PortPilot
Benchmark · pre-generated, always available← back to leaderboard
Migration Readiness Report

inducer/pycuda

https://github.com/inducer/pycuda
Manual blockers present

ROCm Migration Advisory — pycuda

This repo is roughly halfway to ROCm readiness: 3 items already work as-is and 312 are mechanical fixes (e.g., header swaps, API renames, cuda*hip* symbol substitutions), but a single manual blocker (Bucket C) gates the whole effort and must be resolved before any mechanical sweep matters. That blocker almost certainly involves PyCUDA's hard dependency on the CUDA driver/runtime API surface at the C++ extension layer — you'll need to decide between porting the extension to HIP (via hipify + manual

3Works as-is
312Mechanical change
1Manual blocker
Findings
316
Python files
77
Files scanned
123
Custom CUDA kernels
none

Detected but out of scope (not analyzed): C++

Findings by file

316 findings · 70 files
pycuda/compiler.py· 15
B
PyCUDA dependency
from pycuda.driver import CompileError
:55

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.driver import CompileError
:63

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.characterize import platform_bits
:96

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.driver import CompileError
:145

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

A
Device string "cuda"
os.path.pardir, "cuda"))
:204

The literal device string "cuda" resolves to the active AMD GPU on ROCm PyTorch. .to("cuda") / device="cuda" need no edits.

RecommendNo change required on ROCm PyTorch.

B
PyCUDA dependency
from pycuda.driver import Error
:239

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.driver import Context
:242

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.driver import CUDA_DEBUGGING
:248

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.driver import Context
:298

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.driver import module_from_buffer
:353

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.driver import Context
:401

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.driver import Linker
:412

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

C
Inline PTX assembly
def add_source(self, source, nvcc_options=None, name="kernel.ptx"):
:495

PTX is NVIDIA-specific intermediate assembly; ROCm uses GCN/ROCm ISA or HSACO binaries. The add_source method signature itself is benign, but any PTX content passed to it must be recompiled or rewritten as HIP/GCN. The default name="kernel.ptx" is cosmetic but misleading on ROCm.

RecommendRewrite the PTX block in HIP or a portable high-level equivalent.

Suggested change · advisory
--- a/pycuda/compiler.py
+++ b/pycuda/compiler.py
@@ -495,1 +495,1 @@
-def add_source(self, source, nvcc_options=None, name="kernel.ptx"):
+def add_source(self, source, nvcc_options=None, name="kernel.hsaco"):
B
PyCUDA dependency
from pycuda.driver import jit_input_type
:508

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.driver import jit_input_type
:529

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/cai_cupy_arrays.py· 3
B
CuPy dependency
import cupy as cp
:5

CuPy itself is not CUDA-exclusive; AMD provides a ROCm/HIP-backed build of CuPy (cupy-rocm) that exposes the same cupy Python API. The import cupy as cp line can remain unchanged as long as the environment installs the ROCm variant rather than the CUDA variant.

RecommendInstall the ROCm CuPy build (cupy-rocm-6-0) instead of cupy-cudaXXx.

B
PyCUDA dependency
import pycuda.autoinit # noqa
:7

PyCUDA is a CUDA-only Python GPU binding and will not work on AMD Instinct GPUs. For ROCm migration, replace it with PyHIP (hip-python) or CuPy built against ROCm, depending on the downstream usage. This is advisory — verify which GPU Python API the rest of the codebase targets before committing to a replacement.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/cai_cupy_arrays.py
+++ examples/cai_cupy_arrays.py
@@ -4,7 +4,7 @@
import numpy as np
-import pycuda.autoinit # noqa
+# import pycuda.autoinit # noqa # PyCUDA is CUDA-only; use PyHIP or CuPy(ROCm) instead
import cupy as cp
B
PyCUDA dependency
from pycuda.compiler import SourceModule
:8

PyCUDA is NVIDIA-only and will not work on AMD Instinct GPUs. For ROCm, replace PyCUDA with PyHIP (e.g., from pyhip.compiler import SourceModule) or migrate to CuPy/hipify'd PyTorch custom extensions. This is an advisory change; verify the replacement API supports the same SourceModule semantics.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/cai_cupy_arrays.py
+++ examples/cai_cupy_arrays.py
@@ -8,1 +8,1 @@
-from pycuda.compiler import SourceModule
+# ROCm advisory: replace PyCUDA with PyHIP or CuPy equivalent
+# from pyhip.compiler import SourceModule
examples/cai_numba.py· 1
B
PyCUDA dependency
from pycuda import gpuarray
:10

