nvidia-cuda-toolkit 12.0.1-2 source package in Ubuntu

 GPUDirect Storage (GDS) enables a direct data path for direct memory access
 (DMA) transfers between GPU memory and storage, which avoids a bounce buffer
 through the CPU.
 .
 This package contains the GDS tools.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 ACCINJ is the OpenACC internal library for profiling.
 .
 This package contains the 64-bit ACCINJ runtime library.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 The cuBLAS library is an implementation of BLAS (Basic Linear Algebra
 Subprograms) on top of the NVIDIA CUDA runtime. It allows the user to access
 the computational resources of NVIDIA Graphics Processing Unit (GPU), but
 does not auto-parallelize across multiple GPUs.
 .
 This package contains the cuBLAS runtime library.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 The cuBLASLt library is a lightweight GEMM library with a flexible API and
 tensor core support for INT8 inputs and FP16 CGEMM split-complex matrix
 multiplication.
 .
 This package contains the cuBLASLt runtime library.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 This package contains the CUDA Runtime API library for high-level CUDA
 programming, on top of the CUDA Driver API.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 The FFT is a divide-and-conquer algorithm for efficiently computing discrete
 Fourier transforms of complex or real-valued data sets. It is one of the most
 important and widely used numerical algorithms in computational physics and
 general signal processing. The cuFFT library provides a simple interface for
 computing FFTs on an NVIDIA GPU, which allows users to quickly leverage the
 floating-point power and parallelism of the GPU in a highly optimized and
 tested FFT library.
 .
 This package contains the cuFFT runtime library.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 The FFT is a divide-and-conquer algorithm for efficiently computing discrete
 Fourier transforms of complex or real-valued data sets. It is one of the most
 important and widely used numerical algorithms in computational physics and
 general signal processing. The cuFFT library provides a simple interface for
 computing FFTs on an NVIDIA GPU, which allows users to quickly leverage the
 floating-point power and parallelism of the GPU in a highly optimized and
 tested FFT library.
 .
 This package contains the cuFFTW runtime library.

 GPUDirect Storage (GDS) enables a direct data path for direct memory access
 (DMA) transfers between GPU memory and storage, which avoids a bounce buffer
 through the CPU.
 .
 This package contains the development files: headers and libraries.

 GPUDirect Storage (GDS) enables a direct data path for direct memory access
 (DMA) transfers between GPU memory and storage, which avoids a bounce buffer
 through the CPU.
 .
 This package contains the cuFile RDMA runtime library.

 GPUDirect Storage (GDS) enables a direct data path for direct memory access
 (DMA) transfers between GPU memory and storage, which avoids a bounce buffer
 through the CPU.
 .
 This package contains the cuFile runtime library.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 CUINJ is the CUDA internal library for profiling.
 .
 This package contains the 64-bit CUINJ runtime library.

 The CUDA Profiler Tools Interface (CUPTI) enables the creation of
 profiling and tracing tools that target CUDA applications. CUPTI
 provides a set of APIs targeted at ISVs creating profilers and other
 performance optimization tools. The CUPTI APIs are not intended to be
 used by developers in their CUDA applications.
 .
 This package contains the development files: headers and libraries.

 The CUDA Profiler Tools Interface (CUPTI) enables the creation of
 profiling and tracing tools that target CUDA applications. CUPTI
 provides a set of APIs targeted at ISVs creating profilers and other
 performance optimization tools. The CUPTI APIs are not intended to be
 used by developers in their CUDA applications.
 .
 This package contains the documentation and examples.

 The CUDA Profiler Tools Interface (CUPTI) enables the creation of
 profiling and tracing tools that target CUDA applications. CUPTI
 provides a set of APIs targeted at ISVs creating profilers and other
 performance optimization tools. The CUPTI APIs are not intended to be
 used by developers in their CUDA applications.
 .
 This package contains the runtime library.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 The cuRAND library provides facilities that focus on the simple and efficient
 generation of high-quality pseudorandom and quasirandom numbers.
 A pseudorandom sequence of numbers satisfies most of the statistical
 properties of a truly random sequence but is generated by a deterministic
 algorithm. A quasirandom sequence of n-dimensional points is generated by a
 deterministic algorithm designed to fill an n-dimensional space evenly.
 .
 This package contains the cuRAND runtime library.

 The cuSOLVER library contains LAPACK-like functions in dense and sparse
 linear algebra, including linear solver, least-square solver and eigenvalue
 solver.
 .
 This package contains the cuSOLVER runtime library.

