This shows you the differences between two versions of the page.
| Both sides previous revision Previous revision Next revision | Previous revision | ||
|
howto:windows [2024/04/17 22:06] ccrosby |
howto:windows [2024/04/17 22:39] (current) ccrosby [Running Windows applications] |
||
|---|---|---|---|
| Line 3: | Line 3: | ||
| * The Linux-based analysis or simulation software has no GUI, but there is a Windows GUI. | * The Linux-based analysis or simulation software has no GUI, but there is a Windows GUI. | ||
| * The Windows application may be a pre- or post-processor for a native Linux application. | * The Windows application may be a pre- or post-processor for a native Linux application. | ||
| - | * If the user lacks a workstation with sufficient | + | * If the user lacks a workstation with sufficient |
| + | |||
| + | ==== Consider Cloud ==== | ||
| + | The CHPC has its own [[https:// | ||
| ==== Have some Wine ==== | ==== Have some Wine ==== | ||
| - | [[https:// | + | [[https:// |
| - | < | + | |
| - | / | + | |
| - | </ | + | |
| Therefore, if you need to run a Windows application on the cluster, you can do so as follows: | Therefore, if you need to run a Windows application on the cluster, you can do so as follows: | ||
| Line 64: | Line 64: | ||
| </ | </ | ||
| + | |||
| + | The above file listing in '' | ||
| + | - Wine creates a Windows directory structure in your home directory. | ||
| + | - This could use up your home directory quota rather quickly. | ||
| + | - The " | ||
| + | |||
| + | === Wine example === | ||
| + | A Windows application may be installed in exactly the same way. The application will be installed in '' | ||
| + | < | ||
| + | $ singularity run / | ||
| + | </ | ||
| + | |||
| + | {{: | ||
| + | |||
| + | It goes without saying that you need to use either X-forwarding or [[howto: | ||
| + | |||
| + | ==== Final thoughts on using Singularity and Wine ==== | ||
| + | * Don' | ||
| + | * Unless you cannot avoid it. | ||
| + | * The mere fact that something is possible does not automatically make it a good idea. | ||
| + | * We are not aware of a practical way of doing multi-node (MPI) runs. OpenMP shared-memory multi-thread parallelism does, however work, and makes good use of the available resources. | ||
| + | * Raw computational performance is very good, and is only marginally slower than the native Linux application. | ||
| + | * Compatibility is surprisingly good. Many things work, but, of course, there are no guarantees. | ||
| + | |||
| + | |||