Commit 0231466f authored by Ludovic Courtès's avatar Ludovic Courtès

Change URLs to "hpc.guix.info".

* hpcweb-config.scm, posts/*.md, posts/*.sxml: Replace
"guix-hpc.bordeaux.inria.fr" with "hpc.guix.info".
parent 91fc737f
;; Configuration of hpcguix-web for use on
;; "https://guix-hpc.bordeaux.inria.fr/browse".
;; Configuration of hpcguix-web for use on "https://hpc.guix.info/browse".
(use-modules (hpcweb-configuration))
......
......@@ -5,9 +5,9 @@ slug: guix-hpc-activity-report-2018
---
_This document is also available as
[PDF](https://guix-hpc.bordeaux.inria.fr/static/doc/activity-report-2018.pdf)
[PDF](https://hpc.guix.info/static/doc/activity-report-2018.pdf)
([printable
booklet](https://guix-hpc.bordeaux.inria.fr/static/doc/activity-report-2018-booklet.pdf))._
booklet](https://hpc.guix.info/static/doc/activity-report-2018-booklet.pdf))._
Guix-HPC is a collaborative effort to bring reproducible software
deployment to scientific workflows and high-performance computing (HPC).
......@@ -46,7 +46,7 @@ these areas.
# Reproducible Scientific Workflows
![Lab book.](https://guix-hpc.bordeaux.inria.fr/static/images/blog/lab-book.svg)
![Lab book.](https://hpc.guix.info/static/images/blog/lab-book.svg)
Research heavily depends on computational results, which in turn depends
on the ability to reproduce software environments. As key scientific
......@@ -68,7 +68,7 @@ provides information about the currently used revision. Along with the
new _channels_ facility, which allows users to obtain software packages
from third-party repositories, this offers a transparent way to
replicate a Guix setup, as [explained in the release notes of version
0.16.0](https://guix-hpc.bordeaux.inria.fr/blog/2018/12/hpc-reproducible-research-in-guix-0-16-0/).
0.16.0](https://hpc.guix.info/blog/2018/12/hpc-reproducible-research-in-guix-0-16-0/).
Better yet, Guix allows mixing software packages coming from different
Guix revisions through a new mechanism called _inferiors_.
......@@ -121,7 +121,7 @@ dependent on software provided in a Docker application bundle, which
is convenient but ignores the problem of software provenance.
As demonstrated by Pjotr Prins in a [blog
post](https://guix-hpc.bordeaux.inria.fr/blog/2019/01/creating-a-reproducible-workflow-with-cwl/),
post](https://hpc.guix.info/blog/2019/01/creating-a-reproducible-workflow-with-cwl/),
GNU Guix can be used to build reproducible software environments
incrementally or declaratively to prepare the context in which an
existing Common Workflow Language (CWL) workflow is to be executed.
......@@ -167,7 +167,7 @@ Since the Guix-HPC effort was started in September 2017, around 3,000
packages were added to Guix itself; of these many had to do with linear
algebra, computational fluid dynamics, bioinformatics, and statistics,
as reported in the HPC release notes on the [Guix-HPC
blog](https://guix-hpc.bordeaux.inria.fr/blog).
blog](https://hpc.guix.info/blog).
In addition, our institutes have developed their own package
collections, sometimes as a staging area before packages are reviewed
......@@ -226,14 +226,14 @@ Inc](http://www.cray.com/), to better support this architecture.
# Cluster Usage
![High-performance computing cluster.](https://guix-hpc.bordeaux.inria.fr/static/images/blog/hpc-cluster.svg)
![High-performance computing cluster.](https://hpc.guix.info/static/images/blog/hpc-cluster.svg)
GNU Guix has been deployed on clusters at our research institutes and in
other places. One of our first task has been to simplify the deployment
and installation of Guix on clusters, providing new features for
distributed setups to its build daemon and command-line tools, and
[documenting the installation process for system
administrators](https://guix-hpc.bordeaux.inria.fr/blog/2017/11/installing-guix-on-a-cluster/).
administrators](https://hpc.guix.info/blog/2017/11/installing-guix-on-a-cluster/).
