCmdLineFlag, getCmdLineArgumentInt
+from common import common
+from common.helper_cuda import checkCudaErrors, findCudaDevice
+from common.helper_string import checkCmdLineFlag, getCmdLineArgumentInt
 
 blockSize = 16
 class kernels(Enum):

diff --git a/.../3_CUDA_Features/simpleCudaGraphs_test.py → .../3_CUDA_Features/simpleCudaGraphs_test.py b/.../3_CUDA_Features/simpleCudaGraphs_test.py → .../3_CUDA_Features/simpleCudaGraphs_test.py
@@ -10,8 +10,8 @@
 import pytest
 import random as rnd
 from cuda import cuda, cudart
-from cuda.bindings.examples.common import common
-from cuda.bindings.examples.common.helper_cuda import checkCudaErrors, findCudaDevice
+from common import common
+from common.helper_cuda import checkCudaErrors, findCudaDevice
 
 THREADS_PER_BLOCK = 512
 GRAPH_LAUNCH_ITERATIONS = 3

diff --git a/...ies/conjugateGradientMultiBlockCG_test.py → ...ies/conjugateGradientMultiBlockCG_test.py b/...ies/conjugateGradientMultiBlockCG_test.py → ...ies/conjugateGradientMultiBlockCG_test.py
@@ -10,8 +10,8 @@
 import numpy as np
 import sys
 from cuda import cuda, cudart
-from cuda.bindings.examples.common import common
-from cuda.bindings.examples.common.helper_cuda import checkCudaErrors, findCudaDevice
+from common import common
+from common.helper_cuda import checkCudaErrors, findCudaDevice
 from random import random
 
 conjugateGradientMultiBlockCG = '''\

diff --git a/cuda/cuda/bindings/examples/common/common.py → cuda_bindings/examples/common/common.py b/cuda/cuda/bindings/examples/common/common.py → cuda_bindings/examples/common/common.py
@@ -9,7 +9,7 @@
 import numpy as np
 import os
 from cuda import cuda, cudart, nvrtc
-from cuda.bindings.examples.common.helper_cuda import checkCudaErrors
+from common.helper_cuda import checkCudaErrors
 
 class KernelHelper:
     def __init__(self, code, devID):

diff --git a/...a/bindings/examples/common/helper_cuda.py → cuda_bindings/examples/common/helper_cuda.py b/...a/bindings/examples/common/helper_cuda.py → cuda_bindings/examples/common/helper_cuda.py
@@ -6,7 +6,7 @@
 # this software and related documentation outside the terms of the EULA
 # is strictly prohibited.
 from cuda import cuda, cudart, nvrtc
-from cuda.bindings.examples.common.helper_string import getCmdLineArgumentInt, checkCmdLineFlag
+from common.helper_string import getCmdLineArgumentInt, checkCmdLineFlag
 
 def _cudaGetErrorEnum(error):
     if isinstance(error, cuda.CUresult):

diff --git a/...bindings/examples/common/helper_string.py → ...bindings/examples/common/helper_string.py b/...bindings/examples/common/helper_string.py → ...bindings/examples/common/helper_string.py
diff --git a/...ngs/examples/extra/isoFDModelling_test.py → ...ngs/examples/extra/isoFDModelling_test.py b/...ngs/examples/extra/isoFDModelling_test.py → ...ngs/examples/extra/isoFDModelling_test.py
@@ -8,8 +8,8 @@
 import numpy as np
 import time
 from cuda import cuda, cudart
-from cuda.bindings.examples.common import common
-from cuda.bindings.examples.common.helper_cuda import checkCudaErrors
+from common import common
+from common.helper_cuda import checkCudaErrors
 
 isoPropagator = '''\
 extern "C"

diff --git a/...ndings/examples/extra/jit_program_test.py → ...ndings/examples/extra/jit_program_test.py b/...ndings/examples/extra/jit_program_test.py → ...ndings/examples/extra/jit_program_test.py
diff --git a/...ndings/examples/extra/numba_emm_plugin.py → ...ndings/examples/extra/numba_emm_plugin.py b/...ndings/examples/extra/numba_emm_plugin.py → ...ndings/examples/extra/numba_emm_plugin.py
diff --git a/cuda_bindings/examples/pytest.ini b/cuda_bindings/examples/pytest.ini
@@ -0,0 +1,4 @@
+[pytest]
+python_files = *_test.py
+python_functions = main
+pythonpath = .
diff --git a/cuda/pyproject.toml → cuda_bindings/pyproject.toml b/cuda/pyproject.toml → cuda_bindings/pyproject.toml
diff --git a/cuda/requirements.txt → cuda_bindings/requirements.txt b/cuda/requirements.txt → cuda_bindings/requirements.txt
diff --git a/cuda/setup.py → cuda_bindings/setup.py b/cuda/setup.py → cuda_bindings/setup.py
@@ -95,6 +95,10 @@
                      cache='./cache_{}'.format(library.split('.')[0]) if PARSER_CACHING else None,
                      replace=replace)
 
