edits

Author: laurenhughes

articles/virtual-machines/TOC.yml

@@ -12,6 +12,8 @@
     href: workloads/oracle/oracle-considerations.md
   - name: SAP
     href: workloads/sap/get-started.md
+  - name: High performance computing
+    href: workloads/hpc/configure.md
   - name: Mainframe rehosting
     href: workloads/mainframe-rehosting/overview.md
 - name: Classic deployments

articles/virtual-machines/workloads/hpc/hb-hc-known-issues.md

@@ -1,27 +1,31 @@
 ---
-title: Known issues for HB and HC-series VMs - Azure Virtual Machines | Microsoft Docs
-description: Learn about known issues with HB and HC-series VMs in Azure. 
+title: Known issues for HB-series VM sizes in Azure | Microsoft Docs
+description: Learn about known issues with HB-series VM sizes in Azure. 
 services: virtual-machines
 documentationcenter: ''
-author: githubname
+author: vermagit
 manager: jeconnoc
 editor: ''
 tags: azure-resource-manager
 
 ms.service: virtual-machines
 ms.workload: infrastructure-services
 ms.topic: article
-ms.date: 05/15/2019
-ms.author: msalias
+ms.date: 05/07/2019
+ms.author: amverma
 ---
 
 # Known issues with HB and HC-series VM sizes
 
-This article describes known issues with SR-IOV enabled HB and HC-series VMs.
+<Intro>
+
+## Platform MPI
+
+Platform MPI may not automatically know which PKEY to use for InfiniBand. See Slide 8, Step 1 to discover the necessary PKEY.
 
 ## DRAM on HB-series
 
-HB-series VMs can only expose 228 GB of RAM to guest VMs at this time. This is due to a known limitation of Azure hypervisor to prevent pages from being assigned to the local DRAM of AMD CCX’s (NUMA domains) reserved for the guest VM.
+HB-series VMs can only expose 228 GB of RAM to guest VMs at this time. This is due to a known limitation of Azure hypervisor to prevent pages from being assigned to the local DRAM of AMD CCX’s (NUMA domains) reserved for the guest VM. An upcoming update will address this issue.
 
 ## Accelerated Networking
 
@@ -85,4 +89,4 @@ You can ignore this warning. This is due to a known limitation of the Azure hype
 
 ## Next steps
 
-Learn more about [HPC](https://docs.microsoft.com/azure/architecture/topics/high-performance-computing/) in Azure.
+Learn more about [high-performance computing](https://docs.microsoft.com/azure/architecture/topics/high-performance-computing/) in Azure.

articles/virtual-machines/workloads/hpc/hb-series-overview.md

@@ -1,5 +1,5 @@
 ---
-title: Preview of the HB-series - Azure Virtual Machines | Microsoft Docs
+title: Preview of the HB-series in Azure | Microsoft Docs
 description: Learn about the preview support for the HB-series VM size in Azure. 
 services: virtual-machines
 documentationcenter: ''
@@ -11,17 +11,15 @@ tags: azure-resource-manager
 ms.service: virtual-machines
 ms.workload: infrastructure-services
 ms.topic: article
-ms.date: 05/07/2019
+ms.date: 05/16/2019
 ms.author: amverma
 ---
 
 # HB-Series overview
 
-This article provides recommended practices, observed patterns, and recipes to get started with High Performance Computing (HPC) workloads on Azure HC-series VMs.
+Maximizing high performance compute (HPC) application performance on AMD EPYC requires a thoughtful approach memory locality and process placement. Below we outline the AMD EPYC architecture and our implementation of it on Azure for HPC applications. We will use the term “pNUMA” to refer to a physical NUMA domain, and “vNUMA” to refer to a virtualized NUMA domain.
 
