guix-patches
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[bug#64400] [PATCH 5/5] gnu: Add ocaml-owl.


From: Liliana Marie Prikler
Subject: [bug#64400] [PATCH 5/5] gnu: Add ocaml-owl.
Date: Tue, 11 Jul 2023 20:37:49 +0200
User-agent: Evolution 3.46.4

Am Samstag, dem 01.07.2023 um 17:43 +0000 schrieb Anders Ågren Thuné:
> * gnu/packages/ocaml.scm (ocaml-owl): New variable.
> ---
>  gnu/packages/ocaml.scm | 28 ++++++++++++++++++++++++++++
>  1 file changed, 28 insertions(+)
> 
> diff --git a/gnu/packages/ocaml.scm b/gnu/packages/ocaml.scm
> index 83077addb0..f118c1289e 100644
> --- a/gnu/packages/ocaml.scm
> +++ b/gnu/packages/ocaml.scm
> @@ -2797,6 +2797,34 @@ (define-public ocaml-toml
>  OCaml primitive types are also supplied.")
>      (license license:lgpl3)))
>  
> +(define-public ocaml-owl
> +  (package
> +    (name "ocaml-owl")
> +    (version "1.1")
> +    (source (origin
> +              (method git-fetch)
> +              (uri (git-reference
> +                    (url "https://github.com/owlbarn/owl";)
> +                    (commit version)))
> +              (file-name (git-file-name name version))
> +              (sha256
> +               (base32
> +               
> "08jvgf1fd7d28cxxjifx4ikmwcbfbiyw0sivw3xy4vdzvbyc9xw9"))))
> +    (build-system dune-build-system)
> +    (propagated-inputs (list openblas zlib ocaml-ctypes ocaml-npy
> ocaml-compiler-libs))
I'm pretty sure some of those inputs shouldn't be propagated.
> +    (native-inputs (list ocaml-alcotest ocaml-base ocaml-stdio))
> +    (home-page "https://github.com/owlbarn/owl";)
> +    (synopsis "OCaml Scientific and Engineering Computing")
> +    (description
> +     "Owl is an OCaml numerical library.  It supports N-dimensional
> +arrays, both dense and sparse matrix operations, linear algebra,
> +regressions, fast Fourier transforms, and many advanced mathematical
> +and statistical functions (such as Markov chain Monte Carlo
> methods).
> +Recently, Owl has implemented algorithmic differentiation which
> +simplifies developing machine learning and neural network
> +algorithms.")
> +    (license license:expat)))
This package sadly fails to build on aarch64 and probably also has
issues on other architectures, as it assumes some build flags specific
to x86_64 (see the CI build logs).  These should be stripped in a
patch, snippet or pre-build phase.  If these flags are there for
performance reasons, you should instead mark the package as tunable.

Cheers





reply via email to

[Prev in Thread] Current Thread [Next in Thread]