Category: News
-
First stage of MLIR switch is live
We have released new development versions of Adaptyst, the linuxperf Adaptyst module, and the nvgpu Adaptyst module: v0.1.0-dev.2026.03a, v0.1.0-dev.2026.03a, and v0.1.0-dev.2026.03a respectively. These implement the first stage of the MLIR switch described in the previous blog post. The documentation has been updated accordingly as well. At the same time, we are reintroducing Docker and Apptainer/Singularity…
-
Details of the upcoming IR switch
As mentioned in the previous post, the intermediate representation (IR) used by Adaptyst is changing to MLIR. We can share more technical details about this now: the switch will be specifically done in two stages. We hope that this makes the situation clear, especially for module developers. We will publish another blog post once the…
-
IR used by Adaptyst is changing to MLIR
Since the release of the modular design of Adaptyst, the tool has used stateful multiflow datagraphs (SDFGs) as its intermediate representation (IR) in the performance analysis flow. This is achieved through DaCe and the main rationale can be split into two categories: “Why SDFGs?” and “Why DaCe?”. For “Why SDFGs?”, the reason is the potential…
-
New version with LGPL v3+ licensing and setup facilitation delays
We have released new development versions of Adaptyst, Adaptyst Analyser, the linuxperf Adaptyst module, and the linuxperf Adaptyst Analyser module: v0.1.0-dev.2026.02a, v0.1.0.dev5, v0.1.0-dev.2026.02a, and v0.1.0-dev.2026.02a respectively. Please see the attached links for the detailed changelogs. The main update we want to emphasise in this post is the license change of Adaptyst and Adaptyst Analyser. To…
-
Making Adaptyst setup easier
Adaptyst along with the linuxperf and nvgpu modules has been available for installation only from source so far. To increase the accessibility of our tool for users, that situation is about to change. Within the next two weeks (i.e. by 20 February inclusive), three new things will be introduced: Additionally, for the CERN community, we…
-
Adaptyst at FOSDEM 2026
If you are around in Brussels on 1 February, please consider joining the Adaptyst presentation at FOSDEM 2026! The main developer will talk about performance challenges at CERN and how Adaptyst aims to address these. The presentation will also serve as an invitation for everybody to join our open-source development efforts as a contributor. The…
-
Holiday greetings
As the end-of-year closure at CERN has begun, we want to wish you a peaceful and relaxing break along with a merry Christmas and happy New Year to everyone celebrating! Blog posting will resume in the week of 12 January 2026.
-
Adaptyst HPC deployability and AMD GPU support at LUMI-CSCS hackathon in Lugano
This post is published under the same CC BY-SA 3.0 Unported license. On 1-5 December 2025, we participated in the LUMI-CSCS hackathon organised by the Swiss National Supercomputing Centre (CSCS) in Hotel de la Paix in Lugano, Switzerland. The theme of the event was porting and optimising various applications for the LUMI supercomputer in Finland.…
-
IEEE HPEC’25 award-winning Adaptyst paper available on arXiv
IEEE HPEC’25 award-winning Adaptyst paper available on arXiv Some of you may have heard of an Adaptyst-related paper getting the Outstanding Short Paper Award at the IEEE HPEC conference this year. If you are interested in looking at it, it is available now on arXiv: https://arxiv.org/abs/2511.13928 (the organisers do not publish short papers in IEEE…
-
Multiple modules can be used now!
We have just deployed bugfixes and enhancements addressing the issue of not being able to use multiple modules/nodes/edges in Adaptyst and Adaptyst Analyser. These are available in Adaptyst v0.1.0-dev.2025.11b and Adaptyst Analyser v0.1.0.dev3. The new update allows you now to e.g. analyse performance of CPU and basic NVIDIA GPU activity of your workflows at the…