Skip to content
/ oneMKL Public
forked from oneapi-src/oneMKL

oneAPI Math Kernel Library (oneMKL) Interfaces

License

Notifications You must be signed in to change notification settings

Rbiessy/oneMKL

 
 

Repository files navigation

oneAPI Math Kernel Library (oneMKL) Interfaces

oneAPI logo

oneMKL Interfaces is an open-source implementation of the oneMKL Data Parallel C++ (DPC++) interface according to the oneMKL specification. It works with multiple devices (backends) using device-specific libraries underneath.

oneMKL is part of oneAPI.

User Application oneMKL Layer Third-Party Library Hardware Backend
oneMKL interface oneMKL selector Intel(R) oneAPI Math Kernel Library for x86 CPU x86 CPU
Intel(R) oneAPI Math Kernel Library for Intel GPU Intel GPU
NVIDIA cuBLAS for NVIDIA GPU NVIDIA GPU
NVIDIA cuSOLVER for NVIDIA GPU NVIDIA GPU
NVIDIA cuRAND for NVIDIA GPU NVIDIA GPU
NVIDIA cuFFT for NVIDIA GPU NVIDIA GPU
NETLIB LAPACK for x86 CPU x86 CPU
AMD rocBLAS for AMD GPU AMD GPU
AMD rocSOLVER for AMD GPU AMD GPU
AMD rocRAND for AMD GPU AMD GPU
AMD rocFFT for AMD GPU AMD GPU
portBLAS x86 CPU, Intel GPU, NVIDIA GPU, AMD GPU
portFFT x86 CPU, Intel GPU, NVIDIA GPU, AMD GPU

Table of Contents


Support and Requirements

Supported Usage Models:

Host API

There are two oneMKL selector layer implementations:

  • Run-time dispatching: The application is linked with the oneMKL library and the required backend is loaded at run-time based on device vendor (all libraries should be dynamic).

Example of app.cpp with run-time dispatching:

#include "oneapi/mkl.hpp"

...
cpu_dev = sycl::device(sycl::cpu_selector());
gpu_dev = sycl::device(sycl::gpu_selector());

sycl::queue cpu_queue(cpu_dev);
sycl::queue gpu_queue(gpu_dev);

oneapi::mkl::blas::column_major::gemm(cpu_queue, transA, transB, m, ...);
oneapi::mkl::blas::column_major::gemm(gpu_queue, transA, transB, m, ...);

How to build an application with run-time dispatching:

if OS is Linux, use icpx compiler. If OS is Windows, use icx compiler. Linux example:

$> icpx -fsycl –I$ONEMKL/include app.cpp
$> icpx -fsycl app.o –L$ONEMKL/lib –lonemkl
  • Compile-time dispatching: The application uses a templated backend selector API where the template parameters specify the required backends and third-party libraries and the application is linked with the required oneMKL backend wrapper libraries (libraries can be static or dynamic).

Example of app.cpp with compile-time dispatching:

#include "oneapi/mkl.hpp"

...
cpu_dev = sycl::device(sycl::cpu_selector());
gpu_dev = sycl::device(sycl::gpu_selector());

sycl::queue cpu_queue(cpu_dev);
sycl::queue gpu_queue(gpu_dev);

oneapi::mkl::backend_selector<oneapi::mkl::backend::mklcpu> cpu_selector(cpu_queue);

oneapi::mkl::blas::column_major::gemm(cpu_selector, transA, transB, m, ...);
oneapi::mkl::blas::column_major::gemm(oneapi::mkl::backend_selector<oneapi::mkl::backend::cublas> {gpu_queue}, transA, transB, m, ...);

How to build an application with compile-time dispatching:

$> clang++ -fsycl –I$ONEMKL/include app.cpp
$> clang++ -fsycl app.o –L$ONEMKL/lib –lonemkl_blas_mklcpu –lonemkl_blas_cublas

Refer to Selecting a Compiler for the choice between icpx/icx and clang++ compilers.