This is the option we recommend because it gives cluster users a lot of
flexibility: each user can install, upgrade, and remove packages at
will, create software environments on the fly with `guix environment`,
......@@ -244,7 +244,7 @@ installed, and we wanted to offer interoperability with those. As
so-called “container-based solutions” like Docker and Singularity are
being deployed on clusters, we developed `guix pack`, a [tool that can
create “container
images”](https://guix-hpc.bordeaux.inria.fr/blog/2017/10/using-guix-without-being-root/).
images”](https://hpc.guix.info/blog/2017/10/using-guix-without-being-root/).
In this setup, users use `guix pack` on their laptop to generate an
image that contains precisely the software environment they need, and
then send it over to the cluster to run their application. `guix pack`
......@@ -259,13 +259,13 @@ To help cluster users get started with Guix,
developed at the UMC Utrecht, provides a web interface that allows users
to search for software packages and to learn how to install them. A
public instance is visible [on the Guix-HPC web
site](https://guix-hpc.bordeaux.inria.fr/browse). Hpcguix-web is
site](https://hpc.guix.info/browse). Hpcguix-web is
customizable and the instance running in Utrecht provides users with
additional information such as how to use the batch scheduler.
# Outreach and User Support
![Containers are like smoothies.](https://guix-hpc.bordeaux.inria.fr/static/images/blog/container-smoothie.png)
![Containers are like smoothies.](https://hpc.guix.info/static/images/blog/container-smoothie.png)
One aspect of our work on Guix-HPC is to “spread the word” about the
importance of being able not just to replicate software environments,
......@@ -290,7 +290,7 @@ in peer-reviewed conferences:
GigaScience and was presented at the [International Conference on
Genomics (ICG-13)](http://www.icg-13.org/) where it was [awarded
2nd Runner Up in the GigaScience Prize Track
](https://guix-hpc.bordeaux.inria.fr/blog/2019/01/pigx-paper-awarded-at-the-international-conference-on-genomics-icg-13/).
](https://hpc.guix.info/blog/2019/01/pigx-paper-awarded-at-the-international-conference-on-genomics-icg-13/).
Altuna Alkalin, research team leader at the Max Delbrück Center (MDC),
wrote an article entitled [_Scientific Data Analysis Pipelines and
......@@ -300,7 +300,7 @@ compares existing tools to achieve software environment reproducibility:
application bundles (also referred to as “containers”), CONDA, and Guix.
We have published [12 articles on this
blog](https://guix-hpc.bordeaux.inria.fr/blog) touching a range of
blog](https://hpc.guix.info/blog) touching a range of
technical topics: running Guix without being root, on the performance of
pre-built binaries, creating reproducible workflows with CWL or PiGx,
and more.
......@@ -403,7 +403,7 @@ work more closely with initiatives such as
to investigate ways Guix could help improve their reviewing workflows.
We are very much open to new ideas and we’d like to [hear from
you](https://guix-hpc.bordeaux.inria.fr/about)!
you](https://hpc.guix.info/about)!
# Credits
......
......@@ -111,16 +111,16 @@ Problem solved?
It turns out that not all software packages, especially scientific
software, use these techniques. Some do—for instance,
[OpenBLAS](https://guix-hpc.bordeaux.inria.fr/package/openblas) supports
[OpenBLAS](https://hpc.guix.info/package/openblas) supports
run-time selection when compiled with `DYNAMIC_ARCH=1`—but many don’t.
For example, [FFTW](https://guix-hpc.bordeaux.inria.fr/package/fftw)
For example, [FFTW](https://hpc.guix.info/package/fftw)
insists on being compiled with
[`-mtune=native`](https://gcc.gnu.org/onlinedocs/gcc-7.3.0/gcc/x86-Options.html#index-mtune-15)
and provides [configuration
options](http://fftw.org/fftw3_doc/Installation-on-Unix.html#Installation-on-Unix)
to statically select CPU optimizations (*Update:* FFTW 3.3.7+ [can select
optimized routines at run time](https://lists.gnu.org/archive/html/guix-devel/2018-04/msg00091.html));
[ATLAS](https://guix-hpc.bordeaux.inria.fr/package/atlas) optimizes
[ATLAS](https://hpc.guix.info/package/atlas) optimizes
itself for the CPU it is being built on. We can always say that the
“right” solution would be to “fix” these packages upstream so that they
use run-time selection, but how do we handle these _today_ in Guix?