PyCUDA is tightly coupled to the NVIDIA CUDA driver and runtime, making it incompatible with AMD ROCm. To migrate, this dependency must be replaced with a ROCm-compatible array library such as CuPy (which supports ROCm via HIP) or PyOpenCL targeting the ROCm OpenCL backend.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- a/examples/cai_numba.py
+++ b/examples/cai_numba.py
@@ -10,1 +10,1 @@
-from pycuda import gpuarray
+from cupy import ndarray as gpuarray # Advisory: Replace PyCUDA with CuPy for ROCm compatibility
examples/demo_cdpSimplePrint.py· 3
B
PyCUDA dependency
import pycuda.autoinit
:23

PyCUDA is a CUDA-only Python binding and will not work on ROCm. The advisory migration path is to replace it with PyHIP (hip-python) or PyOpenCL targeting the ROCm OpenCL backend, adjusting downstream API calls accordingly. This is advisory only — manual verification of all PyCUDA usage in the file is required.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/demo_cdpSimplePrint.py
+++ examples/demo_cdpSimplePrint.py
@@ -23 +23 @@
-import pycuda.autoinit
+# Advisory: replace PyCUDA with PyHIP or PyOpenCL for ROCm; e.g.:
+# import hip
B
PyCUDA dependency
import pycuda.driver as cuda
:24

PyCUDA is a CUDA-only Python binding and has no ROCm/HIP equivalent, so this import will fail on AMD Instinct GPUs. The recommended migration path is to PyOpenCL (with ROCm OpenCL runtime) or to a HIP-based Python binding such as PyHIP, though APIs differ and code changes will be required beyond a simple import swap.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/demo_cdpSimplePrint.py
+++ examples/demo_cdpSimplePrint.py
@@ -24,1 +24,1 @@
-import pycuda.driver as cuda
+# Advisory: PyCUDA is CUDA-only; migrate to PyOpenCL or PyHIP for ROCm.
+# import pyopencl as cl # or equivalent HIP Python binding
B
PyCUDA dependency
from pycuda.compiler import DynamicSourceModule
:25

PyCUDA is NVIDIA-only and has no ROCm backend; it must be replaced with PyHIP (or another ROCm-compatible Python kernel-launch layer) to run on AMD Instinct GPUs. The DynamicSourceModule API maps conceptually to PyHIP's SourceModule, but kernel-launch syntax and context handling differ, so surrounding code will also need review.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/demo_cdpSimplePrint.py
+++ examples/demo_cdpSimplePrint.py
@@ -25,1 +25,1 @@
-from pycuda.compiler import DynamicSourceModule
+# Advisory: replace PyCUDA with PyHIP for ROCm; e.g.
+# from pyhip.compiler import SourceModule
examples/demo_elementwise.py· 3
B
PyCUDA dependency
from pycuda import gpuarray
:5

PyCUDA is tightly coupled to the NVIDIA CUDA driver and runtime, making it incompatible with AMD ROCm. To migrate this dependency, you should replace PyCUDA with a ROCm-compatible alternative such as CuPy (which supports ROCm) or PyOpenCL.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/demo_elementwise.py
+++ examples/demo_elementwise.py
@@ -5,1 +5,1 @@
-from pycuda import gpuarray
+import cupy as cp # Advisory: Replace PyCUDA with CuPy for ROCm support
B
PyCUDA dependency
from pycuda.curandom import rand as curand
:6

PyCUDA is NVIDIA-only and will not work on AMD Instinct GPUs. This import should be replaced with a ROCm-compatible alternative such as PyOpenCL, pyhip, or a framework-agnostic library like CuPy (which supports ROCm) or Numba-Rocm.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/demo_elementwise.py
+++ examples/demo_elementwise.py
@@ -6,1 +6,1 @@
-from pycuda.curandom import rand as curand
+# Advisory: replace PyCUDA with a ROCm-compatible option, e.g.:
+# from cupy.random import rand as curand # CuPy with ROCm backend
B
PyCUDA dependency
from pycuda.elementwise import ElementwiseKernel
:12