Device API

Header-based and backend-independent Device API can be called within sycl kernel or work from Host code (device-rng-usage-model-example). Currently, the following domains support the Device API:

  • RNG. To use RNG Device API functionality it's required to include oneapi/mkl/rng/device.hpp header file.

Supported Configurations:

Supported domains: BLAS, LAPACK, RNG, DFT, SPARSE_BLAS

Linux*

Domain Backend Library Supported Link Type Supported Compiler
BLAS x86 CPU Intel(R) oneAPI Math Kernel Library Dynamic, Static DPC++, LLVM*, hipSYCL
Intel GPU Dynamic, Static DPC++
NVIDIA GPU NVIDIA cuBLAS Dynamic, Static LLVM*, hipSYCL
x86 CPU NETLIB LAPACK Dynamic, Static DPC++, LLVM*, hipSYCL
AMD GPU AMD rocBLAS Dynamic, Static LLVM*, hipSYCL
x86 CPU, Intel GPU, NVIDIA GPU, AMD GPU portBLAS Dynamic, Static DPC++, LLVM*
LAPACK x86 CPU Intel(R) oneAPI Math Kernel Library Dynamic, Static DPC++, LLVM*
Intel GPU Dynamic, Static DPC++
NVIDIA GPU NVIDIA cuSOLVER Dynamic, Static LLVM*
AMD GPU AMD rocSOLVER Dynamic, Static LLVM*
RNG x86 CPU Intel(R) oneAPI Math Kernel Library Dynamic, Static DPC++, LLVM*, hipSYCL
Intel GPU Dynamic, Static DPC++
NVIDIA GPU NVIDIA cuRAND Dynamic, Static LLVM*, hipSYCL
AMD GPU AMD rocRAND Dynamic, Static LLVM*, hipSYCL
DFT Intel GPU Intel(R) oneAPI Math Kernel Library Dynamic, Static DPC++
x86 CPU Dynamic, Static DPC++
NVIDIA GPU NVIDIA cuFFT Dynamic, Static DPC++
AMD GPU AMD rocFFT Dynamic, Static DPC++
x86 CPU, Intel GPU, NVIDIA GPU, AMD GPU portFFT (limited API support) Dynamic, Static DPC++
SPARSE_BLAS Intel GPU Intel(R) oneAPI Math Kernel Library Dynamic, Static DPC++
x86 CPU Dynamic, Static DPC++

Windows*

Domain Backend Library Supported Link Type Supported Compiler
BLAS x86 CPU Intel(R) oneAPI Math Kernel Library Dynamic, Static DPC++, LLVM*
Intel GPU Dynamic, Static DPC++
x86 CPU NETLIB LAPACK Dynamic, Static DPC++, LLVM*
LAPACK x86 CPU Intel(R) oneAPI Math Kernel Library Dynamic, Static DPC++, LLVM*
Intel GPU Dynamic, Static DPC++
RNG x86 CPU Intel(R) oneAPI Math Kernel Library Dynamic, Static DPC++, LLVM*
Intel GPU Dynamic, Static DPC++

* LLVM - Intel project for LLVM* technology with support for NVIDIA CUDA


Hardware Platform Support

  • CPU
    • Intel Atom(R) Processors
    • Intel(R) Core(TM) Processor Family
    • Intel(R) Xeon(R) Processor Family
  • Accelerators
    • Intel(R) Arc(TM) A-Series Graphics
    • Intel(R) Data Center GPU Max Series
    • NVIDIA(R) A100 (Linux* only)
    • AMD(R) GPUs see here tested on AMD Vega 20 (gfx906)

Supported Operating Systems

Linux*

Operating System CPU Host/Target Integrated Graphics from Intel (Intel GPU) NVIDIA GPU
Ubuntu 18.04.3, 19.04 18.04.3, 19.10 18.04.3, 20.04
SUSE Linux Enterprise Server* 15 Not supported Not supported
Red Hat Enterprise Linux* (RHEL*) 8 Not supported Not supported
Linux* kernel N/A 4.11 or higher N/A

Windows*

Operating System CPU Host/Target Integrated Graphics from Intel (Intel GPU)
Microsoft Windows* 10 (64-bit version only) 10 (64-bit version only)
Microsoft Windows* Server 2016, 2019 Not supported

Software Requirements

What should I download?