......@@ -131,7 +131,7 @@ simply don’t distribute pre-built binaries for it. Instead, running
`guix package -i atlas` unconditionally builds it locally, as upstream
authors intended.
For FFTW, [BLIS](https://guix-hpc.bordeaux.inria.fr/package/blis), and
For FFTW, [BLIS](https://hpc.guix.info/package/blis), and
other packages where optimizations are selected at configure-time, we
simply build the generic version, like Debian and others do. This is
the most unsatisfactory situation: we have portable binaries at the cost
......@@ -154,8 +154,8 @@ corresponds to their machine.
Having optimized package variants is nice, but how can users take
advantage of them? For instance, the
[`julia`](https://guix-hpc.bordeaux.inria.fr/package/julia) and
[`octave`](https://guix-hpc.bordeaux.inria.fr/package/octave) packages
[`julia`](https://hpc.guix.info/package/julia) and
[`octave`](https://hpc.guix.info/package/octave) packages
depend on the generic (unoptimized) `fftw` package—this allows us to
distribute pre-built binaries. What if you want Octave to use the
AVX2-optimized FFTW?
......
......@@ -5,7 +5,7 @@ tags: cluster installation sysadmin
---
Previously we
[discussed](https://guix-hpc.bordeaux.inria.fr/blog/2017/10/using-guix-without-being-root)
[discussed](https://hpc.guix.info/blog/2017/10/using-guix-without-being-root)
ways to use Guix-produced packages on a cluster where Guix is not
installed. In this post we look at how a cluster sysadmin can install
Guix for system-wide use, and discuss the various tradeoffs.
......
......@@ -13,7 +13,7 @@ describes workflows that can run on specifically supported backend hardware syst
principle, this combination of tools should be enough to provide
reproducibility with provenance and improved security.
![A CWL provenance graph](https://guix-hpc.bordeaux.inria.fr/static/images/blog/cwl-provenance-graph.png)
![A CWL provenance graph](https://hpc.guix.info/static/images/blog/cwl-provenance-graph.png)
*Note: this work was mostly executed during the Biohackathon 2018 in
Matsue, Japan.*
......@@ -58,7 +58,7 @@ different institutes), and can locate and serve files in a
(reasonably) scalable manner. IPFS is content-addressable (more on this
below) and allows for deduplication, local caching, and swarm-like
software downloads over the network. IPFS is free software and
[available through GNU Guix](https://guix-hpc.bordeaux.inria.fr/package/go-ipfs).
[available through GNU Guix](https://hpc.guix.info/package/go-ipfs).
## GNU Guix
......@@ -141,12 +141,12 @@ An [install script](https://git.savannah.gnu.org/cgit/guix.git/plain/etc/guix-in
option. More installation instructions are here at [GitLab](https://gitlab.com/pjotrp/guix-notes/blob/master/INSTALL.org) and here at
[GNU.org](https://www.gnu.org/software/guix/manual/html_node/Binary-Installation.html). The short of it is that the Guix (daemon) needs to be installed
as root, but runs with user land privileges. For those who cannot get
access to root there are [workarounds](https://guix-hpc.bordeaux.inria.fr/blog/2017/10/using-guix-without-being-root/), including the use of
access to root there are [workarounds](https://hpc.guix.info/blog/2017/10/using-guix-without-being-root/), including the use of
Docker. Ricardo Wurmus describes how MDC deploys GNU Guix on their [HPC](https://guix.mdc-berlin.de/documentation.html)
and [here](https://elephly.net/posts/2015-04-17-gnu-guix.html) (essentially use one build host and copy files to the
rest). For HPC we typically use a build host that has privileges, but
all other HPC nodes simply mount one directory under /gnu/store using
a network mount. More HPC blogs on this topic [here](https://guix-hpc.bordeaux.inria.fr/blog/). If you don't think
a network mount. More HPC blogs on this topic [here](https://hpc.guix.info/blog/). If you don't think
it can be done on HPC, think again: [Compute Canada](https://archive.fosdem.org/2018/schedule/event/computecanada/) deploys Nix on
their HPCs (over 120,000 cores). And if you can do Nix, you can do
Guix. Same principles.
