Skip to content

Commit d991104

Browse files
American-DipperAmerican-Dipper
authored
Merge pull request #275596 from mattmcinnes/pr/275592
ND MI300X v5 docs (re-based into release branch)
2 parents 5d54915 + c097def commit d991104

File tree

5 files changed

+93
-2
lines changed

5 files changed

+93
-2
lines changed

‎articles/virtual-machines/TOC.yml

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -351,6 +351,8 @@
351351
href: ndv2-series.md
352352
- name: ND-H100-v5-series
353353
href: nd-h100-v5-series.md
354+
- name: ND-MI300X-v5-series
355+
href: ./sizes/gpu-accelerated/nd-mi300x-v5-series.md
354356
- name: NGads V620-series
355357
href: ngads-v-620-series.md
356358
- name: NV-series

‎articles/virtual-machines/sizes/gpu-accelerated/includes/nd-mi300x-v5-series-specs.md

Lines changed: 20 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,20 @@
1+
---
2+
title: ND MI300X v5-series specs include
3+
description: Include file containing specifications of ND MI300X v5-series virtual machine (VM) sizes.
4+
services: virtual-machines
5+
author: marccharest
6+
ms.topic: include
7+
ms.service: virtual-machines
8+
ms.subservice: sizes
9+
ms.date: 05/21/2024
10+
ms.author: marccharest
11+
ms.custom: include file
12+
---
13+
| Part | Quantity <br>Count Units | Specs <br>SKU ID, Performance Units, etc. |
14+
|---|---|---|
15+
| Processor | 96 vCores | Intel® Xeon® Scalable (Sapphire Rapids) |
16+
| Memory | 1850 GiB | |
17+
| Local Storage | 1 Disk | 1000 GiB
18+
| Remote Disks | 32 Disks | 40800 IOPS <br> 612 MBps |
19+
| Network | 8 NICs | 80000 Mbps |
20+
| Accelerators | 8 GPUs | AMD MI300X 192 GiB <br> 1535 GiB per VM |

‎articles/virtual-machines/sizes/gpu-accelerated/includes/nd-mi300x-v5-series-summary.md

Lines changed: 13 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,13 @@
1+
---
2+
title: ND MI300X v5-series summary include
3+
description: Include file containing a summary of the ND MI300X v5-series size family.
4+
services: virtual-machines
5+
author: marccharest
6+
ms.topic: include
7+
ms.service: virtual-machines
8+
ms.subservice: sizes
9+
ms.date: 05/21/2024
10+
ms.author: marccharest
11+
ms.custom: include file
12+
---
13+
The ND MI300X v5 series virtual machine (VM) is a new flagship addition to the Azure GPU family. It was designed for high-end Deep Learning training and tightly coupled scale-up and scale-out Generative AI and HPC workloads. The ND MI300X v5 series VM starts with eight AMD Instinct MI300 GPUs and two fourth Gen Intel Xeon Scalable processors for a total 96 physical cores. Each GPU within the VM is then connected to one another via 4th-Gen AMD Infinity Fabric links with 128 GB/s bandwidth per GPU and 896 GB/s aggregate bandwidth. ND MI300X v5-based deployments can scale up to thousands of GPUs with 3.2 Tb/s of interconnect bandwidth per VM. Each GPU within the VM is provided with its own dedicated, topology-agnostic 400 Gb/s NVIDIA Quantum-2 CX7 InfiniBand connection. These connections are automatically configured between VMs occupying the same virtual machine scale set, and support GPUDirect RDMA. These instances provide excellent performance for many AI, ML, and analytics tools that support GPU acceleration "out-of-the-box," such as TensorFlow, Pytorch, and other frameworks. Additionally, the scale-out InfiniBand interconnect supports a large set of existing AI and HPC tools that are built on AMD’s ROCm Communication Collectives Library (RCCL) for seamless clustering of GPUs.