+    if library == 'driver':
+        CUDA_VERSION = parser.defs['macros']['CUDA_VERSION'] if 'CUDA_VERSION' in parser.defs['macros'] else 'Unknown'
+        print(f'Found CUDA_VERSION: {CUDA_VERSION}')
+
     # Combine types with others since they sometimes get tangled
     found_types += {key for key in parser.defs['types']}
     found_types += {key for key in parser.defs['structs']}
@@ -225,7 +229,7 @@ def do_cythonize(extensions):
     # public (deprecated, to be removed)
     ["cuda/*.pyx"],
     # tests
-    ["cuda/bindings/tests/*.pyx"],
+    ["tests/*.pyx"],
 ]
 
 for sources in sources_list:
@@ -256,9 +260,9 @@ def finalize_options(self):
 setup(
     version=versioneer.get_version(),
     ext_modules=do_cythonize(extensions),
-    packages=find_packages(include=["cuda.cuda", "cuda.cuda.*", "cuda.cuda.bindings", "cuda.cuda.bindings._bindings", "cuda.cuda.bindings._lib", "cuda.cuda.bindings._lib.cyruntime", "cuda.cuda.bindings.tests"]),
+    packages=find_packages(include=["cuda.cuda", "cuda.cuda.*", "cuda.cuda.bindings", "cuda.cuda.bindings._bindings", "cuda.cuda.bindings._lib", "cuda.cuda.bindings._lib.cyruntime", "tests"]),
     package_data=dict.fromkeys(
-        find_packages(include=["cuda.cuda", "cuda.cuda.*", "cuda.cuda.bindings", "cuda.cuda.bindings._bindings", "cuda.cuda.bindings._lib", "cuda.cuda.bindings._lib.cyruntime", "cuda.cuda.bindings.tests"]),
+        find_packages(include=["cuda.cuda", "cuda.cuda.*", "cuda.cuda.bindings", "cuda.cuda.bindings._bindings", "cuda.cuda.bindings._lib", "cuda.cuda.bindings._lib.cyruntime", "tests"]),
         ["*.pxd", "*.pyx", "*.py", "*.h", "*.cpp"],
     ),
     cmdclass=cmdclass,

diff --git a/cuda/cuda/bindings/tests/test_ccuda.pyx → cuda_bindings/tests/test_ccuda.pyx b/cuda/cuda/bindings/tests/test_ccuda.pyx → cuda_bindings/tests/test_ccuda.pyx
diff --git a/cuda/cuda/bindings/tests/test_ccudart.pyx → cuda_bindings/tests/test_ccudart.pyx b/cuda/cuda/bindings/tests/test_ccudart.pyx → cuda_bindings/tests/test_ccudart.pyx
diff --git a/cuda/cuda/bindings/tests/test_cuda.py → cuda_bindings/tests/test_cuda.py b/cuda/cuda/bindings/tests/test_cuda.py → cuda_bindings/tests/test_cuda.py
diff --git a/cuda/cuda/bindings/tests/test_cudart.py → cuda_bindings/tests/test_cudart.py b/cuda/cuda/bindings/tests/test_cudart.py → cuda_bindings/tests/test_cudart.py
diff --git a/cuda/cuda/bindings/tests/test_cython.py → cuda_bindings/tests/test_cython.py b/cuda/cuda/bindings/tests/test_cython.py → cuda_bindings/tests/test_cython.py
@@ -22,9 +22,9 @@ def wrapped(*args, **kwargs):
     return wrapped
 
