Skip to content
Snippets Groups Projects
Commit 47b0afde authored by Claudia Frauen's avatar Claudia Frauen
Browse files

Added to OpenMP vs MPI comparison

parent d02c71ab
No related branches found
No related tags found
1 merge request!72Draft: Compute devices lecture
Pipeline #70709 passed
Stage: test
Stage: build
Hide whitespace changes
Inline Side-by-side
lectures/hardware/slides.qmd
+ 2
− 1
View file @ 47b0afde
...@@ -179,10 +179,11 @@ If we want to scale beyond one node, what do we need in our example? ...@@ -179,10 +179,11 @@ If we want to scale beyond one node, what do we need in our example?
:::{.smaller} :::{.smaller}
| OpenMP | MPI | | OpenMP | MPI |
| ------ | --- | | --- | --- |
| shared memory | distributed memory | | shared memory | distributed memory |
| can't scale beyond one node | can scale across many nodes | | can't scale beyond one node | can scale across many nodes |
| uses independent threads| uses ranks that need to communicate with each other | | uses independent threads| uses ranks that need to communicate with each other |
| automatic boundary exchange due to shared memory | explicit boundary exchange needed |
| pragmas inserted in the code tell the compiler where to parallelise | API to vendor specific libraries | | pragmas inserted in the code tell the compiler where to parallelise | API to vendor specific libraries |
* OpenMPI is a MPI implementation and has nothing to do with OpenMP * OpenMPI is a MPI implementation and has nothing to do with OpenMP
......
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

Imprint | Privacy Policy