Learn Editor: Update compiling-scaling-applications.md

Author: mattmcinnes

articles/virtual-machines/compiling-scaling-applications.md

@@ -26,15 +26,13 @@ The [azurehpc repo](https://github.com/Azure/azurehpc) contains many examples of
 
 The following suggestions apply for optimal application scaling efficiency, performance, and consistency:
 
-- For smaller scale jobs (< 256 K connections) use the option:
+- For smaller scale jobs (< 256 K connections) use:
    ```bash  UCX_TLS=rc,sm  ```
-- For larger scale jobs (> 256 K connections) use the option:
+- For larger scale jobs (> 256 K connections) use:
    ```bash  UCX_TLS=dc,sm  ```
-- In the above, to calculate the number of connections for your MPI job, use:
-   ```bash
-   Max Connections = (processes per node) x (number of nodes per job) x (number of nodes per job) 
-   ```
-
+- To calculate the number of connections for your MPI job, use:
+   ```bash  Max Connections = (processes per node) x (number of nodes per job) x (number of nodes per job)   ```
+   
 ## Adaptive Routing
 Adaptive Routing (AR) allows Azure Virtual Machines (VMs) running EDR and HDR InfiniBand to automatically detect and avoid network congestion by dynamically selecting optimal network paths. As a result, AR offers improved latency and bandwidth on the InfiniBand network, which in turn drives higher performance and scaling efficiency. For more information, see [TechCommunity article](https://techcommunity.microsoft.com/t5/azure-compute/adaptive-routing-on-azure-hpc/ba-p/1205217).
 
@@ -45,7 +43,7 @@ Adaptive Routing (AR) allows Azure Virtual Machines (VMs) running EDR and HDR In
 - For hybrid parallel applications (OpenMP+MPI), use 4 threads and 1 MPI rank per [CCX]([HB-series virtual machines overview including info on CCXs](/azure/virtual-machines/hb-series-overview)) on HB and HBv2 VM sizes.
 - For pure MPI applications, experiment with 1-4 MPI ranks per CCX for optimal performance on HB and HBv2 VM sizes.
 - Some applications with extreme sensitivity to memory bandwidth may benefit from using a reduced number of cores per CCX. For these applications, using three or two cores per CCX may reduce memory bandwidth contention and yield higher real-world performance or more consistent scalability. In particular, MPI 'Allreduce' may benefit from this approach.
-- For larger scale runs, it's recommended to use UD or hybrid RC+UD transports. Many MPI libraries/runtime libraries do this internally (such as UCX or MVAPICH2). Check your transport configurations for large-scale runs.
+- For larger scale runs, it's recommended to use UD or hybrid RC+UD transports. Many MPI libraries/runtime libraries use these transports internally (such as UCX or MVAPICH2). Check your transport configurations for large-scale runs.
 
 ## Compiling applications
 <br>
@@ -80,14 +78,15 @@ $ gfortran [gFortran flags]
    -c xyz.f90 $ clang -O3 -lgfortran -o xyz xyz.o $./xyz
 ```
 ### PGI Compiler
-PGI Community Edition 17 is confirmed to work with AMD EPYC. A PGI-compiled version of STREAM does deliver full memory bandwidth of the platform. The newer Community Edition 18.10 (Nov 2018) should likewise work well. A sample CLI to compiler optimally with the Intel Compiler:
+PGI Community Edition 17 is confirmed to work with AMD EPYC. A PGI-compiled version of STREAM does deliver full memory bandwidth of the platform. The newer Community Edition 18.10 (Nov 2018) should likewise work well. Use this CLI command to compile with the Intel Compiler:
+
 
 ```bash
 pgcc $(OPTIMIZATIONS_PGI) $(STACK) -DSTREAM_ARRAY_SIZE=800000000 stream.c -o stream.pgi
 ```
 
 ### Intel Compiler
-Intel Compiler 18 is confirmed to work with AMD EPYC. Below is sample CLI to compiler optimally with the Intel Compiler.
+Intel Compiler 18 is confirmed to work with AMD EPYC. Use this CLI command to compile with the Intel Compiler.
 
 ```bash
 icc -o stream.intel stream.c -DSTATIC -DSTREAM_ARRAY_SIZE=800000000 -mcmodel=large -shared-intel -Ofast –qopenmp