Merge branch 'master' into application_testing

.github/workflows/pages-html.yml

@@ -0,0 +1,25 @@
+name: Deploy HTML slides to Pages
+
+on:
+  # Runs on pushes targeting the default branch
+  push:
+    branches:
+      - "master"
+    paths:
+      - "*/docs/**"
+      - ".github/workflows/pages.yml"
+
+  # Allows you to run this workflow manually from the Actions tab
+  workflow_dispatch:
+
+# Sets permissions of the GITHUB_TOKEN to allow deployment to GitHub Pages
+permissions:
+  contents: read
+  pages: write
+  id-token: write
+
+jobs:
+  pages-html:
+    uses: ./.github/workflows/pages.yml
+    with:
+      include_pdf: false

.github/workflows/pages-pdf.yml

@@ -0,0 +1,17 @@
+name: Deploy HTML and PDF slides to Pages
+
+on:
+  # Allows you to run this workflow manually from the Actions tab
+  workflow_dispatch:
+
+# Sets permissions of the GITHUB_TOKEN to allow deployment to GitHub Pages
+permissions:
+  contents: read
+  pages: write
+  id-token: write
+
+jobs:
+  pages-pdf:
+    uses: ./.github/workflows/pages.yml
+    with:
+      include_pdf: true

.github/workflows/pages.yml

@@ -0,0 +1,65 @@
+# Script based on examples in https://github.com/actions/starter-workflows/tree/main/pages
+name: Deploy slides to Pages
+
+on:
+  workflow_call:
+    inputs:
+      include_pdf:
+        required: true
+        type: boolean
+
+# Sets permissions of the GITHUB_TOKEN to allow deployment to GitHub Pages
+permissions:
+  contents: read
+  pages: write
+  id-token: write
+
+# Allow only one concurrent deployment, skipping runs queued between the run in-progress and latest queued.
+# However, do NOT cancel in-progress runs as we want to allow these production deployments to complete.
+concurrency:
+  group: "pages"
+  cancel-in-progress: false
+
+jobs:
+  build:
+    timeout-minutes: 30
+    runs-on: ubuntu-latest
+    container:
+      image: ghcr.io/csc-training/slidefactory:3.1.0-beta.4
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+      - name: Setup Pages
+        id: pages
+        uses: actions/configure-pages@v4
+      - name: Build slides
+        env:
+          INCLUDE_PDF: ${{ inputs.include_pdf }}
+        run: |
+          git config --global --add safe.directory $PWD
+          GIT_SHORT_SHA=$(git rev-parse --short $GITHUB_SHA)
+          GIT_DATE=$(git show -s --format=%ci $GITHUB_SHA)
+
+          ARGS=""
+          [[ "$INCLUDE_PDF" == "true" ]] && ARGS="--with-pdf"
+
+          slidefactory pages about.yml build \
+              --filters tools/mpi_links.py \
+              --info_content "Updated for [$GIT_SHORT_SHA]($GITHUB_SERVER_URL/$GITHUB_REPOSITORY/commit/$GITHUB_SHA) ($GIT_DATE)" \
+              $ARGS
+
+      - name: Upload artifact
+        uses: actions/upload-pages-artifact@v3
+        with:
+          path: ./build
+
+  deploy:
+    environment:
+      name: github-pages
+      url: ${{ steps.deployment.outputs.page_url }}
+    runs-on: ubuntu-latest
+    needs: build
+    steps:
+      - name: Deploy to GitHub Pages
+        id: deployment
+        uses: actions/deploy-pages@v4

.gitignore

@@ -7,6 +7,8 @@
 *.html
 *.pdf
 *.sif
+build/
+core
 
 #ignore default binary name
 a.out

README.md

@@ -1,12 +1,15 @@
-# CSC Summer School in High-Performance Computing 2023
+# CSC Summer School in High-Performance Computing 2024
 
-This is the material repository for the high-performance computing summer school by CSC - Finnish IT Center for Science.
+This is the material repository for the High-Performance Computing Summer School organized by [CSC - IT Center for Science](https://csc.fi/en/).
 
-Feel free to fork this repository and work through the exercises, see more details in 
-[exercise instructions](exercise-instructions.md). You can also add general notes to yourself (like how to compile files etc.) at the end of this readme file (`README.md`). 
+Feel free to fork this repository to work through the exercises.
 
 Versions from previous years can be found in tags.
 