PyCUDA is NVIDIA-only and will not work on AMD Instinct GPUs. The closest ROCm-compatible replacement for ElementwiseKernel is CuPy's cupy.ElementwiseKernel, which supports HIP/ROCm backends. This is advisory only — verify CuPy is installed with ROCm support and that kernel syntax is compatible.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/demo_elementwise.py
+++ examples/demo_elementwise.py
@@ -12,1 +12,1 @@
-from pycuda.elementwise import ElementwiseKernel
+import cupy as cp # ROCm-compatible; use cp.ElementwiseKernel as replacement
examples/demo_meta_codepy.py· 2
B
PyCUDA dependency
import pycuda.driver as cuda
:6

PyCUDA is NVIDIA-specific and has no ROCm equivalent; it must be replaced with PyHIP (e.g., import pyhip.driver as hip) or another ROCm-compatible runtime Python binding. This is an advisory change—API names and semantics differ, so downstream calls to cuda.* will also need review.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/demo_meta_codepy.py
+++ examples/demo_meta_codepy.py
@@ -6,1 +6,1 @@
-import pycuda.driver as cuda
+import pyhip.driver as hip # ADVISORY: PyCUDA -> PyHIP migration; review downstream cuda.* calls
B
PyCUDA dependency
from pycuda.compiler import SourceModule
:7

PyCUDA is NVIDIA-only and will not work on AMD Instinct GPUs under ROCm. The import must be replaced with a ROCm-compatible alternative such as PyHIP (which mirrors the PyCUDA API for HIP) or PyOpenCL, and any kernel source strings may need CUDA-to-HIP translation. This is advisory only — verify the chosen replacement library is installed and that downstream SourceModule usage is compatible.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- a/examples/demo_meta_codepy.py
+++ b/examples/demo_meta_codepy.py
@@ -4,7 +4,7 @@
# Advisory: PyCUDA is NVIDIA-only; migrate to PyHIP for ROCm.
-from pycuda.compiler import SourceModule
+from pyhip.compiler import SourceModule
examples/demo_meta_template.py· 2
B
PyCUDA dependency
import pycuda.driver as cuda
:6

PyCUDA is a CUDA-only Python GPU driver wrapper and has no direct ROCm equivalent. Migration requires switching to a ROCm-compatible Python GPU interface such as PyOpenCL (with AMD OpenCL runtime) or HIP Python bindings, which will require rewriting driver API calls accordingly.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/demo_meta_template.py
+++ examples/demo_meta_template.py
@@ -6,1 +6,1 @@
-import pycuda.driver as cuda
+# Advisory: PyCUDA is CUDA-only; migrate to PyOpenCL or a HIP Python binding (requires API rewrite)
B
PyCUDA dependency
from pycuda.compiler import SourceModule
:7

PyCUDA is NVIDIA-specific and will not work on AMD Instinct GPUs. The migration path is to replace PyCUDA's SourceModule with PyOpenCL (using ROCm's OpenCL backend) or with CuPy's RawKernel (CuPy has ROCm support), or to rewrite the kernel compilation using HIP directly. This is a manual migration step that requires updating both the import and the surrounding kernel compilation/launch logic.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/demo_meta_template.py
+++ examples/demo_meta_template.py
@@ -4,7 +4,8 @@
# ADVISORY: PyCUDA is not supported on ROCm. Replace with PyOpenCL or CuPy.
- from pycuda.compiler import SourceModule
+ # from pycuda.compiler import SourceModule # CUDA-only, not ROCm-compatible
+ # Consider: import pyopencl as cl (with ROCm OpenCL backend)
+ # Or: import cupy as cp (CuPy supports ROCm)
examples/demo_struct.py· 2
B
PyCUDA dependency
import pycuda.driver as cuda
:8

PyCUDA is NVIDIA-only and will not work on AMD Instinct GPUs under ROCm. It must be replaced with a ROCm-compatible Python GPU runtime such as PyHIP or migrated to a backend-agnostic framework like CuPy (with ROCm builds) or PyTorch-ROCm. The import and all downstream API calls (e.g., cuda.Event, cuda.mem_alloc) will need corresponding PyHIP equivalents.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/demo_struct.py
+++ examples/demo_struct.py
@@ -8,1 +8,1 @@
-import pycuda.driver as cuda
+import pyhip.driver as hip # advisory: PyHIP mirrors PyCUDA's API; rename downstream 'cuda.' references to 'hip.'
B
PyCUDA dependency
from pycuda.compiler import SourceModule
:9