......@@ -185,7 +185,7 @@ packages on Guix trunk needed to be updated, including [`python-rdflib`
and `python-setuptools`](https://gitlab.com/genenetwork/guix/commit/1204258ca29bba9966934507287eb320a64afe8f). This leads to the following dependency graph
for cwltool generated by Guix itself:
![Dependencies of cwtool.](https://guix-hpc.bordeaux.inria.fr/static/images/blog/cwltool-dependencies.svg)
![Dependencies of cwtool.](https://hpc.guix.info/static/images/blog/cwltool-dependencies.svg)
If Guix is correctly installed most packages get downloaded and
installed as binaries. Guix only builds packages when it cannot find
......@@ -706,7 +706,7 @@ spirit of functional programming Guix does not allow for side-effects.
This figure shows the dependency graph for running the workflow, and
includes fastqc, trimmomatic-jar, bwa, ipfs-go, and cwltool itself.
![Dependency graph of the tools needed to run the workflow](https://guix-hpc.bordeaux.inria.fr/static/images/blog/workflow-example.svg)
![Dependency graph of the tools needed to run the workflow](https://hpc.guix.info/static/images/blog/workflow-example.svg)
This is a huge graph (but not abnormally so). GNU Guix keeps track of all these
dependencies (here we show versions, but can also show the hash values) and can
......@@ -721,7 +721,7 @@ and this is what GNU Guix provides.
## CWL provenance graph
![A CWL provenance graph](https://guix-hpc.bordeaux.inria.fr/static/images/blog/cwl-provenance-graph.png)
![A CWL provenance graph](https://hpc.guix.info/static/images/blog/cwl-provenance-graph.png)
The figure was created by adding the workflow to the CWL viewer online
(simply by pasting the github link). See
......@@ -822,7 +822,7 @@ which writes out a container that can be uploaded to docker hub or
some other repo without using Docker. See also
<https://github.com/genenetwork/guix-cwl> where we achieved exactly that. Note that a
recent version of Docker itself is [packaged in
GNU Guix](https://guix-hpc.bordeaux.inria.fr/package/docker).
GNU Guix](https://hpc.guix.info/package/docker).
# Future work
......
......@@ -47,7 +47,7 @@ us summarize the salient features of Guix.
[Singularity](https://www.sylabs.io/singularity/).
With [`guix pull` and `guix
describe`](https://guix-hpc.bordeaux.inria.fr/blog/2018/12/hpc-reproducible-research-in-guix-0-16-0/),
describe`](https://hpc.guix.info/blog/2018/12/hpc-reproducible-research-in-guix-0-16-0/),
you can reproduce the exact same environment on a different machine,
or at a different point in time. Our [collaboration with Software
Heritage](https://www.softwareheritage.org/2019/04/18/software-heritage-and-gnu-guix-join-forces-to-enable-long-term-reproducibility/)
......@@ -57,12 +57,12 @@ us summarize the salient features of Guix.
- *Packages.* No need to wait for cluster sysadmins to provide the
software you’re interested in as
[modules](http://modules.sourceforge.net/): almost [10,000
packages](https://guix-hpc.bordeaux.inria.fr/browse) are one `guix
packages](https://hpc.guix.info/browse) are one `guix
install` command away, maintained openly by [an active
community](https://www.openhub.net/p/gnuguix/), including lots of
scientific software packages—from
[OpenFOAM](https://guix-hpc.bordeaux.inria.fr/package/openfoam) and
[TensorFlow](https://guix-hpc.bordeaux.inria.fr/package/tensorflow),
[OpenFOAM](https://hpc.guix.info/package/openfoam) and
[TensorFlow](https://hpc.guix.info/package/tensorflow),
to linear algebra software, and to statistics and bioinformatics
packages.
......@@ -89,7 +89,7 @@ us summarize the salient features of Guix.
- *Container images.* You have a Guix-managed package collection and
would like to use it on a cluster that lacks Guix? Create an
application bundle with [`guix
pack`](https://guix-hpc.bordeaux.inria.fr/blog/2017/10/using-guix-without-being-root/),
pack`](https://hpc.guix.info/blog/2017/10/using-guix-without-being-root/),
in the [Docker (OCI) or Singularity
format](https://www.gnu.org/software/guix/manual/en/html_node/Invoking-guix-pack.html).
......@@ -126,5 +126,5 @@ us summarize the salient features of Guix.