 
-cython_test_modules = ["cuda.bindings.tests.test_ccuda",
-                       "cuda.bindings.tests.test_ccudart",
-                       "cuda.bindings.tests.test_interoperability_cython"]
+cython_test_modules = ["test_ccuda",
+                       "test_ccudart",
+                       "test_interoperability_cython"]
 
 
 for mod in cython_test_modules:

diff --git a/...a/bindings/tests/test_interoperability.py → cuda_bindings/tests/test_interoperability.py b/...a/bindings/tests/test_interoperability.py → cuda_bindings/tests/test_interoperability.py
diff --git a/...gs/tests/test_interoperability_cython.pyx → ...gs/tests/test_interoperability_cython.pyx b/...gs/tests/test_interoperability_cython.pyx → ...gs/tests/test_interoperability_cython.pyx
diff --git a/.../cuda/bindings/tests/test_kernelParams.py → cuda_bindings/tests/test_kernelParams.py b/.../cuda/bindings/tests/test_kernelParams.py → cuda_bindings/tests/test_kernelParams.py
diff --git a/cuda/cuda/bindings/tests/test_nvrtc.py → cuda_bindings/tests/test_nvrtc.py b/cuda/cuda/bindings/tests/test_nvrtc.py → cuda_bindings/tests/test_nvrtc.py
diff --git a/docs/_sources/release/12.6.1-notes.md.txt b/docs/_sources/release/12.6.1-notes.md.txt
@@ -3,7 +3,27 @@
 Released on October 7, 2024
 
 ## Hightlights
-- Resolve [Issue #90](https://github.com/NVIDIA/cuda-python/issues/90): Use new layout in preperation for cuda-python becoming a metapackage
+- Resolve [Issue #90](https://github.com/NVIDIA/cuda-python/issues/90): Use new layout in preparation for cuda-python becoming a metapackage
+- Resolve [Issue #75](https://github.com/NVIDIA/cuda-python/issues/75): CUDA namespace cleanup
+
+## CUDA namespace cleanup with a new module layout
+
+[Issue #75](https://github.com/NVIDIA/cuda-python/issues/75) explains in detail what the new module layout is, what problem it fixes and how it impacts the users. However for the sake of completeness, this release notes will highlight key points of this change.
+
+Before this change, `cuda-python` was tightly coupled to CUDA Toolkit releases and all new features would inherit this coupling regardless of their applicability. As we develop new features, this coupling was becoming overly restrictive and motivated a new solution: Convert `cuda-python` into a metapackage where we use `cuda` as a namespace with existing bindings code moved to a `cuda_bindings` subpackage.
+
+This patch release applies the new module layout for the bindings as follows:
+- `cuda.cuda` -> `cuda.bindings.driver`
+- `cuda.ccuda` -> `cuda.bindings.cydriver`
+- `cuda.cudart` -> `cuda.bindings.runtime`
+- `cuda.ccudart` -> `cuda.bindings.cyruntime`
+- `cuda.nvrtc` -> `cuda.bindings.nvrtc`
+- `cuda.cnvrtc` -> `cuda.bindings.cynvrtc`
+
+Deprecation warnings are turned on as a notice to switch to the new module layout.
+
+```{note} This is non-breaking, backwards compatible change. All old module path will continue work as they "forward" user calls towards the new layout.
+```
 
 ## Limitations
 

diff --git a/docs/release.html b/docs/release.html
@@ -236,6 +236,7 @@ <h1>Release Notes<a class="headerlink" href="#release-notes" title="Permalink to
 <ul>
 <li class="toctree-l1"><a class="reference internal" href="release/12.6.1-notes.html">    12.6.1</a><ul>
 <li class="toctree-l2"><a class="reference internal" href="release/12.6.1-notes.html#hightlights">Hightlights</a></li>
+<li class="toctree-l2"><a class="reference internal" href="release/12.6.1-notes.html#cuda-namespace-cleanup-with-a-new-module-layout">CUDA namespace cleanup with a new module layout</a></li>
 <li class="toctree-l2"><a class="reference internal" href="release/12.6.1-notes.html#limitations">Limitations</a><ul>
 <li class="toctree-l3"><a class="reference internal" href="release/12.6.1-notes.html#cuda-functions-not-supported-in-this-release">CUDA Functions Not Supported in this Release</a></li>
 </ul>