 The cuSOLVER library contains LAPACK-like functions in dense and sparse
 linear algebra, including linear solver, least-square solver and eigenvalue
 solver.
 .
 This package contains the cuSOLVERmg runtime library.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 The cuSPARSE library contains a set of basic linear algebra subroutines used
 for handling sparse matrices. It is implemented on top of the NVIDIA CUDA
 runtime and is designed to be called from C and C++. The library routines can
 be classified into four categories:
  * Level 1: operations between a vector in sparse format and a vector in dense
 format
  * Level 2: operations between a matrix in sparse format and a vector in dense
 format
  * Level 3: operations between a matrix in sparse format and a set of vectors
 in dense format
  * Conversion: operations that allow conversion between different matrix
 formats
 .
 This package contains the cuSPARSE runtime library.

 NVIDIA NPP is a library of functions for performing CUDA accelerated
 processing. The initial set offunctionality in the library focuses on imaging
 and video processing and is widely applicable for developers in these areas.
 NPP will evolve over time to encompass more of the compute heavy tasks in a
 variety of problem domains. The NPP library is written to maximize
 flexibility, while maintaining high performance.
 .
 This package contains the NVIDIA Performance Primitives core runtime library.

 NVIDIA NPP is a library of functions for performing CUDA accelerated
 processing.
 .
 This package contains the NVIDIA Performance Primitives runtime library for
 Image Arithmetic and Logic operations, which is a sub-library of nppi.

 NVIDIA NPP is a library of functions for performing CUDA accelerated
 processing.
 .
 This package contains the NVIDIA Performance Primitives runtime library for
 Image Color and sampling Conversion, which is a sub-library of nppi.

 NVIDIA NPP is a library of functions for performing CUDA accelerated
 processing.
 .
 This package contains the NVIDIA Performance Primitives runtime library for
 Image Data Exchange and Initialization, which is a sub-library of nppi.

 NVIDIA NPP is a library of functions for performing CUDA accelerated
 processing.
 .
 This package contains the NVIDIA Performance Primitives runtime library for
 Image Filters, which is a sub-library of nppi.

 NVIDIA NPP is a library of functions for performing CUDA accelerated
 processing.
 .
 This package contains the NVIDIA Performance Primitives runtime library for
 Image Geometry transforms, which is a sub-library of nppi.

 NVIDIA NPP is a library of functions for performing CUDA accelerated
 processing.
 .
 This package contains the NVIDIA Performance Primitives runtime library for
 Image Morphological operations, which is a sub-library of nppi.

 NVIDIA NPP is a library of functions for performing CUDA accelerated
 processing.
 .
 This package contains the NVIDIA Performance Primitives runtime library for
 Image Statistics and Linear Transformation, which is a sub-library of nppi.

 NVIDIA NPP is a library of functions for performing CUDA accelerated
 processing.
 .
 This package contains the NVIDIA Performance Primitives runtime library for
 Image Support, which is a sub-library of nppi.

 NVIDIA NPP is a library of functions for performing CUDA accelerated
 processing.
 .
 This package contains the NVIDIA Performance Primitives runtime library for
 Image Threshold and Compare, which is a sub-library of nppi.

 NVIDIA NPP is a library of functions for performing CUDA accelerated
 processing. The initial set offunctionality in the library focuses on imaging
 and video processing and is widely applicable for developers in these areas.
 NPP will evolve over time to encompass more of the compute heavy tasks in a
 variety of problem domains. The NPP library is written to maximize
 flexibility, while maintaining high performance.
 .
 This package contains the NVIDIA Performance Primitives runtime library for
 signal processing.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 The NVBLAS Library is a GPU-accelerated Library that implements BLAS (Basic
 Linear Algebra Subprograms). It can accelerate most BLAS Level-3 routines by
 dynamically routing BLAS calls to one or more NVIDIA GPUs present in the
 system, when the characteristics of the call make it to speedup on a GPU.

 The NVIDIA Management Library (NVML) provides a monitoring and management API.
 It provides a direct access to the queries and commands exposed via nvidia-smi.
 .
 This package contains the header file and depends on the driver-provided
 library.

 NVIDIA's CUDA Compiler (NVCC) may apply Just-in-Time (JIT)
 Link-Time-Optimization (LTO) to improve performance of CUDA kernels
 comprised of multiple separate compilation units.
 .
 This package contains the NVIDIA Compiler JIT LTO runtime library.