-## Performance and topology
-
-Maximizing high performance compute (HPC) application performance on AMD EPYC requires a thoughtful approach memory locality and process placement. Below we outline the AMD EPYC architecture and our implementation of it on Azure for HPC applications. We will use the term “pNUMA” to refer to a physical NUMA domain, and “vNUMA” to refer to a virtualized NUMA domain. Physically, a HB-series is 2 * 32-core EPYC 7551 CPUs for a total of 64 physical cores. These 64 cores are divided into 16 pNUMA domains (8 per socket), each of which is four cores and known as a “CPU Complex” (or “CCX”). Each CCX has its own L3 cache, which is how an OS will see a pNUMA/vNUMA boundary. A pair of adjacent CCXs shares access to two channels of physical DRAM (32 GB of DRAM in HB-series servers).
+Physically, an HB-series is 2 * 32-core EPYC 7551 CPUs for a total of 64 physical cores. These 64 cores are divided into 16 pNUMA domains (8 per socket), each of which is four cores and known as a “CPU Complex” (or “CCX”). Each CCX has its own L3 cache, which is how an OS will see a pNUMA/vNUMA boundary. A pair of adjacent CCXs shares access to two channels of physical DRAM (32 GB of DRAM in HB-series servers).
 
 To provide room for the Azure hypervisor to operate without interfering with the VM, we reserve physical pNUMA domain 0 (the first CCX). We then assign pNUMA domains 1-15 (the remaining CCX units) for the VM. The VM will see:
 
@@ -33,9 +31,11 @@ Process pinning will work on HB-series VMs because we expose the underlying sili
 
 See more on [AMD EPYC architecture](https://bit.ly/2Epv3kC) and [multi-chip architectures](https://bit.ly/2GpQIMb) on LinkedIn. For more detailed information, see the [HPC Tuning Guide for AMD EPYC Processors](https://bit.ly/2T3AWZ9).
 
-See the [HC-series](https://docs.microsoft.com/azure/virtual-machines/linux/sizes-hpc) VM size article for detailed specifications.
+Topology of EPYC 2P Server
+
+![Topology of EPYC 2P Server](./media/hb-series-overview/dual-socket.png)
 
-The following diagram represents the segregation of cores reserved for Azure Hypervisor and for the HB-series VM.
+Segregation of cores reserved for Azure Hypervisor and HB-series VM
 
 ![Segregation of cores reserved for Azure Hypervisor and HB-series VM](./media/hb-series-overview/segregation-cores.png)
 
@@ -65,6 +65,6 @@ The following diagram represents the segregation of cores reserved for Azure Hyp
 
 ## Next steps
 
-* Learn more about HPC VM sizes for [Linux](https://docs.microsoft.com/azure/virtual-machines/linux/sizes-hpc) and [Windows](https://docs.microsoft.com/azure/virtual-machines/windows/sizes-hpc) in Azure.
+* Learn more about HPC VM sizes for [Linux](../linux/sizes-hpc.md) and [Windows](../windows/sizes-hpc.md) in Azure.
 
 * Learn more about [HPC](https://docs.microsoft.com/azure/architecture/topics/high-performance-computing/) in Azure.

articles/virtual-machines/workloads/hpc/hc-series-overview.md

@@ -1,5 +1,5 @@
 ---
-title: Preview of the HC-series - Azure Virtual Machines | Microsoft Docs
+title: Preview of the HC-series in Azure | Microsoft Docs
 description: Learn about the preview support for the HC-series VM size in Azure. 
 services: virtual-machines
 documentationcenter: ''
@@ -11,16 +11,12 @@ tags: azure-resource-manager
 ms.service: virtual-machines
 ms.workload: infrastructure-services
 ms.topic: article
-ms.date: 05/15/2019
+ms.date: 05/07/2019
 ms.author: amverma
 ---
 
 # HC-Series overview
 
-This article provides recommended practices, observed patterns, and recipes to get started with High Performance Computing (HPC) workloads on Azure HC-series VMs.
-
-## Performance and topology
-
 Maximizing HPC application performance on Intel Xeon Scalable Processors requires a thoughtful approach to process placement on this new architecture. Here, we outline our implementation of it on Azure HC-series VMs for HPC applications. We will use the term “pNUMA” to refer to a physical NUMA domain, and “vNUMA” to refer to a virtualized NUMA domain. Similarly, we will use the term “pCore” to refer to physical CPU cores, and “vCore” to refer to virtualized CPU cores.
 