diff --git a/docs/release/12.6.1-notes.html b/docs/release/12.6.1-notes.html
@@ -236,9 +236,29 @@ <h1>CUDA Python 12.6.1 Release notes<a class="headerlink" href="#cuda-python-12-
 <section id="hightlights">
 <h2>Hightlights<a class="headerlink" href="#hightlights" title="Permalink to this heading">#</a></h2>
 <ul class="simple">
-<li><p>Resolve <a class="reference external" href="https://github.com/NVIDIA/cuda-python/issues/90">Issue #90</a>: Use new layout in preperation for cuda-python becoming a metapackage</p></li>
+<li><p>Resolve <a class="reference external" href="https://github.com/NVIDIA/cuda-python/issues/90">Issue #90</a>: Use new layout in preparation for cuda-python becoming a metapackage</p></li>
+<li><p>Resolve <a class="reference external" href="https://github.com/NVIDIA/cuda-python/issues/75">Issue #75</a>: CUDA namespace cleanup</p></li>
 </ul>
 </section>
+<section id="cuda-namespace-cleanup-with-a-new-module-layout">
+<h2>CUDA namespace cleanup with a new module layout<a class="headerlink" href="#cuda-namespace-cleanup-with-a-new-module-layout" title="Permalink to this heading">#</a></h2>
+<p><a class="reference external" href="https://github.com/NVIDIA/cuda-python/issues/75">Issue #75</a> explains in detail what the new module layout is, what problem it fixes and how it impacts the users. However for the sake of completeness, this release notes will highlight key points of this change.</p>
+<p>Before this change, <code class="docutils literal notranslate"><span class="pre">cuda-python</span></code> was tightly coupled to CUDA Toolkit releases and all new features would inherit this coupling regardless of their applicability. As we develop new features, this coupling was becoming overly restrictive and motivated a new solution: Convert <code class="docutils literal notranslate"><span class="pre">cuda-python</span></code> into a metapackage where we use <code class="docutils literal notranslate"><span class="pre">cuda</span></code> as a namespace with existing bindings code moved to a <code class="docutils literal notranslate"><span class="pre">cuda_bindings</span></code> subpackage.</p>
+<p>This patch release applies the new module layout for the bindings as follows:</p>
+<ul class="simple">
+<li><p><code class="docutils literal notranslate"><span class="pre">cuda.cuda</span></code> -&gt; <code class="docutils literal notranslate"><span class="pre">cuda.bindings.driver</span></code></p></li>
+<li><p><code class="docutils literal notranslate"><span class="pre">cuda.ccuda</span></code> -&gt; <code class="docutils literal notranslate"><span class="pre">cuda.bindings.cydriver</span></code></p></li>
+<li><p><code class="docutils literal notranslate"><span class="pre">cuda.cudart</span></code> -&gt; <code class="docutils literal notranslate"><span class="pre">cuda.bindings.runtime</span></code></p></li>
+<li><p><code class="docutils literal notranslate"><span class="pre">cuda.ccudart</span></code> -&gt; <code class="docutils literal notranslate"><span class="pre">cuda.bindings.cyruntime</span></code></p></li>
+<li><p><code class="docutils literal notranslate"><span class="pre">cuda.nvrtc</span></code> -&gt; <code class="docutils literal notranslate"><span class="pre">cuda.bindings.nvrtc</span></code></p></li>
+<li><p><code class="docutils literal notranslate"><span class="pre">cuda.cnvrtc</span></code> -&gt; <code class="docutils literal notranslate"><span class="pre">cuda.bindings.cynvrtc</span></code></p></li>
+</ul>
+<p>Deprecation warnings are turned on as a notice to switch to the new module layout.</p>
+<div class="admonition note">
+<p class="admonition-title">Note</p>
+<p>This is non-breaking, backwards compatible change. All old module path will continue work as they “forward” user calls towards the new layout.</p>
+</div>
+</section>
 <section id="limitations">
 <h2>Limitations<a class="headerlink" href="#limitations" title="Permalink to this heading">#</a></h2>
 <section id="cuda-functions-not-supported-in-this-release">
@@ -334,6 +354,7 @@ <h3>CUDA Functions Not Supported in this Release<a class="headerlink" href="#cud
             <ul>
 <li><a class="reference internal" href="#">CUDA Python 12.6.1 Release notes</a><ul>
 <li><a class="reference internal" href="#hightlights">Hightlights</a></li>
+<li><a class="reference internal" href="#cuda-namespace-cleanup-with-a-new-module-layout">CUDA namespace cleanup with a new module layout</a></li>
 <li><a class="reference internal" href="#limitations">Limitations</a><ul>
 <li><a class="reference internal" href="#cuda-functions-not-supported-in-this-release">CUDA Functions Not Supported in this Release</a></li>
 </ul>

diff --git a/docs/searchindex.js b/docs/searchindex.js