‎articles/virtual-machines/sizes/gpu-accelerated/nd-family.md

Lines changed: 9 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,5 @@
11
---
2-
title: ND sub-family VM size series
2+
title: ND sub-family virtual machine size series
33
description: Overview of the 'ND' sub-family of virtual machine sizes
44
author: mattmcinnes
55
ms.service: virtual-machines
@@ -9,7 +9,7 @@ ms.date: 04/18/2024
99
ms.author: mattmcinnes
1010
---
1111

12-
# 'ND' sub-family GPU accelerated VM size series
12+
# 'ND' sub-family GPU accelerated virtual machine size series
1313

1414
**Applies to:** :heavy_check_mark: Linux VMs :heavy_check_mark: Windows VMs :heavy_check_mark: Flexible scale sets :heavy_check_mark: Uniform scale sets
1515

@@ -60,5 +60,12 @@ ms.author: mattmcinnes
6060

6161
[!INCLUDE [nd-h100-v5-series-specs](./includes/nd-h100-v5-series-specs.md)]
6262

63+
### ND_MI300X_v5-series
64+
[!INCLUDE [nd-mi300x-v5-series-summary](./includes/nd-mi300x-v5-series-summary.md)]
65+
66+
[View the full ND_MI300X_v5-series page](./nd-mi300x-v5-series.md).
67+
68+
[!INCLUDE [nd-mi300x-v5-series-specs](./includes/nd-mi300x-v5-series-specs.md)]
69+
6370

6471
[!INCLUDE [sizes-footer](../includes/sizes-footer.md)]

‎articles/virtual-machines/sizes/gpu-accelerated/nd-mi300x-v5-series.md

Lines changed: 49 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,49 @@
1+
---
2+
title: ND MI300X v5-series
3+
description: Specifications for the ND MI300X v5-series VMs
4+
author: charest
5+
ms.author: marccharest
6+
ms.reviewer: mattmcinnes
7+
ms.service: virtual-machines
8+
ms.topic: conceptual
9+
ms.date: 05/21/2024
10+
---
11+
12+
# ND MI300X v5-series
13+
14+
**Applies to:** :heavy_check_mark: Linux VMs :heavy_check_mark: Flexible scale sets :heavy_check_mark: Uniform scale sets
15+
16+
[!INCLUDE [Size series summary](./includes/nd-mi300x-v5-series-summary.md)]
17+
18+
## Host specifications
19+
[!INCLUDE [series-specs](./includes/nd-mi300x-v5-series-specs.md)]
20+
21+
## Feature support
22+
[Premium Storage](../../premium-storage-performance.md): Supported<br>
23+
[Premium Storage caching](../../premium-storage-performance.md): Supported<br>
24+
[Ultra disk](../../disks-types.md#ultra-disks): Supported [Learn more](https://techcommunity.microsoft.com/t5/azure-compute/ultra-disk-storage-for-hpc-and-gpu-vms/ba-p/2189312) about availability, usage, and performance <br>
25+
[Live Migration](../../maintenance-and-updates.md): Not Supported<br>
26+
[Memory Preserving Updates](../../maintenance-and-updates.md): Not Supported<br>
27+
[VM Generation Support](../../generation-2.md): Generation 2<br>
28+
[Accelerated Networking](../../../virtual-network/create-vm-accelerated-networking-cli.md): Supported <br>
29+
[Ephemeral OS Disks](../../ephemeral-os-disks.md): Supported <br>
30+
Infiniband: Supported, GPUDirect RDMA, 8x400 Gigabit NDR <br>
31+
NVIDIA NVLink Interconnect: Supported <br>
32+
[Nested Virtualization](/virtualization/hyper-v-on-windows/user-guide/nested-virtualization): Not Supported <br>
33+
<br>
34+
35+
>[!IMPORTANT]
36+
>To get started with ND MI300X v5 VMs, refer to HPC Workload Configuration and Optimization for steps including driver and network configuration. Due to increased GPU memory I/O footprint, the ND MI300X v5 requires the use of Generation 2 VMs and marketplace images.
37+
38+
## Sizes in series
39+
40+
| Size | vCPU | Memory: GiB | Temp storage (SSD) GiB | GPU | GPU Memory GiB | Max data disks | Max uncached disk throughput: IOPS/MBps | Max network bandwidth | Max NICs |
41+
|---------------------|------|------------|------------------------|----------------------------|----------------|----------------|-----------------------------------------|------------------------------|----------|
42+
| Standard_ND96isr_MI300X_v5 | 96 | 1850 | 1000 | 8 MI300X | 80 | 32 | 40800/612 | 80,000 Mbps | 8 |
43+
44+
>[!NOTE]
45+
>The ND MI300X v5 series supports the following kernel version: Ubuntu 20.04: 5.4.0-1046-azure
46+
47+
[!INCLUDE [virtual-machines-common-sizes-table-defs](../../../../includes/virtual-machines-common-sizes-table-defs.md)]
48+
49+
[!INCLUDE [sizes-footer](../includes/sizes-footer.md)]

0 commit comments

Comments
 (0)