Whatever your HPC or scientific use case is, we hope that you’ll find in
GNU Guix 1.0 the tools to further your goals. We’d love to hear from
you! You can email us at
[`guix-hpc@gnu.org`](https://guix-hpc.bordeaux.inria.fr/about/) or on
[`guix-hpc@gnu.org`](https://hpc.guix.info/about/) or on
the Guix [mailing lists](https://www.gnu.org/software/guix/contact/).
......@@ -17,50 +17,50 @@ Dave Love of the University of Manchester made a number of contributions
on the *high-performance networking* side:
- Dave added
[several](https://guix-hpc.bordeaux.inria.fr/package/opensm)
[InfiniBand](https://guix-hpc.bordeaux.inria.fr/package/ibutils)
[packages](https://guix-hpc.bordeaux.inria.fr/package/infiniband-diags),
added [PSM](https://guix-hpc.bordeaux.inria.fr/package/psm) support
[several](https://hpc.guix.info/package/opensm)
[InfiniBand](https://hpc.guix.info/package/ibutils)
[packages](https://hpc.guix.info/package/infiniband-diags),
added [PSM](https://hpc.guix.info/package/psm) support
in
[libfabric](https://guix-hpc.bordeaux.inria.fr/package/libfabric).
- [Open MPI](https://guix-hpc.bordeaux.inria.fr/package/openmpi)
[libfabric](https://hpc.guix.info/package/libfabric).
- [Open MPI](https://hpc.guix.info/package/openmpi)
gained support for
[RDMA](https://guix-hpc.bordeaux.inria.fr/package/rdma-core),
[RDMA](https://hpc.guix.info/package/rdma-core),
libfabric, and PSM.
- The [Intel MPI
Benchmarks](https://guix-hpc.bordeaux.inria.fr/package/imb-openmpi)
Benchmarks](https://hpc.guix.info/package/imb-openmpi)
are now available as a package.
- The size of the closure of the `openmpi` package with all its
dependencies was [significantly
reduced](https://debbugs.gnu.org/cgi/bugreport.cgi?bug=27905). This
is particularly appreciable if you’re [building application bundles
with `guix
pack`](https://guix-hpc.bordeaux.inria.fr/blog/2017/10/using-guix-without-being-root/).
pack`](https://hpc.guix.info/blog/2017/10/using-guix-without-being-root/).
The collection of *HPC profiling tools* has grown:
- [PAPI](https://guix-hpc.bordeaux.inria.fr/package/papi);
- [Score-P](https://guix-hpc.bordeaux.inria.fr/package/scorep-openmpi);
- [otf2](https://guix-hpc.bordeaux.inria.fr/package/otf2);
- [OPARI2](https://guix-hpc.bordeaux.inria.fr/package/opari2);
- [CUBE](https://guix-hpc.bordeaux.inria.fr/package/cube).
- [PAPI](https://hpc.guix.info/package/papi);
- [Score-P](https://hpc.guix.info/package/scorep-openmpi);
- [otf2](https://hpc.guix.info/package/otf2);
- [OPARI2](https://hpc.guix.info/package/opari2);
- [CUBE](https://hpc.guix.info/package/cube).
*Linear algebra* packages have seen some changes:
- [LAPACK](https://guix-hpc.bordeaux.inria.fr/package/lapack) was
- [LAPACK](https://hpc.guix.info/package/lapack) was
upgraded to 3.7.1, though 3.5.0 has been kept around for packages
that are not ready for the latest API.
- [Scotch](https://guix-hpc.bordeaux.inria.fr/package/scotch) is now
- [Scotch](https://hpc.guix.info/package/scotch) is now
built in 64-bit mode on 64-bit machines, and the new
[`scotch32`](https://guix-hpc.bordeaux.inria.fr/package/scotch32)
[`scotch32`](https://hpc.guix.info/package/scotch32)
package provides a 32-bit version for applications that need it.
