use github pages for the docs (#4)

.github/workflows/actions.yml → .github/workflows/lint_test.yml

@@ -1,4 +1,4 @@
-name: GitHub Actions Demo
+name: Lint and test
 run-name: CI/CD for ${{ github.ref_name }}
 on:
   pull_request:

.github/workflows/publish.yml

@@ -0,0 +1,34 @@
+name: Publish repo
+
+on:
+  push:
+    tags:
+      - 'v[0-9]+.[0-9]+.[0-9]+'
+  workflow_dispatch: {}
+
+concurrency:
+  group: fast-${{ github.head_ref || github.run_id }}
+  cancel-in-progress: true
+
+jobs:
+  deploy_docs:
+    name: Deploy docs (ubuntu)
+    runs-on: ubuntu-latest
+    steps:
+    - name: Checkout emulators
+      uses: actions/checkout@v4
+    - name: Set up Python 3.10
+      uses: actions/setup-python@v5
+      with:
+        python-version: '3.10'
+        cache: 'pip'
+    - name: Setup environment
+      run: |
+        pip install -r doc_requirements.txt
+        pip install -e .
+    - name: Deploy docs
+      run: |
+        git config user.name "GitHub Actions"
+        git config user.email "[email protected]"
+        git fetch origin gh-pages
+        mike deploy --push --update-aliases ${{ github.ref_name }} latest

LICENSE

@@ -0,0 +1,15 @@
+PASQAL OPEN-SOURCE SOFTWARE LICENSE AGREEMENT (MIT-derived)
+
+The author of the License is:
+  Pasqal, a Société par Actions Simplifiée (Simplified Joint Stock Company) registered under number 849 441 522 at the Registre du commerce et des sociétés (Trade and Companies Register) of Evry – France, headquartered at 7 rue Leonard de Vinci – 91300 – Massy – France, duly represented by its Président, M. Georges-Olivier REYMOND,
+
+Hereafter referred to as « the Licensor »
+
+- Permission is hereby granted, free of charge, to any person obtaining a copy of this software (the “Licensee”) and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
+The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. The Software is “as is”, without warranty of any kind, express or implied, including but not limited to the warranties of merchantability, fitness for a particular purpose and non-infringement. In no event shall the authors or copyright holders be liable for any claim, damages or other liability, whether in an action of contract, tort or otherwise arising from, out of or in connection with the Software or the use or other dealings in the Software.
+
+- If use of the Software leads to the necessary use of any patent of the Licensor and/or any of its Affiliates (defined as a company owned or controlled by the Licensor), the Licensee is granted a royalty-free license, in any country where such patent is in force, to use the object of such patent; or use the process covered by such patent,
+
+- Such a patent license is granted for internal research or academic use of the Licensee's, which includes use by employees and students of the Licensee, acting on behalf of the Licensee, for research purposes only.
+
+- The License is governed by the laws of France. Any dispute relating to the License, notably its execution, performance and/or termination shall be brought to, heard and tried by the Tribunal Judiciaire de Paris, regardless of the rules of jurisdiction in the matter.

doc_requirements.txt

@@ -8,3 +8,4 @@ mkdocs-exclude
 notebook<7
 markdown-exec
 jupyter_contrib_nbextensions
+mike

docs/benchmarks/performance.md

@@ -28,7 +28,7 @@ In the right column (both CPU and GPU figure), we explore the available register
 Let us now analyze the performance of the simulation for a much larger system. We show results for the adiabatic sequence with 49 qubits arranged in a 7x7 grid. The parameters of the simulation were `max_bond_dim = 1600`, `extra_krylov_tolerance=1e-5` and `precision=1e-6`, and we ran the simulation on 2 GPUs. The maximum bond dimension of the state, its size in memory, the total memory consumption of the program on GPU 1, and the time taken per emulation timestep (there are 390 timesteps of `dt=10 ns` each) are shown in the graph below.
 
 <div style="text-align:center;">
-<img src="./images/performance.png" width="50%" style="background-color:white;">
+<img src="../benchmark_plots/afm_performance_big.png" width="50%" style="background-color:white;">
 </div>
 
 First off, note that the peak memory consumption on GPU 1 reaches almost 30GB at the end of the simulation and the memory profile on GPU 2 will be very similar. Note that this memory consumption can be estimated ([see here](../advanced/memory.md#an-example)), and that the simulation would not have fit on a single GPU. Next, the memory consumption stops increasing as quickly when the maximum bond dimension plateaus, but it does not stop increasing entirely. This is because when the maximum bond dimension reaches the cutoff value of `1600`, most of the tensors in the MPS will not have reached maximum size yet. However, the rate of memory consumption growth will decrease as more of the tensors reach this maximum size.
@@ -67,7 +67,9 @@ The unshuffled register ordering is that given by `Register.rectangle` as used i
  <tr><td> 5 </td><td> 1 </td><td> 9 </td></tr>
  <tr><td> 8 </td><td> 3 </td><td> 6 </td></tr>
 </table>
-<img src="../benchmark_plots/qubit_shuffling_cpu.png"  width="49.7%">
+<div style="text-align:center;">
+<img src="../benchmark_plots/qubit_shuffling_cpu.png"  width="90%">
+</div>
 
 The left column of the image shows no accuracy degradation from the qubit shuffling, returning equivalent observables. That is expected since both runs were able to converge to the desired precision.
 

mkdocs.yml

@@ -1,5 +1,5 @@
 site_name: Emu-MPS
-repo_url: "https://gitlab.pasqal.com/emulation/rydberg-atoms/emu-ct"
+repo_url: "https://github.com/pasqal-io/emulators"
 repo_name: "emu-mps"
 
 nav: