AMD Core Math Library: Difference between revisions

Revision as of 21:28, 15 December 2008

Latest revision as of 20:05, 27 June 2025

(41 intermediate revisions by 31 users not shown)

Line 1:

'''AMD Core Math Library''' ('''ACML''') is aan end-of-life<ref>[https://developer.amd.com/tools-and-sdks/archive/acml-product-features/ AMD Core Math Library Product Features]{{Dead link|date=March 2018}}</ref> [[software development]] library released by [[AMD]], replaced by many open source libraries, including AMD libm 4.0. This library provides useful mathematical routines optimized for AMD processors.

{{orphan|date=December 2008}}

{{Expand|date=January 2007}}

The successor to ACML is the '''AMD Optimizing CPU Libraries''' ('''AOCL'''), a set of mostly [[open source]] libraries compiled for [[AMD64]] processors. It includes the open source [[BLIS (software)|BLIS]], [[LAPACK#Implementations|libFLAME]], [[ScaLAPACK]], [[Fastest Fourier Transform in the West|FFTW]], and AOCL-Sparse, plus the original closed-source AMD LibM, [[memcpy]], and RNG.<ref>{{cite web |title=AMD Optimizing CPU Libraries (AOCL) |url=https://developer.amd.com/amd-aocl/ |website=AMD}}</ref>

'''AMD Core Math Library''' (ACML) is a [[software development]] library released by [[AMD]]. This library provides useful mathematical routines optimized for AMD processors.

<!-- updated by AMD on 6/1/2007 -->

Originally developed in 2002 for use in [[HPC]] ([[high performance computing|high-performance computing]]) scientific computing, ACML allows nearly optimal use of [[AMD]] [[Opteron]] processors in compute-intensive applications.

==Features==

ACML consists of the following main components:<ref name="AMD-Core-Math-Library-dotcom">{{Cite web |url=http://developer.amd.com/tools-and-sdks/cpu-development/amd-core-math-library-acml/ |title=ACML |access-date=2014-10-19 |archive-url=https://web.archive.org/web/20141015020116/http://developer.amd.com/tools-and-sdks/cpu-development/amd-core-math-library-acml/ |archive-date=2014-10-15 |url-status=dead }}</ref>

ACML consists of the following main components:

* A full implementation of Level 1, 2 and 3 [[Basic Linear Algebra SubroutinesSubprograms]] (BLAS), with key routines optimizedoptimizations for high performance on AMD Opteron processors.

<ref name="AMD-Core-Math-Library-dotcom">{{cite web

* A full suite of [[Linear Algebra]] ([[LAPACK]]) routines.

|url=http://developer.amd.com/acml.jsp

* A comprehensive suite of [[Fast Fourier Transformstransform]] ([[FFT]]sFFTs) in both single-, double-, single-complex and double-complex data types.

|title=AMD Core Math Library (ACML)

* Fast scalar, vector, and array math transcendental library routines optimized for high performance on AMD Opteron processors.

|publisher=[[AMD]]

* [[Random Number GeneratorsGenerator]]s in both single- and double-precision.

|date=[[April 20]], [[2005]]

|accessdate = 2007-06-01

==Supported platforms==

|}}</ref>

AMD offers pre-compiled binaries for [[Linux]], [[Solaris (operating system)|Solaris]], and [[Microsoft Windows|Windows]] available for download. Supported compilers include [[GNU Fortran]], [[Intel Fortran Compiler]], [[Microsoft Visual Studio]], [[Numerical Algorithms Group|NAG]], [[PathScale]], [[PGI compiler]], and [[Sun Studio (software)|Sun Studio]].<ref>[https://developer.amd.com/tools-and-sdks/archive/acml-downloads-resources/ ACML Downloads & Resources]{{Dead link|date=November 2018}}</ref>

* A full implementation of Level 1, 2 and 3 Basic Linear Algebra Subroutines (BLAS), with key routines optimized for high performance on AMD Opteron processors.

* A full suite of [[Linear Algebra]] ([[LAPACK]]) routines. As well as taking advantage of the highly-tuned [[BLAS]] kernels, a key set of [[LAPACK]] routines has been further optimized to achieve considerably higher performance than standard LAPACK implementations.

==License==

* A comprehensive suite of Fast Fourier Transforms ([[FFT]]s) in both single-, double-, single-complex and double-complex data types.

ACML has a proprietary [[freeware]] license. The library is distributed in binary form free of charge, but cannot be freely redistributed.<ref>[http://developer.amd.com/tools-and-sdks/cpu-development/amd-core-math-library-acml/acml-redistribution-agreements/ Redistribution Agreements] {{Webarchive|url=https://web.archive.org/web/20141015020107/http://developer.amd.com/tools-and-sdks/cpu-development/amd-core-math-library-acml/acml-redistribution-agreements/ |date=2014-10-15 }}, ACML Redistribution Agreements</ref><ref>[http://amd-dev.wpengine.netdna-cdn.com/wordpress/media/2012/10/ACML-EULA.pdf ACML EULA], End-user license agreement</ref>

* Fast scalar, vector, and array math transcendental library routines optimized for high performance on AMD Opteron processors.

* Random Number Generators in both single- and double-precision.

==See also==

* [[GPUOpen]] - Open-source software suite for visual effects, HPC, and GPGPU

* Framewave - formerly the AMD Performance Library

* [[Open64]] - AMD has an Open64 compiler distribution that can be used with ACML

* [[Math Kernel Library]] (MKL)

==References==

{{Reflist}}

<div class="references-small" style="-moz-column-count: 3; column-count: 3;">

<references/>

</div>

==External links==

*[https://web.archive.org/web/20081010124022/http://ati.amd.com/technology/streamcomputing/Stream_Computing_User_Guide.pdf Streaming Computing (User Guide)]

*[http://developer.amd.com/acml.jsp AMD webpage for ACML developers]

*Replacement: [https://www.amd.com/en/developer/aocl/libm.html AMD Math Library (LibM)], [https://github.com/amd/aocl-libm-ose github]

*[http://ati.amd.com/technology/streamcomputing/Stream_Computing_User_Guide.pdf Streaming Computing (User Guide)]

{{Numerical linear algebra}}

[[Category:Advanced Micro Devices products]]

[[ruCategory:AMD Core Math Librarysoftware]]

[[Category:Numerical libraries]]