- [BLIS](https://guix-hpc.bordeaux.inria.fr/package/blis) has been
- [BLIS](https://hpc.guix.info/package/blis) has been
added. Since it [does not support instruction set selection at
run-time yet](https://github.com/flame/blis/issues/129), we created
several packages specialized for each ISA extension:
[`blis-haswell`](https://guix-hpc.bordeaux.inria.fr/package/blis-haswell)
[`blis-haswell`](https://hpc.guix.info/package/blis-haswell)
is the variant optimized for Haswell CPUs,
[`blis-knl`](https://guix-hpc.bordeaux.inria.fr/package/blis-knl) is
[`blis-knl`](https://hpc.guix.info/package/blis-knl) is
the KNL variant, and so on. This is simplified by the
implementation of a [high-level `make-blis` function that returns
specialized
......@@ -68,7 +68,7 @@ The collection of *HPC profiling tools* has grown:
As far as *numerical simulation* is concerned:
- The [OpenFOAM](https://guix-hpc.bordeaux.inria.fr/package/openfoam)
- The [OpenFOAM](https://hpc.guix.info/package/openfoam)
C++ computational fluid dynamics package was contributed by Paul
Garlick of Tourbillion Technology.
......@@ -76,19 +76,19 @@ There’s been a lot of activity in *bioinformatics*, leading to a total
of almost 300 bioinformatics packages, notably with the addition of the
following packages:
- [f-seq](https://guix-hpc.bordeaux.inria.fr/package/f-seq);
- [GEMMA](https://guix-hpc.bordeaux.inria.fr/package/gemma);
- [GESS](https://guix-hpc.bordeaux.inria.fr/package/gess);
- [Kallisto](https://guix-hpc.bordeaux.inria.fr/package/kallisto);
- [kentutils](https://guix-hpc.bordeaux.inria.fr/package/kentutils);
- [Ritornello](https://guix-hpc.bordeaux.inria.fr/package/ritornello);
- [f-seq](https://hpc.guix.info/package/f-seq);
- [GEMMA](https://hpc.guix.info/package/gemma);
- [GESS](https://hpc.guix.info/package/gess);
- [Kallisto](https://hpc.guix.info/package/kallisto);
- [kentutils](https://hpc.guix.info/package/kentutils);
- [Ritornello](https://hpc.guix.info/package/ritornello);
- many, many R packages (428 R package are available today).
Changes to *core Guix* relevant to HPC include:
- `guix-daemon` gained the `--listen` command-line option that [we
covered
recently](https://guix-hpc.bordeaux.inria.fr/blog/2017/11/installing-guix-on-a-cluster/),
recently](https://hpc.guix.info/blog/2017/11/installing-guix-on-a-cluster/),
and which is handy when installing Guix cluster-wide.
- The new [`guix import
json`](https://www.gnu.org/software/guix/manual/html_node/Invoking-guix-import.html)
......
......@@ -14,49 +14,49 @@ sysadmins will enjoy. This release brings us close to our goals for
Over the 1,200+ packages added and 2,200+ packages updated since the
0.14.0 release, there are many noteworthy HPC additions.
- [Open MPI](https://guix-hpc.bordeaux.inria.fr/package/openmpi) has
- [Open MPI](https://hpc.guix.info/package/openmpi) has
been updated to 3.0.1. It now comes with plugins for Intel
OmniPath and Intel TrueScale _via_ the
[PSM2](https://guix-hpc.bordeaux.inria.fr/package/psm2) and
[PSM](https://guix-hpc.bordeaux.inria.fr/package/psm) packages.
- [hwloc](https://guix-hpc.bordeaux.inria.fr/package/hwloc) 2.0 is now
[PSM2](https://hpc.guix.info/package/psm2) and
[PSM](https://hpc.guix.info/package/psm) packages.
- [hwloc](https://hpc.guix.info/package/hwloc) 2.0 is now
available alongside the 1.x series, which some applications such as
[Slurm](https://guix-hpc.bordeaux.inria.fr/package/slurm) still
[Slurm](https://hpc.guix.info/package/slurm) still
require.
- [fftw](https://guix-hpc.bordeaux.inria.fr/package/fftw) was upgraded
- [fftw](https://hpc.guix.info/package/fftw) was upgraded
to 3.3.7. It is now built [with support for SIMD
extensions](https://git.savannah.gnu.org/cgit/guix.git/commit/?id=65bb22796f854cbc3eae053a80b1d64365dad376)
selected at run-time, which addressed performance concerns we
[discussed a while
back](https://guix-hpc.bordeaux.inria.fr/blog/2018/01/pre-built-binaries-vs-performance/).