+## Presentation slides
+
+The slides are available [here](https://github.com/csc-training/summerschool/releases/).
+
 ## Exercises
 
  - [General instructions](exercise-instructions.md)
@@ -16,7 +19,3 @@ Versions from previous years can be found in tags.
  - [Hybrid MPI/OpenMP](hybrid)
  - [GPU programming with OpenMP](gpu-openmp)
  - [GPU programming with HIP](gpu-hip)
-
----
-## Notes
-- [x] Have fun!

about.yml

@@ -0,0 +1,13 @@
+# This file is used in the generation of the web page
+title: CSC Summer School in High-Performance Computing
+modules:
+  - intro-to-hpc
+  - computer-platforms
+  - mpi
+  - parallel-io
+  - hybrid
+  - gpu-openmp
+  - gpu-hip
+  - application-performance
+  - application-testing
+  - build-systems

application-design/README.md

@@ -1,3 +1,3 @@
-## Exercieses
+## Exercises
 
 - [Continuous integration](ci/)

application-design/about.yml

@@ -0,0 +1,3 @@
+# This file is used in the generation of the web page
+title: Application Design
+slidesdir: docs

application-design/docs/01_design_choices.md

@@ -1,6 +1,6 @@
 ---
 title:  Application design
-event:  CSC Summer School in High-Performance Computing 2023
+event:  CSC Summer School in High-Performance Computing 2024
 lang:   en
 ---
 

application-design/docs/02-documentation-testing.md

@@ -1,6 +1,6 @@
 ---
 title:  Documenting and testing
-event:  CSC Summer School in High-Performance Computing 2023
+event:  CSC Summer School in High-Performance Computing 2024
 lang:   en
 ---
 

application-design/docs/03-collaboration-release.md

@@ -1,6 +1,6 @@
 ---
 title:  Collaboration and release
-event:  CSC Summer School in High-Performance Computing 2023
+event:  CSC Summer School in High-Performance Computing 2024
 lang:   en
 ---
 

application-performance/about.yml

@@ -0,0 +1,3 @@
+# This file is used in the generation of the web page
+title: Application Performance
+slidesdir: docs

application-performance/docs/01-application-performance-intro.md

@@ -1,6 +1,6 @@
 ---
 title:  Introduction to Application Performance
-event:  CSC Summer School in High-Performance Computing 2023
+event:  CSC Summer School in High-Performance Computing 2024
 lang:   en
 ---
 
@@ -41,7 +41,7 @@ Have you profiled your code?
 
 - Profiled the code: 99.9% of the execution time was being spent on these lines:
 
-```fortran
+```fortranfree
 do i=1,n          ! Removing these unnecessary loop iterations reduced the
  do j=1,m         ! wall-time of one simulation run from 17 hours to 3 seconds…
     do k=1,fact(x)

application-performance/docs/02-gpu-performance.md

@@ -1,27 +1,30 @@
 ---
-title:  Single node performance optimization
-event:  CSC Summer School in High-Performance Computing 2023
+title:  GPU performance optimization
+event:  CSC Summer School in High-Performance Computing 2024
 lang:   en
 ---
 
 # GPU performance optimization {.section}
 
 # Introduction
-- GPUs (Graphics Processing Units) are widely used in High-Performance Computing (HPC) applications.
-- GPUs are powerful and complex processors designed for parallel computing.
-- GPUs require explicit expression of parallelism by the programmer.
+
+- GPUs (Graphics Processing Units) are widely used in High-Performance Computing (HPC) applications
+- GPUs are powerful and complex processors designed for parallel computing
+- GPUs require explicit expression of parallelism by the programmer
 
 # General Principles for High GPU Performance 
 
-<div class=column>
+:::::: {.columns}
+::: {.column width="50%"}
 - Keep all the compute resources busy (idle resources are a waste)
 - Minimize the synchronization at all levels
 - Minimize the data transfers between host and device 
 - Keep the data in faster memory and use an appropriate access pattern
-</div>
-<div class=column>
-![](img/lumi_node.png){.center width=40%}
-</div>
+:::
+::: {.column width="50%"}
+![](img/lumi_node.png){.center width=50%}
+:::
+::::::
 
 # GPU performance analysis {.section}
 
@@ -30,6 +33,7 @@ lang:   en
 ![](img/perf-analysis-single-gpu.svg){.center width=60%}
 
