Skip to content

Commit 0a374f9

Browse files
Merge pull request #241761 from jvenka/docs-editor/hbv4-series-overview-1686876352
Update HBv4 overview page
2 parents 5fe6fab + 139b221 commit 0a374f9

File tree

1 file changed

+10
-16
lines changed

1 file changed

+10
-16
lines changed

articles/virtual-machines/hbv4-series-overview.md

Lines changed: 10 additions & 16 deletions
Original file line numberDiff line numberDiff line change
@@ -67,27 +67,15 @@ lstopo-no-graphics --no-io --no-legend --of txt
6767
![lstopo output for HBv4-176 VM](./media/hpc/architecture/hbv4/hbv4-176-lstopo.png)
6868
</details>
6969

70-
<details>
71-
<summary>Click to view lstopo output for Standard_HB176-144rs_v4</summary>
72-
7370
![lstopo output for HBv4-144 VM](./media/hpc/architecture/hbv4/hbv4-144-lstopo.png)
7471
</details>
7572

76-
<details>
77-
<summary>Click to view lstopo output for Standard_HB176-96rs_v4</summary>
78-
7973
![lstopo output for HBv4-64 VM](./media/hpc/architecture/hbv4/hbv4-96-lstopo.png)
8074
</details>
8175

82-
<details>
83-
<summary>Click to view lstopo output for Standard_HB176-48rs_v4</summary>
84-
8576
![lstopo output for HBv4-32 VM](./media/hpc/architecture/hbv4/hbv4-48-lstopo.png)
8677
</details>
8778

88-
<details>
89-
<summary>Click to view lstopo output for Standard_HB176-24rs_v4</summary>
90-
9179
![lstopo output for HBv4-16 VM](./media/hpc/architecture/hbv4/hbv4-24-lstopo.png)
9280
</details>
9381

@@ -128,14 +116,20 @@ When paired in a striped array, the NVMe SSD provides up to 12 GB/s reads and 7
128116
| Orchestrator Support | Azure CycleCloud, Azure Batch, AKS; [cluster configuration options](sizes-hpc.md#cluster-configuration-options) |
129117

130118
> [!NOTE]
131-
> These VMs support only Generation 2.
132-
> There is no official kernel level support from AMD on CentOS. Support starts at RHEL 8.6 and a derivative of RHEL which is AlmaLinux 8.6.
133-
> Windows Server 2012 R2 is not supported on HBv4 and other VMs with more than 64 (virtual or physical) cores. For more information, see [Supported Windows guest operating systems for Hyper-V on Windows Server](/windows-server/virtualization/hyper-v/supported-windows-guest-operating-systems-for-hyper-v-on-windows). Windows Server 2022 is required for 144 and 176 core sizes, Windows Server 2016 also works for 24, 48, and 96 core sizes, Windows Server works for only 24 and 48 core sizes.
119+
> * These VMs support only Generation 2.
120+
> * There is no official kernel level support from AMD on CentOS. Support starts at RHEL 8.6 and a derivative of RHEL which is AlmaLinux 8.6.
121+
> * Windows Server 2012 R2 is not supported on HBv4 and other VMs with more than 64 (virtual or physical) cores. For more information, see [Supported Windows guest operating systems for Hyper-V on Windows Server](/windows-server/virtualization/hyper-v/supported-windows-guest-operating-systems-for-hyper-v-on-windows). Windows Server 2022 is required for 144 and 176 core sizes, Windows Server 2016 also works for 24, 48, and 96 core sizes, Windows Server works for only 24 and 48 core sizes.
134122
135123
> [!IMPORTANT]
136-
> Recommended image URN: almalinux:almalinux-hpc:8_7-hpc-gen2:8.7.2023060101, To deploy this image over Azure CLI, esnure the following parameters are included **--plan 8_7-hpc-gen2 --product almalinux-hpc --publisher almalinux**. For scaling tests please use the recommended URN along with the new [HPC-X tarball] (https://github.com/Azure/azhpc-images/blob/c8db6de3328a691812e58ff56acb5c0661c4d488/alma/alma-8.x/alma-8.6-hpc/install_mpis.sh#L16).
124+
> Recommended image URN: almalinux:almalinux-hpc:8_7-hpc-gen2:8.7.2023060101, To deploy this image over Azure CLI, ensure the following parameters are included **--plan 8_7-hpc-gen2 --product almalinux-hpc --publisher almalinux**. For scaling tests please use the recommended URN along with the new [HPC-X tarball](https://github.com/Azure/azhpc-images/blob/c8db6de3328a691812e58ff56acb5c0661c4d488/alma/alma-8.x/alma-8.6-hpc/install_mpis.sh#L16).
125+
126+
> [!NOTE]
127+
> * NDR support is added in UCX 1.13 or later. Older UCX versions will report the above runtime error. UCX Error: Invalid active speed `[1677010492.951559] [updsb-vm-0:2754 :0] ib_iface.c:1549 UCX ERROR Invalid active_speed on mlx5_ib0:1: 128`.
128+
> * Ibstat shows low speed (SDR): Older Mellanox OFED (MOFED) versions do not support NDR and it may report slower IB speeds. Please use MOFED versions MOFED 5.6-1.0.3.3 or above.
137129
138130
## Next steps
139131

140132
- Read about the latest announcements, HPC workload examples, and performance results at the [Azure Compute Tech Community Blogs](https://techcommunity.microsoft.com/t5/azure-compute/bg-p/AzureCompute).
141133
- For a higher level architectural view of running HPC workloads, see [High Performance Computing (HPC) on Azure](/azure/architecture/topics/high-performance-computing/).
134+
135+

0 commit comments

Comments
 (0)