General:

Using Conan Using CMake Directly
Functional Testing Build Only Documentation
Linux* : GNU* GCC 5.1 or higher
Windows* : MSVS* 2017 or MSVS* 2019 (version 16.5 or newer)
Python 3.6 or higher CMake
Ninja (optional)
Conan C++ package manager GNU* FORTRAN Compiler - Sphinx
NETLIB LAPACK - -

Hardware and OS Specific:

Operating System Device Package Installed by Conan
Linux*/Windows* x86 CPU Intel(R) oneAPI DPC++ Compiler
or
Intel project for LLVM* technology
No
Intel(R) oneAPI Math Kernel Library Yes
Intel GPU Intel(R) oneAPI DPC++ Compiler No
Intel GPU driver No
Intel(R) oneAPI Math Kernel Library Yes
Linux* only NVIDIA GPU Intel project for LLVM* technology
or
hipSYCL with CUDA backend and dependencies
No
AMD GPU Intel project for LLVM* technology
or
hipSYCL with ROCm backend and dependencies
No

If Building with Conan, above packages marked as "No" must be installed manually.

If Building with CMake, above packages must be installed manually.

Notice for Use of Conan Package Manager

LEGAL NOTICE: By downloading and using this container or script as applicable (the "Software Package") and the included software or software made available for download, you agree to the terms and conditions of the software license agreements for the Software Package, which may also include notices, disclaimers, or license terms for third party software (together, the "Agreements") included in this README file.

If the Software Package is installed through a silent install, your download and use of the Software Package indicates your acceptance of the Agreements.

Product and Version Information:

Product Supported Version Installed by Conan Conan Package Source Package Install Location on Linux* License
Python 3.6 or higher No N/A Pre-installed or Installed by user PSF
Conan C++ Package Manager 1.24 or higher No N/A Installed by user MIT
CMake 3.13 or higher Yes
(3.15 or higher)
conan-center ~/.conan/data or $CONAN_USER_HOME/.conan/data The OSI-approved BSD 3-clause License
Ninja 1.10.0 Yes conan-center ~/.conan/data or $CONAN_USER_HOME/.conan/data Apache License v2.0
GNU* FORTRAN Compiler 7.4.0 or higher Yes apt /usr/bin GNU General Public License, version 3
Intel(R) oneAPI DPC++ Compiler latest No N/A Installed by user End User License Agreement for the Intel(R) Software Development Products
hipSYCL later than 2cfa530 No N/A Installed by user BSD-2-Clause License
Intel project for LLVM* technology binary for x86 CPU Daily builds No N/A Installed by user Apache License v2
Intel project for LLVM* technology source for NVIDIA and AMD GPUs Daily source releases No N/A Installed by user Apache License v2
Intel(R) oneAPI Math Kernel Library latest Yes apt /opt/intel/inteloneapi/mkl Intel Simplified Software License
NVIDIA CUDA SDK 10.2 No N/A Installed by user End User License Agreement
AMD rocBLAS 4.5 No N/A Installed by user AMD License
AMD rocRAND 5.1.0 No N/A Installed by user AMD License
AMD rocSOLVER 5.0.0 No N/A Installed by user AMD License
AMD rocFFT rocm-5.4.3 No N/A Installed by user AMD License
NETLIB LAPACK 3.7.1 Yes conan-community ~/.conan/data or $CONAN_USER_HOME/.conan/data BSD like license
Sphinx 2.4.4 Yes pip ~/.local/bin (or similar user local directory) BSD License
portBLAS 0.1 No N/A Installed by user Apache License v2.0
portFFT 0.1 No N/A Installed by user or automatically by cmake Apache License v2.0

conan-center: https://api.bintray.com/conan/conan/conan-center

conan-community: https://api.bintray.com/conan/conan-community/conan


Documentation


Contributing

See CONTRIBUTING for more information.