 # Measuring performance
+
 - Don’t speculate about performance – measure it!
 - Performance analysis tools help to
     - Find hot-spots
@@ -43,12 +47,10 @@ lang:   en
 
 # Hardware performance counters
 
-- Hardware performance counters are special registers on CPU \& GPU that count
-  hardware events
-- They enable more accurate statistics and low overhead
+- Special registers on CPU \& GPU that count hardware events
+- Enable more accurate statistics and low overhead
     - In some cases they can be used for tracing without any extra
       instrumentation
-
 - Number of counters is much smaller than the number of events that can be
   recorded
 - Different devices have different counters
@@ -71,71 +73,65 @@ lang:   en
 - Start with an overview!
     - Call tree information, what routines are most expensive?
 
-# Sampling vs. Tracing
+# [Sampling]{.underline} vs. Tracing
 
-<div class=column>
-Sampling
+:::::: {.columns}
+::: {.column width="50%"}
 
 - Application is stopped at predetermined intervals 
     - Information is collected about the state of application
 - Lightweight, but may give skewed results
     - Statistical information
 
-</div>
-<div class=column>
-Tracing
+:::
+::: {.column width="50%"}
 
-- Records events, e.g., every function call
-- Requires usually modification to the executable *i.e.* instrumentation
-- More accurate, but may affect program behavior
-- Generates often lots of data
-
-</div>
+![](img/sampling.png){.left width=100%}
 
-# Sampling vs. Tracing
+:::
+::::::
 
-<div class=column>
-Sampling
+# Sampling vs. [Tracing]{.underline}
 
-![](img/sampling.png){.left width=80%}
+:::::: {.columns}
+::: {.column width="50%"}
 
-</div>
-<div class=column>
-Tracing
+- Records events, e.g., every function call
+- Requires usually modification to the executable: code instrumentation
+- More accurate than sampling, but may affect program behavior
+- Generates often lots of data
 
-![](img/tracing.png){.left width=80%}
+:::
+::: {.column width="50%"}
 
-</div>
+![](img/tracing.png){.left width=100%}
 
-# Tau  Analysis Utilities
+:::
+::::::
 
-<small> 
+# Tau Analysis Utilities
 
-- TAU is a powerful performance evaluation toolkit
+- <https://www.cs.uoregon.edu/research/tau/home.php>
+- A performance evaluation toolkit
 - Runs on all HPC platforms, relatively easy to install
 - Targets all parallel programming/execution paradigms (GPU, MPI, OpenMP, pthreads, ...)
-- Programming languages: Fortran, C, C++, UPC,  Java, Python, ...
+- Programming languages: Fortran, C, C++, UPC, Java, Python, ...
+
+# Tau Analysis Utilities cont.
+
 - TAU has instrumentation, measurement and analysis tools 
 - User-friendly graphical interface
 - Profiling: Measures total time spent in each routine
 - Tracing: Shows events and their timings across processes on a timeline
 - I/O performance evaluation
 - Memory debugging
 
-</small>
-
 # Omniperf Tools
 
+- <https://amdresearch.github.io/omniperf/getting_started.html>
  - system performance profiling tool for machine
-learning/HPC workloads running on AMD MI GPUs. 
- - presently targets usage on MI100 and MI200 accelerators.
+learning/HPC workloads running on AMD MI GPUs 
+ - presently targets usage on MI100 and MI200 accelerators
  - profiling, roofline model, tracing
  - built on top of `roctracer` and `rocprof`  
- - supports both a web-based GUI and a command-line analyzer for user convenience.
-
-
-# Web Resources
- - TAU homepage
-    - <https://www.cs.uoregon.edu/research/tau/home.php>
- - Omniperf
-    - <https://amdresearch.github.io/omniperf/getting_started.html>
+ - supports both a web-based GUI and a command-line analyzer for user convenience

application-performance/docs/03-mpi-performance.md

@@ -1,6 +1,6 @@
 ---
 title:  MPI performance analysis
-event:  CSC Summer School in High-Performance Computing 2023
+event:  CSC Summer School in High-Performance Computing 2024
 lang:   en
 ---
 

application-performance/docs/03-single-node-performance.md → application-performance/docs/04-single-node-performance.md

@@ -1,6 +1,6 @@
 ---
 title:  Single node performance optimization
-event:  CSC Summer School in High-Performance Computing 2023
+event:  CSC Summer School in High-Performance Computing 2024
 lang:   en
 ---