- [superlu-dist](https://guix-hpc.bordeaux.inria.fr/package/superlu-dist)
back](https://hpc.guix.info/blog/2018/01/pre-built-binaries-vs-performance/).
- [superlu-dist](https://hpc.guix.info/package/superlu-dist)
was upgraded to 5.3.0.
- The [libpfm4](https://guix-hpc.bordeaux.inria.fr/package/libpfm4)
- The [libpfm4](https://hpc.guix.info/package/libpfm4)
performance monitoring library was added.
- [Spindle](https://guix-hpc.bordeaux.inria.fr/package/spindle), a
- [Spindle](https://hpc.guix.info/package/spindle), a
tool to improve the performance of dynamic library and Python module
loading on HPC clusters, as well as the companion
[LaunchMON](https://guix-hpc.bordeaux.inria.fr/package/launchmon)
[LaunchMON](https://hpc.guix.info/package/launchmon)
tool are now available, thanks to Pierre-Antoine Rouby who is
currently working as an [intern at
Inria](https://www.gnu.org/software/guix/blog/2018/guix-welcomes-outreachy-gsoc-and-guix-hpc-interns/).
- OpenCL headers,
[POCL](https://guix-hpc.bordeaux.inria.fr/package/pocl), and
[Beignet](https://guix-hpc.bordeaux.inria.fr/package/beignet) were
[POCL](https://hpc.guix.info/package/pocl), and
[Beignet](https://hpc.guix.info/package/beignet) were
contributed by “Fis Trivial”, without support for the lock-in NVIDIA
devices, but with support for modern Intel CPUs and GPUs.
- Other noteworthy changes include the addition of
[Elemental](https://guix-hpc.bordeaux.inria.fr/package/elemental)
and [QD](https://guix-hpc.bordeaux.inria.fr/package/qd), as well as
[Elemental](https://hpc.guix.info/package/elemental)
and [QD](https://hpc.guix.info/package/qd), as well as
upgrades to
[CERES](https://guix-hpc.bordeaux.inria.fr/package/ceres-solver) and
[Eigen](https://guix-hpc.bordeaux.inria.fr/package/eigen), thanks to
[CERES](https://hpc.guix.info/package/ceres-solver) and
[Eigen](https://hpc.guix.info/package/eigen), thanks to
Eric Bavier of Cray, Inc.
There have been many changes among the several hundreds of
*bioinformatics* and statistics packages available in Guix and we won’t
list them here. As part of the [work on reproducible bioinformatics
pipeline](https://guix-hpc.bordeaux.inria.fr/blog/2018/05/paper-on-reproducible-bioinformatics-pipelines-with-guix/),
pipeline](https://hpc.guix.info/blog/2018/05/paper-on-reproducible-bioinformatics-pipelines-with-guix/),
Ricardo Wurmus of the Max Delbrück Center for Molecular Medicine
contributed lots of fixes enabling bit-reproducible builds of packages,
in particular Python and R packages.
......@@ -67,7 +67,7 @@ Changes to *core Guix* relevant to HPC include:
executables](https://www.gnu.org/software/guix/blog/2018/tarballs-the-ultimate-container-image-format/).
The technique relies on _user namespaces_ so it’s unfortunately [not
directly usable on some
clusters](https://guix-hpc.bordeaux.inria.fr/blog/2017/09/reproducibility-and-root-privileges/).
clusters](https://hpc.guix.info/blog/2017/09/reproducibility-and-root-privileges/).
Nevertheless, it’s a simple and powerful way to distribute
applications for use on non-Guix systems.
- `guix pack` can now produce application bundles in the form of
......@@ -84,14 +84,14 @@ Changes to *core Guix* relevant to HPC include:
environments.
- `guix-daemon` will now reject attempts to run `guix gc` from a
remote node. This is done such that, on [cluster
installations](https://guix-hpc.bordeaux.inria.fr/blog/2017/11/installing-guix-on-a-cluster/),
installations](https://hpc.guix.info/blog/2017/11/installing-guix-on-a-cluster/),
garbage collection can only be triggered from the master node, which
is typically only accessible to sysadmins. This was contributed by
Roel Janssen of the Utrecht Bioinformatics Center.