License

Distributed under the Apache license 2.0. See [LICENSE](LICENSE) for more

information.


FAQs

oneMKL

  1. What is the difference between the following oneMKL items?

Answer:

  • The oneAPI Specification for oneMKL defines the DPC++ interfaces for performance math library functions. The oneMKL specification can evolve faster and more frequently than implementations of the specification.

  • The oneAPI Math Kernel Library (oneMKL) Interfaces Project is an open source implementation of the specification. The project goal is to demonstrate how the DPC++ interfaces documented in the oneMKL specification can be implemented for any math library and work for any target hardware. While the implementation provided here may not yet be the full implementation of the specification, the goal is to build it out over time. We encourage the community to contribute to this project and help to extend support to multiple hardware targets and other math libraries.

  • The Intel(R) oneAPI Math Kernel Library (oneMKL) product is the Intel product implementation of the specification (with DPC++ interfaces) as well as similar functionality with C and Fortran interfaces, and is provided as part of Intel® oneAPI Base Toolkit. It is highly optimized for Intel CPU and Intel GPU hardware.

  1. I'm trying to use oneMKL Interfaces in my project using FetchContent, but I keep running into ONEMKL::SYCL::SYCL target was not found problem when I try to build the project. What should I do?

Answer:

Make sure you set the compiler when you configure your project. E.g. cmake -Bbuild . -DCMAKE_CXX_COMPILER=icpx.

  1. I'm trying to use oneMKL Interfaces in my project using find_package(oneMKL). I set oneMKL/oneTBB and Compiler environment first, then I built and installed oneMKL Interfaces, and finally I tried to build my project using installed oneMKL Interfaces (e.g. like this cmake -Bbuild -GNinja -DCMAKE_CXX_COMPILER=icpx -DoneMKL_ROOT=<path_to_installed_oneMKL_interfaces> .) and I noticed that cmake includes installed oneMKL Interfaces headers as a system include which ends up as a lower priority than the installed oneMKL package includes which I set before for building oneMKL Interfaces. As a result, I get conflicts between oneMKL and installed oneMKL Interfaces headers. What should I do?

Answer:

Having installed oneMKL Interfaces headers as -I instead on system includes (as -isystem) helps to resolve this problem. We use INTERFACE_INCLUDE_DIRECTORIES to add paths to installed oneMKL Interfaces headers (check oneMKLTargets.cmake in lib/cmake to find it). It's a known limitation that INTERFACE_INCLUDE_DIRECTORIES puts headers paths as system headers. To avoid that:

  • Option 1: Use CMake >=3.25. In this case oneMKL Interfaces will be built with EXPORT_NO_SYSTEM property set to true and you won't see the issue.
  • Option 2: If you use CMake < 3.25, set PROPERTIES NO_SYSTEM_FROM_IMPORTED true for your target. E.g: set_target_properties(test PROPERTIES NO_SYSTEM_FROM_IMPORTED true).

Conan

  1. I am behind a proxy. How can Conan download dependencies from external network?

    • ~/.conan/conan.conf has a [proxies] section where you can add the list of proxies. For details refer to Conan proxy settings.
  2. I get an error while installing packages via APT through Conan.

    dpkg: warning: failed to open configuration file '~/.dpkg.cfg' for reading: Permission denied
    Setting up intel-oneapi-mkl-devel (2021.1-408.beta07) ...
    E: Sub-process /usr/bin/dpkg returned an error code (1)
    
    • Although your user session has permissions to install packages via sudo apt, it does not have permissions to update debian package configuration, which throws an error code 1, causing a failure in conan install command.
    • The package is most likely installed correctly and can be verified by:
      1. Running the conan install command again.
      2. Checking /opt/intel/inteloneapi for mkl and/or tbb directories.

About

oneAPI Math Kernel Library (oneMKL) Interfaces

Resources

License

Code of conduct

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C++ 98.0%
  • CMake 1.7%
  • Python 0.3%