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/demo.py· 4
B
PyCUDA dependency
import pycuda.autoinit # noqa
:6

PyCUDA is NVIDIA-specific and will not work on AMD Instinct GPUs under ROCm. It must be replaced with a ROCm-compatible alternative such as PyHIP (pyhip) or migrated to a HIP-based kernel launch path. Note that PyHIP's API surface differs from PyCUDA, so surrounding kernel compilation and launch code will also need adaptation.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/demo.py
+++ examples/demo.py
@@ -6,1 +6,1 @@
-import pycuda.autoinit # noqa
+import pyhip.autoinit # noqa # Advisory: replace PyCUDA with PyHIP for ROCm; API differences apply
B
PyCUDA dependency
import pycuda.driver as cuda
:7

PyCUDA is NVIDIA-only and will not work on AMD Instinct GPUs. It should be replaced with PyOpenCL, CuPy (with ROCm build), or PyHIP for ROCm compatibility. This is an advisory recommendation; manual testing is required.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/demo.py
+++ examples/demo.py
@@ -7,1 +7,1 @@
-import pycuda.driver as cuda
+import cupy as cuda # ADVISORY: replace PyCUDA with CuPy (ROCm build) or PyHIP
B
PyCUDA dependency
from pycuda.compiler import SourceModule
:8

PyCUDA is NVIDIA-specific and will not work on AMD Instinct GPUs under ROCm. The SourceModule import and all PyCUDA GPU array/compiler usage must be replaced with a ROCm-compatible alternative such as PyHIP (hip-python) or PyOpenCL, or the kernel logic should be ported to a HIP C++ extension.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/demo.py
+++ examples/demo.py
- from pycuda.compiler import SourceModule
+ # Advisory: PyCUDA is not supported on ROCm. Replace with PyHIP or a HIP C++ extension.
+ # from hip_python.compiler import SourceModule # requires hip-python package
B
PyCUDA dependency
from pycuda import gpuarray
:45

PyCUDA is NVIDIA-only and will not work on AMD Instinct GPUs. Replace it with a ROCm-compatible alternative such as PyTorch (torch) or PyOpenCL, or use CuPy with ROCm support. The gpuarray import should be swapped for the equivalent array abstraction in the chosen replacement library.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

Suggested change · advisory
--- examples/demo.py
+++ examples/demo.py
@@ -45,1 +45,1 @@
-from pycuda import gpuarray
+# Advisory: PyCUDA is NVIDIA-only; replace with a ROCm-compatible alternative
+# e.g., torch (import torch), cupy (import cupy), or pyopencl
+import torch # or: import cupy as cp; use cp.array / torch.tensor for GPU arrays
examples/dump_properties.py· 1
B
PyCUDA dependency
import pycuda.driver as drv
:3

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/fill_gpu_with_nans.py· 2
B
PyCUDA dependency
import pycuda.driver as cuda
:5

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:6

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/2d_fft.py· 2
B
PyCUDA dependency
import pycuda.driver as drv
:10

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.gpuarray as garray
:11

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/arithmetic_example.py· 3
B
PyCUDA dependency
import pycuda.driver as cuda
:7

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:8

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:9

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/c++_function_templates.py· 2
B
PyCUDA dependency
from pycuda import gpuarray
:6

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:7

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/convolution.py· 2
B
PyCUDA dependency
import pycuda.driver as cuda
:51

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:52

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/demo_complex.py· 2
B
PyCUDA dependency
from pycuda import gpuarray
:6

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.elementwise import ElementwiseKernel
:17

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/demo_meta_cgen.py· 2
B
PyCUDA dependency
import pycuda.driver as cuda
:7

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:8

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/demo_meta_matrixmul_cheetah.py· 1
B
PyCUDA dependency
from pycuda import compiler, gpuarray
:18

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/distance_element_wise3d.py· 3
B
PyCUDA dependency
import pycuda.driver as cuda
:8

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:9

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.elementwise import ElementwiseKernel as Elementwise
:10

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/game_of_life.py· 2
B
PyCUDA dependency
from pycuda import gpuarray
:17

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:18

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/gl_interop.py· 5
B
PyCUDA dependency
import pycuda.driver as cuda_driver
:17

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.gl as cuda_gl
:18

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:19

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.gl
:280

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.gl.autoinit
:281

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/gpu_scalar_mult.py· 8
B
PyCUDA dependency
import pycuda.driver as drv
:6

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:7

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import context_dependent_memoize
:8

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.elementwise import get_linear_combination_kernel
:12

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand
:21

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.elementwise import ScalarArg, VectorArg, get_elwise_module
:48

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import dtype_to_ctype
:49

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand
:91

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/kernel_concurrency.py· 3
B
PyCUDA dependency
import pycuda.autoinit
:10

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.driver as drv
:11

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:12

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/light_field_3d_viewer.py· 3
B
PyCUDA dependency
import pycuda.autoinit
:33

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.compiler
:34

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.driver as cuda
:35

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/mandelbrot_interactive.py· 2
B
PyCUDA dependency
import pycuda.driver as drv
:38

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:39

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/mandelbrot.py· 3
B
PyCUDA dependency
import pycuda.driver as drv
:18

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:19

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.elementwise import ElementwiseKernel
:26

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/manhattan_distance_for_2D_array.py· 2
B
PyCUDA dependency
import pycuda.driver as cuda
:6

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:7

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/matrix_transpose.py· 7
B
PyCUDA dependency
import pycuda.autoinit
:11

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.driver as cuda
:12

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:13

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:14

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import context_dependent_memoize
:15

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand
:142

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand
:162

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/matrixmul_simple.py· 1
B
PyCUDA dependency
from pycuda import compiler, gpuarray
:14

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/matrixmul_tiled.py· 1
B
PyCUDA dependency
from pycuda import compiler, gpuarray
:18

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/measure_gpuarray_speed_random.py· 3
B
PyCUDA dependency
import pycuda.driver as drv
:11

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import curandom
:12

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:16

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/multiple_threads.py· 2
B
PyCUDA dependency
import pycuda.driver as cuda
:10

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:11

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/plot_random_data.py· 1
B
PyCUDA dependency
from pycuda import curandom
:5

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/rotate.py· 3
B
PyCUDA dependency
import pycuda.autoinit
:14

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.compiler
:15

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.driver as cuda
:16

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/select_to_list.py· 3
B
PyCUDA dependency
import pycuda.driver as cuda
:10

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:11

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:12

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/simple_rgb2gray.py· 2
B
PyCUDA dependency
import pycuda.driver as drv
:11

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:12

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/simple_speed_test.py· 7
B
PyCUDA dependency
import pycuda.autoinit
:48

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.cumath
:49

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.driver as drv
:50

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.tools
:51

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:52

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:53

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.elementwise import ElementwiseKernel
:54

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/sobel_filter.py· 4
B
PyCUDA dependency
import pycuda
:36

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.driver as cuda_driver
:37

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.gl as cuda_gl
:38

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:41

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/sparse_solve.py· 6
B
PyCUDA dependency
import pycuda.driver as drv
:7

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:8

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import DeviceMemoryPool, PageLockedMemoryPool
:20

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.sparse.packeted import PacketedSpMV
:30

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.sparse.operator import DiagonalPreconditioner
:34

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.sparse.cg import solve_pkt_with_cg
:42

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/threads_and_blocks.py· 1
B
PyCUDA dependency
from pycuda.compiler import SourceModule
:4

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/thrust_interop.py· 1
B
PyCUDA dependency
from pycuda import gpuarray
:9

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/from-wiki/using_printf.py· 1
B
PyCUDA dependency
from pycuda.compiler import SourceModule
:4

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

examples/hello_gpu.py· 2
B
PyCUDA dependency
import pycuda.driver as drv
:5

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:6

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/autoinit.py· 3
B
PyCUDA dependency
import pycuda.driver as cuda
:5

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import make_default_context
:11

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import clear_context_caches
:23

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/autoprimaryctx.py· 3
B
PyCUDA dependency
import pycuda.driver as cuda
:5

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import make_default_context
:11

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import clear_context_caches
:29