- Roel Janssen’s
[hpcguix-web](https://guix-hpc.bordeaux.inria.fr/package/hpcguix-web),
[hpcguix-web](https://hpc.guix.info/package/hpcguix-web),
the web interface to browse packages that you can see [on this
site](https://guix-hpc.bordeaux.inria.fr/browse), is now available
site](https://hpc.guix.info/browse), is now available
[as a Guix system
system](https://www.gnu.org/software/guix/manual/en/html_node/Web-Services.html#index-hpcguix_002dweb_002dservice_002dtype),
thanks to the work of Pierre-Antoine Rouby at Inria.
......
......@@ -15,26 +15,26 @@ noteworthy packages and features for HPC and reproducible research.
Of the 985 packages were added and 1,945 upgraded, here are the
noteworthy HPC and bioinformatics changes:
- [FEniCS](https://guix-hpc.bordeaux.inria.fr/package/fenics) and
- [FEniCS](https://hpc.guix.info/package/fenics) and
related packages were contributed by Paul Garlick of Tourbillion
Technology.
- [`petsc-openmpi`](https://guix-hpc.bordeaux.inria.fr/package/petsc-openmpi)
- [`petsc-openmpi`](https://hpc.guix.info/package/petsc-openmpi)
is now configured with support for the HYPRE preconditioner, for
the MUMPS solver, and for the HDF5 data format.
- [SLEPc](https://guix-hpc.bordeaux.inria.fr/package/slepc) and
[PETSc](https://guix-hpc.bordeaux.inria.fr/package/petsc) were
- [SLEPc](https://hpc.guix.info/package/slepc) and
[PETSc](https://hpc.guix.info/package/petsc) were
upgraded to 3.10.1 and 3.10.2.
- [MPI4Py](https://guix-hpc.bordeaux.inria.fr/package/python-mpi4py)
- [MPI4Py](https://hpc.guix.info/package/python-mpi4py)
was added.
- [ngless](https://guix-hpc.bordeaux.inria.fr/package/ngless), a
- [ngless](https://hpc.guix.info/package/ngless), a
domain-specific language (DSL) for working with next generation
sequencing data is now available.
- A handful of packages for Nanopore processing have been added,
including
[Filtlong](https://guix-hpc.bordeaux.inria.fr/package/filtlong),
[Nanopolish](https://guix-hpc.bordeaux.inria.fr/package/nanopolish),
[Poretools](https://guix-hpc.bordeaux.inria.fr/package/poretools),
and [Porechop](https://guix-hpc.bordeaux.inria.fr/package/porechop).
[Filtlong](https://hpc.guix.info/package/filtlong),
[Nanopolish](https://hpc.guix.info/package/nanopolish),
[Poretools](https://hpc.guix.info/package/poretools),
and [Porechop](https://hpc.guix.info/package/porechop).
- Hundreds of bioinformatics and R packages and upgrades were
contributed in particular by Mădălin Ionel Patrașcu and Ricardo
Wurmus of the Max Delbrück Center for Molecular Medicine (MDC).
......@@ -55,7 +55,7 @@ solves several problems:
pull`](https://www.gnu.org/software/guix/manual/en/html_node/Invoking-guix-pull.html)
pull not just Guix, but also external package repositories such as
[the HPC and bioinfo repositories of our
institutes](https://guix-hpc.bordeaux.inria.fr/about/).
institutes](https://hpc.guix.info/about/).
- It allows you to _replicate_ a Guix setup with all its channels.
As an example, if I want to use packages from the [Guix-HPC repository
......
......@@ -5,7 +5,7 @@
(use-modules (srfi srfi-19))
(define %video-url
"https://guix-hpc.bordeaux.inria.fr/static/videos/pigx.webm")
"https://hpc.guix.info/static/videos/pigx.webm")
`((title . "PiGx paper awarded at the International Conference on Genomics (ICG-13)")
(author . "Ricardo Wurmus")
......
......@@ -17,7 +17,7 @@ them to their needs. This aspect is key to reproducible computational
experiments: scientists need to reproduce software environments before
they can reproduce experimental results, and this is one of the things
we are focusing on in the context of the
[Guix-HPC](https://guix-hpc.bordeaux.inria.fr) effort. At a lower
[Guix-HPC](https://hpc.guix.info) effort. At a lower
level, the project, along with others in the [Reproducible
Builds](https://reproducible-builds.org) community, is working to ensure
that software build outputs are [reproducible,
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment