# cupy | CuPy | Knowledge Base | Instagit

NumPy & SciPy for GPU

GitHub Stars: 11.8k

Repository: https://github.com/cupy/cupy

---

## Articles

### [How to Contribute to the CuPy Project: A Complete Developer Guide](/cupy/cupy/how-to-contribute-cupy)

Learn how to contribute to the CuPy project with this developer guide. Fork the repository, set up your environment, follow coding standards, write tests, and submit a pull request.

- Tags: how-to-guide
- Published: 2026-06-30

### [How to Build CuPy from Source: A Complete Developer’s Guide](/cupy/cupy/build-cupy-from-source)

Build CuPy from source with this developer's guide. Learn the essential steps including Python, Cython, toolkit requirements, cloning, and installation for optimal GPU performance.

- Tags: how-to-guide
- Published: 2026-06-30

### [How to Use CuPy for Linear Algebra Operations: A GPU-Accelerated Guide](/cupy/cupy/cupy-linear-algebra)

Accelerate linear algebra operations with CuPy. This guide shows how to leverage GPU power for solutions like matrix inversion and least-squares using a NumPy compatible API.

- Tags: how-to-guide
- Published: 2026-06-30

### [What Are CuPy Kernels and How to Write Custom Kernels](/cupy/cupy/cupy-custom-kernels)

Learn what CuPy kernels are and how to write custom CUDA C/C++ functions for GPU acceleration in Python. Unlock direct CUDA programming with Cupy RawKernel.

- Tags: deep-dive
- Published: 2026-06-30

### [How to Perform Basic Array Operations in CuPy: A Complete GPU-Accelerated Guide](/cupy/cupy/cupy-basic-array-operations)

Learn to perform basic array operations in CuPy with this GPU-accelerated guide. Leverage CuPy's NumPy-compatible API for faster numerical computing on NVIDIA GPUs.

- Tags: tutorial
- Published: 2026-06-30

### [Using CUDA Profiler with CuPy for Performance Analysis](/cupy/cupy/using-cuda-profiler-with-cupy-for-performance-analysis)

Analyze CuPy performance with the CUDA Profiler. Use cupyx.profiler for Pythonic wrappers, benchmarking, and GPU kernel timing.

- Tags: performance
- Published: 2026-06-29

### [CuPy GPU-Accelerated Linear Algebra Operations: A Complete Guide to NumPy-Compatible GPU Computing](/cupy/cupy/cupy-gpu-accelerated-linear-algebra-operations)

Accelerate linear algebra with CuPy, the NumPy replacement for NVIDIA and AMD GPUs. Leverage cuBLAS and cuSOLVER for faster array computations. Explore our complete guide.

- Tags: tutorial
- Published: 2026-06-29

### [How to Use CuPy linalg for GPU Linear Algebra: A Practical Guide to cuSOLVER and cuBLAS](/cupy/cupy/how-to-use-cupy-linalg-for-gpu-linear-algebra)

Accelerate your linear algebra with CuPy linalg. This guide shows how to use cuSOLVER and cuBLAS on NVIDIA GPUs for matrix operations, solving linear systems, and decompositions with cupy.ndarray.

- Tags: how-to-guide
- Published: 2026-06-29

### [How to Debug CuPy GPU Memory Leaks: A Complete Guide](/cupy/cupy/how-to-debug-cupy-gpu-memory-leaks)

Debug CuPy GPU memory leaks effectively. Monitor memory pools, trace allocations with DebugPrintHook, and free blocks to fix reference issues. Get your GPU code running smoothly.

- Tags: how-to-guide
- Published: 2026-06-29

### [CuPy FFT on GPU for Fast Fourier Transforms: Architecture, Plan Caching, and Performance Tuning](/cupy/cupy/cupy-fft-on-gpu-for-fast-fourier-transforms)

Accelerate Fast Fourier Transforms with CuPy FFT on GPU. Leverage NVIDIA cuFFT and AMD rocFFT via a NumPy-compatible API for zero-copy, batched GPU acceleration. Optimize performance with plan caching.

- Tags: deep-dive
- Published: 2026-06-29

### [How to Use the CuPy FFT Module for GPU‑Accelerated Transforms](/cupy/cupy/how-to-use-cupy-fft-module-for-gpu-accelerated-transforms)

Accelerate your Fast Fourier Transforms with CuPy's FFT module. Leverage GPU power for NumPy-compatible, high-performance transforms with automatic caching and multi-GPU support.

- Tags: how-to-guide
- Published: 2026-06-29

### [How to Optimize CuPy Memory Management Using Memory Pools: Tuning, Limits, and Custom Allocators](/cupy/cupy/how-to-optimize-cupy-memory-management-using-memory-pools)

Optimize CuPy memory management with memory pools. Learn about tuning, limits, and custom allocators to reduce GPU memory overhead and improve performance.

- Tags: how-to-guide
- Published: 2026-06-29

### [CuPy Low-Level GPU Programming with RawKernels: Architecture, Examples, and Best Practices](/cupy/cupy/cupy-low-level-gpu-programming-with-rawkernels)

Explore CuPy low-level GPU programming with RawKernels. Write, compile, and launch custom CUDA kernels directly from Python for maximum control and performance.

- Tags: deep-dive
- Published: 2026-06-29

### [How to Implement Custom CUDA Kernels with CuPy RawKernels](/cupy/cupy/how-to-implement-custom-cuda-kernels-with-cupy-rawkernels)

Learn to implement custom CUDA kernels in Python effortlessly with CuPy RawKernels. Compile and launch your own kernels for accelerated GPU computing.

- Tags: how-to-guide
- Published: 2026-06-29

### [How to Optimize CuPy Performance with CUDA Streams: 5 Proven Techniques](/cupy/cupy/how-to-optimize-cupy-performance-with-cuda-streams)

Boost CuPy performance with CUDA streams. Overlap computation and data transfer using non-blocking streams, priorities, and CUDA graphs for maximum efficiency. Optimize your GPU code today.

- Tags: performance
- Published: 2026-06-29

### [How to Use CuPy CUDA Streams for Asynchronous GPU Operations: A Complete Guide](/cupy/cupy/how-to-use-cupy-cuda-streams-for-asynchronous-gpu-operations)

Master CuPy CUDA streams for asynchronous GPU operations. Learn to queue operations and maximize GPU utilization with concurrent kernel execution and memory transfers. A complete guide.

- Tags: how-to-guide
- Published: 2026-06-29