Summary
In this module, you learned how to:
- Select the right HPC VM for different types of HPC applications.
- Optimally pin processes and threads to maximize performance.
- Select the correct numerical libraries, compilers, and flags to maximize performance.
- Select the right MPI library and flags to maximize HPC application performance.
Important
In the optional exercises for this module, you created resources by using your own Azure subscription. Clean up these resources so that you won't continue to be charged for them.
Learn more
For more information about the Azure technologies we discussed in this module, see the following articles:
- Azure image builder
- Azure HPC VM images
- High-performance computing VM sizes
- HB-series VMs
- HBv2-series VMs
- HC-series VMs
For more information about non-Microsoft products we mentioned in this module, see the following links:
- AMD Optimizing Compiler suite (AOCC)
- AMD Optimized CPU libraries (AOCL)
- AMD Optimizing CPU libraries (AMD-AOCL)
- Basic Linear Algebra Subprograms (BLAS)
- Fastest Fourier Transform in the West (FFTW)
- HashiCorp Packer
- Intel Math Kernel Library (MKL)
- Intel Memory Latency Checker (MLC)
- Linear Algebra PACKage (LAPACK)
- Mellanox Openfabrics Enterprise Distribution (MOFED)
- UCX framework