 The nvJPEG 1.0 library provides high-performance, GPU accelerated JPEG
 decoding functionality for image formats commonly used in deep learning and
 hyperscale multimedia applications. The library offers single and batched JPEG
 decoding capabilities which efficiently utilize the available GPU resources
 for optimum performance; and the flexibility for users to manage the memory
 allocation needed for decoding.

 CUDA Runtime Compilation library (nvrtc) provides an API to compile
 CUDA-C++ device source code at runtime.
 .
 The resulting compiled PTX can be launched on a GPU using the CUDA
 Driver API.
 .
 This package contains the NVRTC Builtins library.

 CUDA Runtime Compilation library (nvrtc) provides an API to compile
 CUDA-C++ device source code at runtime.
 .
 The resulting compiled PTX can be launched on a GPU using the CUDA
 Driver API.
 .
 This package contains the NVRTC library.

 The NVIDIA Tools Extension SDK (NVTX) is a C-based API for marking events and
 ranges in your applications. Applications which integrate NVTX can use Nsight
 to capture and visualize these events and ranges.
 .
 This package contains the NVIDIA Tools Extension runtime library.

 NVIDIA's CUDA Compiler (NVCC) is based on the widely used LLVM open source
 compiler infrastructure.
 .
 The NVVM library is used by NVCC to compile CUDA binary code to run on NVIDIA
 GPUs.
 .
 This package contains the NVIDIA NVVM runtime library.

 NVIDIA Nsight Compute is an interactive kernel profiler for CUDA applications.
 It provides detailed performance metrics and API debugging via a user
 interface and command line tool. In addition, its baseline feature allows
 users to compare results within the tool. Nsight Compute provides a
 customizable and data-driven user interface and metric collection and can be
 extended with analysis scripts for post-processing results.

 NVIDIA Nsight Compute is an interactive kernel profiler for CUDA applications.
 It provides detailed performance metrics and API debugging via a user
 interface and command line tool. In addition, its baseline feature allows
 users to compare results within the tool. Nsight Compute provides a
 customizable and data-driven user interface and metric collection and can be
 extended with analysis scripts for post-processing results.
 .
 This package contains the target specific libraries.

 NVIDIA Nsight Systems is a system-wide performance analysis tool designed to
 visualize an application’s algorithms, help you identify the largest
 opportunities to optimize, and tune to scale efficiently across any quantity
 or size of CPUs and GPUs; from large server to smallest SoCs.

 NVIDIA Nsight Systems is a system-wide performance analysis tool designed to
 visualize an application’s algorithms, help you identify the largest
 opportunities to optimize, and tune to scale efficiently across any quantity
 or size of CPUs and GPUs; from large server to smallest SoCs.
 .
 This package contains the target specific libraries.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 This package contains the development files: headers and libraries.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 This package contains the cuda-gdb debugger.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 This package contains the nvcc compiler and other tools needed for building
 CUDA applications.
 .
 Running CUDA applications requires a supported NVIDIA GPU and the NVIDIA
 driver kernel module.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 OpenCL (Open Computing Language) is a multi-vendor open standard for
 general-purpose parallel programming of heterogeneous systems that include
 CPUs, GPUs and other processors.
 .
 Note that CUDA documentation is no longer bundled with CUDA toolkit releases.
 Visit https://docs.nvidia.com/cuda for the latest documentation on CUDA.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 This package provides the /usr/bin/cuda-gcc, /usr/bin/cuda-g++ symlinks to
 simplify building packages that need to be built with a CUDA-compatible
 compiler.
 .
 This package ensures a deterministic nvcc/gcc combination is used by default
 and is therefore recommended to be used as build dependency for all source
 packages needing nvcc for building.

 GPUDirect Storage (GDS) enables a direct data path for direct memory access
 (DMA) transfers between GPU memory and storage, which avoids a bounce buffer
 through the CPU.
 .
 This package builds the nvidia-fs.ko kernel driver.

 OpenCL (Open Computing Language) is a multi-vendor open standard for
 general-purpose parallel programming of heterogeneous systems that include
 CPUs, GPUs and other processors.
 .
 This metapackage provides the development files: headers and libraries.

 The Compute Unified Device Architecture (CUDA) enables NVIDIA
 graphics processing units (GPUs) to be used for massively parallel
 general purpose computation.
 .
 OpenCL (Open Computing Language) is a multi-vendor open standard for
 general-purpose parallel programming of heterogeneous systems that include
 CPUs, GPUs and other processors.
 .
 This package contains the nvprof profiler.

 The NVIDIA Visual Profiler is a cross-platform performance profiling tool
 that delivers developers vital feedback for optimizing CUDA C/C++ and OpenCL
 applications.