 Physically, an HC server is 2 * 24-core Intel Xeon Platinum 8168 CPUs for a total of 48 physical cores. Each CPU is a single pNUMA domain, and has unified access to six channels of DRAM. Intel Xeon Platinum CPUs feature a 4x larger L2 cache than in prior generations (256 KB/core -> 1 MB/core), while also reducing the L3 cache compared to prior Intel CPUs (2.5 MB/core -> 1.375 MB/core).
@@ -31,13 +27,13 @@ The above topology carries over to the HC-series hypervisor configuration as wel
 
 The VM has no knowledge that pCores 0-1 and 24-25 weren't given to it. Thus, it exposes each vNUMA as if it natively had 22 cores.
 
-Intel Xeon Platinum, Gold, and Silver CPUs also introduce an on-die 2D mesh network for communication within and external to the CPU socket. We strongly recommend process pinning for optimal performance and consistency. Process pinning will work on HC-series VMs because the underlying silicon is exposed as-is to the guest VM.
+Intel Xeon Platinum, Gold, and Silver CPUs also introduce an on-die 2D mesh network for communication within and external to the CPU socket. We strongly recommend process pinning for optimal performance and consistency. Process pinning will work on HC-series VMs because the underlying silicon is exposed as-is to the guest VM. More on Intel Xeon SP architecture at: https://bit.ly/2RCYkiE
 
-For more detailed information, see the [Intel Xeon SP architecture](https://bit.ly/2T3AWZ9) overview.
+Topology of Intel Xeon SP Server
 
-See the [HC-series](https://docs.microsoft.com/azure/virtual-machines/linux/sizes-hpc) VM size article for detailed specifications.
+![Topology of Intel Xeon SP Server](./media/hc-series-overview/xeon-topology.png)
 
-The following diagram represents the segregation of cores reserved for Azure Hypervisor and for the HC-series VM.
+Segregation of cores reserved for Azure Hypervisor and HC-series VMs
 
 ![Segregation of cores reserved for Azure Hypervisor and HC-series VM](./media/hc-series-overview/segregation-cores.png)
 
@@ -66,7 +62,7 @@ The following diagram represents the segregation of cores reserved for Azure Hyp
 | Azure Batch Support         | Yes                         |
 
 ## Next steps
+* Learn more about HPC VM sizes for [Linux](https://docs.microsoft.com/en-us/azure/virtual-machines/linux/sizes-hpc) and [Windows](https://docs.microsoft.com/en-us/azure/virtual-machines/windows/sizes-hpc) in Azure.
 
-* Learn more about HPC VM sizes for [Linux](https://docs.microsoft.com/azure/virtual-machines/linux/sizes-hpc) and [Windows](https://docs.microsoft.com/azure/virtual-machines/windows/sizes-hpc) in Azure.
-
+https://docs.microsoft.com/en-us/azure/virtual-machines/windows/sizes-hpc
 * Learn more about [HPC](https://docs.microsoft.com/azure/architecture/topics/high-performance-computing/) in Azure.

articles/virtual-machines/workloads/hpc/toc.yml

@@ -1,4 +1,4 @@
-- name: High-perfomance Computing
+- name: High perfomance Computing
   items:
   - name: Overview
     href: https://docs.microsoft.com/azure/architecture/topics/high-performance-computing/

articles/virtual-machines/workloads/workloads.md

@@ -13,14 +13,14 @@ ms.topic: article
 ms.workload: infrastructure-services
 ms.tgt_pltfrm: vm-linux
 ms.devlang: azurecli
-ms.date: 02/21/2019
+ms.date: 06/06/2019
 ms.author: cynthn
 ---
 # Workloads on Azure Virtual Machines
 
 Virtual machines in Azure can support many workloads. Each workload can have specialized guidance around the optimal design for deployments on VMs in Azure. Here is some of the guidance we can offer for specialized workloads.
 
-- [High-performance computing] (https://docs.microsoft.com/azure/architecture/topics/high-performance-computing/)
+- [High performance computing](https://docs.microsoft.com/azure/architecture/topics/high-performance-computing/)
 - [SAP](./sap/get-started.md) in Azure can run your mission critical SAP workloads and scenarios on a scalable, compliant, and enterprise-proven platform.
 - [Oracle](./oracle/oracle-considerations.md)  supports running Oracle DB 12.1 Standard and Enterprise editions in Azure on virtual machine images based on Oracle Linux. 
 - [Mainframe rehosting](./mainframe-rehosting/overview.md)