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/characterize.py· 5
B
PyCUDA dependency
from pycuda.tools import context_dependent_memoize
:5

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.driver import Context
:18

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.driver import Context
:24

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:31

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:46

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/cumath.py· 2
B
PyCUDA dependency
from pycuda import elementwise, gpuarray
:7

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.driver import Stream
:8

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/curandom.py· 11
B
PyCUDA dependency
import pycuda.compiler
:7

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.driver as drv
:8

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.gpuarray as array
:9

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.elementwise import get_elwise_kernel
:189

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.gpuarray import GPUArray
:190

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda._driver as _curand # used to be separate module
:262

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.characterize import has_double_support
:424

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.characterize import sizeof
:570

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.characterize import has_stack
:775

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.characterize import has_stack
:1096

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.characterize import has_stack
:1195

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/debug.py· 1
B
PyCUDA dependency
import pycuda.driver
:7

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/driver.py· 9
B
PyCUDA dependency
from pycuda._driver import * # noqa
:75

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda._pvt_struct import pack
:242

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda._pvt_struct import calcsize
:355

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda._pvt_struct import pack
:384

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda._pvt_struct import pack
:416

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda._pvt_struct import pack
:460

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda._pvt_struct import pack
:580

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda._pvt_struct import pack
:596

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda._pvt_struct import pack
:638

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/elementwise.py· 7
B
PyCUDA dependency
from pycuda.tools import ScalarArg, VectorArg, context_dependent_memoize, dtype_to_ctype
:37

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:50

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:100

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import parse_c_arg
:162

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.gpuarray import splay
:279

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.elementwise import ScalarArg, VectorArg, get_elwise_module
:418

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import dtype_to_ctype
:419

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/gl/__init__.py· 1
B
PyCUDA dependency
import pycuda._driver as _drv
:3

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/gl/autoinit.py· 3
B
PyCUDA dependency
import pycuda.driver as cuda
:5

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.gl as cudagl
:6

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import make_default_context
:12

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/gpuarray.py· 15
B
PyCUDA dependency
import pycuda.driver as drv
:17

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import elementwise
:18

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.characterize import has_double_support
:19

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compyte.array import (
:20

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.characterize import platform_bits
:53

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import get_or_register_dtype
:64

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import DeviceData
:113

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.reduction import get_sum_kernel
:2060

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.reduction import get_any_kernel
:2067

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.reduction import get_all_kernel
:2074

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.reduction import get_subset_sum_kernel
:2081

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.reduction import get_dot_kernel
:2088

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.reduction import get_subset_dot_kernel
:2097

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.reduction import get_minmax_kernel
:2105

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.reduction import get_subset_minmax_kernel
:2121

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/reduction.py· 3
B
PyCUDA dependency
from pycuda.tools import context_dependent_memoize, dtype_to_ctype
:63

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:79

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import get_arg_type
:202

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/scan.py· 5
B
PyCUDA dependency
import pycuda._mymako as mako
:30

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import driver, gpuarray
:31

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda._cluda import CLUDA_PREAMBLE
:32

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:33

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import dtype_to_ctype
:34

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/sparse/cg.py· 8
B
PyCUDA dependency
import pycuda.driver as drv
:7

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:8

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.sparse.inner import AsyncInnerProduct
:9

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.sparse.operator import IdentityOperator
:19

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.elementwise import get_linear_combination_kernel
:30

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.elementwise import get_elwise_kernel
:75

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import dtype_to_ctype
:76

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.gpuarray import _get_common_dtype
:90

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/sparse/coordinate.py· 6
B
PyCUDA dependency
import pycuda.driver as drv
:7

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:8

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:9

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import DeviceData
:160

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import dtype_to_ctype
:185

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import dtype_to_ctype
:208

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/sparse/inner.py· 2
B
PyCUDA dependency
import pycuda.driver as drv
:5

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:6

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/sparse/packeted.py· 4
B
PyCUDA dependency
from pycuda import gpuarray
:7

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:8

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import DeviceData
:89

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import dtype_to_ctype
:320

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/sparse/pkt_build.py· 2
B
PyCUDA dependency
from pycuda import gpuarray
:5

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.sparse.pkt_build_cython import (
:79

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pycuda/tools.py· 10
B
PyCUDA dependency
import pycuda._driver as _drv
:34

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.driver as cuda
:35

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compyte.dtypes import ( # noqa: F401
:36

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.driver import Device
:155

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.driver as drv
:270

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compyte.dtypes import parse_c_arg_backend
:435

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.driver
:495

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import clear_context_caches
:509

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

A
.cuda() tensor/module move
return mark_test.cuda(f)
:522

.cuda() on tensors and modules is honoured by ROCm PyTorch and moves data to the AMD GPU. No change needed.

RecommendNo change required on ROCm PyTorch.

B
PyCUDA dependency
import pycuda.driver as cuda
:526

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

pyproject.toml· 4
B
PyCUDA dependency
"pycuda",
:59

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
"pycuda/sparse/coordinate.py" = [
:81

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
"pycuda/driver.py" = [
:84

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
"pycuda/curandom.py" = [
:88

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

test/test_cumath.py· 5
B
PyCUDA dependency
import pycuda.driver as drv # noqa
:7

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import cumath, gpuarray
:8

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import mark_cuda_test
:9

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.driver import Stream
:155

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.autoinit # noqa
:221

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

test/test_driver.py· 15
B
PyCUDA dependency
import pycuda.driver as drv
:13

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:14

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:15

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import dtype_to_ctype, mark_cuda_test
:16

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:144

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import blas
:175

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:678

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import bitlog2
:693

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import DeviceMemoryPool
:704

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import DeviceMemoryPool, bitlog2
:718

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import DynamicSourceModule
:1030

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import DEFAULT_NVCC_FLAGS
:1064

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import DynamicModule
:1083

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.driver # noqa
:1172

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.autoinit # noqa
:1177

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

test/test_gpuarray.py· 56
B
PyCUDA dependency
import pycuda.driver as drv
:10

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:11

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.characterize import has_double_support
:12

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.compiler import SourceModule
:13

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import init_cuda_context_fixture
:14

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:314

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import get_curand_version
:325

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import (
:334

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import (
:343

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import MRG32k3aRandomNumberGenerator
:357

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import get_curand_version
:399

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import (
:408

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import (
:417

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import MRG32k3aRandomNumberGenerator
:431

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import get_curand_version
:476

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import (
:485

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import (
:494

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import MRG32k3aRandomNumberGenerator
:508

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:653

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.elementwise import ElementwiseKernel
:658

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.elementwise import ElementwiseKernel
:672

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:743

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:771

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:833

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:851

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.tools
:895

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:947

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:961

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:979

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:998

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:999

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:1025

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:1037

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:1045

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:1081

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:1103

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.scan import ExclusiveScanKernel, InclusiveScanKernel
:1170

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:1338

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import register_dtype
:1343

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.reduction import ReductionKernel
:1347

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:1367

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.reduction import ReductionKernel
:1372

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.tools
:1393

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.tools
:1417

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:1443

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:1475

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:1481

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:1490

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:1498

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:1510

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:1548

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import cumath
:1676

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.elementwise import ElementwiseKernel
:1714

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.reduction import ReductionKernel
:1733

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.scan import InclusiveScanKernel
:1751

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
import pycuda.autoinit # noqa
:1785

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

test/undistributed/elwise-perf.py· 1
B
PyCUDA dependency
from pycuda import gpuarray
:12

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

test/undistributed/measure_gpuarray_speed.py· 3
B
PyCUDA dependency
import pycuda.driver as drv
:6

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda import gpuarray
:10

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.tools import bitlog2
:18

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

test/undistributed/reduction-perf.py· 3
B
PyCUDA dependency
import pycuda.driver as cuda
:5

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.curandom import rand as curand
:19

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

B
PyCUDA dependency
from pycuda.reduction import get_dot_kernel
:26

PyCUDA binds the CUDA driver API. The equivalent on AMD is PyHIP / HIP Python; kernels port mechanically via HIPify.

RecommendPort to HIP Python (PyHIP) and run kernels through hipify-perl.

setup.py· 1
A
Device string "cuda"
Libraries("CUDADRV", ["cuda"]),
:93

The literal device string "cuda" resolves to the active AMD GPU on ROCm PyTorch. .to("cuda") / device="cuda" need no edits.

RecommendNo change required on ROCm PyTorch.