Compute Hardware Overview: Difference between revisions
Jump to navigation
Jump to search
mNo edit summary |
Updated specs |
||
| Line 1: | Line 1: | ||
Anunna's compute nodes are grouped into [[Partitions / Queues | partitions]] by the kind of hardware they provide. This page summarises what each node type offers; for how to request a particular node or feature in a job, see [[Choosing a node (constraints)]]. | Anunna's compute nodes are grouped into [[Partitions / Queues | partitions]] by the kind of hardware they provide. This page summarises what each node type offers; for how to request a particular node or feature in a job, see [[Choosing a node (constraints)]]. | ||
| Line 7: | Line 5: | ||
Standard compute nodes for jobs that do not need a GPU. These make up the <code>main</code> partition and handle the bulk of the cluster's work. | Standard compute nodes for jobs that do not need a GPU. These make up the <code>main</code> partition and handle the bulk of the cluster's work. | ||
We have 3 generations in use right now: | We have 3 generations in use right now: | ||
* gen2 : 29 nodes with Intel CPUs, 32 cores and 375G RAM. 2 nodes with Intel CPUs, 64 cores and 1.5T RAM | * gen2 : 29 nodes with Intel CPUs, 32 cores and 375G RAM. 2 nodes with Intel CPUs, 64 cores and 1.5T RAM | ||
* gen3: 73 nodes with AMD CPUs, 128 cores and 1T RAM. 2 nodes with AMD CPUs, 96 cores and 4T RAM | * gen3: 73 nodes with AMD CPUs, 128 cores and 1T RAM. 2 nodes with AMD CPUs, 96 cores and 4T RAM | ||
* gen4: 3 nodes with AMD CPUSs, 192 cores and 2. | * gen4: 3 nodes with AMD CPUSs, 192 cores and 2.25T RAM | ||
== GPU nodes == | == GPU nodes == | ||
Nodes equipped with GPUs, for accelerated workloads such as deep learning. Request a GPU with <code>--gres=gpu:<n></code> and the appropriate partition (see [[Choosing a node (constraints)]]). | Nodes equipped with GPUs, for accelerated workloads such as deep learning. Request a GPU with <code>--gres=gpu:<n></code> and the appropriate partition (see [[Choosing a node (constraints)]]). | ||
These also come in different flavours: | |||
=== NVIDIA GPUs === | === NVIDIA GPUs === | ||
| Line 22: | Line 21: | ||
Available in the <code>gpu</code> partition. The cluster has several NVIDIA GPU models — A100, A6000, and V100. The scheduler hands out A100s first, then A6000s, then V100s; the price per GPU-hour is the same for all of them. As a rough guide, the A100 (80 GB) is about twice as fast as the A6000 (48 GB) or V100 (16 GB), depending on whether your workload can use the extra memory. Constrain a job to a specific model with, for example, <code>--constraint='nvidia&A100'</code>. | Available in the <code>gpu</code> partition. The cluster has several NVIDIA GPU models — A100, A6000, and V100. The scheduler hands out A100s first, then A6000s, then V100s; the price per GPU-hour is the same for all of them. As a rough guide, the A100 (80 GB) is about twice as fast as the A6000 (48 GB) or V100 (16 GB), depending on whether your workload can use the extra memory. Constrain a job to a specific model with, for example, <code>--constraint='nvidia&A100'</code>. | ||
* gpu100: 32 cores, 375G RAM, 4 Nvidia V100 GPUs with 16G VRAM | |||
* gpun20{0,1,2,3): 32 cores, 500G RAM, each has 4 Nvidia A100 with 80G VRAM | |||
* gpuxn200: 64 cores, 500G RAM, 2 Nvidia RTX A6000 with 48G VRAM | |||
* node301: 192 cores, 2.25T RAM, 4 Nvidia L40S with 46G VRAM | |||
=== AMD GPUs === | === AMD GPUs === | ||
| Line 28: | Line 30: | ||
Available in the <code>gpu_amd</code> partition. Requested the same way as NVIDIA GPUs, using the <code>gpu_amd</code> partition. | Available in the <code>gpu_amd</code> partition. Requested the same way as NVIDIA GPUs, using the <code>gpu_amd</code> partition. | ||
* gpua201: 32 cores, 375G RAM, 4 AMD MI210 with 64G VRAM<br /> | |||
== See also == | == See also == | ||
Latest revision as of 12:24, 9 July 2026
Anunna's compute nodes are grouped into partitions by the kind of hardware they provide. This page summarises what each node type offers; for how to request a particular node or feature in a job, see Choosing a node (constraints).
CPU nodes
Standard compute nodes for jobs that do not need a GPU. These make up the main partition and handle the bulk of the cluster's work.
We have 3 generations in use right now:
- gen2 : 29 nodes with Intel CPUs, 32 cores and 375G RAM. 2 nodes with Intel CPUs, 64 cores and 1.5T RAM
- gen3: 73 nodes with AMD CPUs, 128 cores and 1T RAM. 2 nodes with AMD CPUs, 96 cores and 4T RAM
- gen4: 3 nodes with AMD CPUSs, 192 cores and 2.25T RAM
GPU nodes
Nodes equipped with GPUs, for accelerated workloads such as deep learning. Request a GPU with --gres=gpu:<n> and the appropriate partition (see Choosing a node (constraints)).
These also come in different flavours:
NVIDIA GPUs
Available in the gpu partition. The cluster has several NVIDIA GPU models — A100, A6000, and V100. The scheduler hands out A100s first, then A6000s, then V100s; the price per GPU-hour is the same for all of them. As a rough guide, the A100 (80 GB) is about twice as fast as the A6000 (48 GB) or V100 (16 GB), depending on whether your workload can use the extra memory. Constrain a job to a specific model with, for example, --constraint='nvidia&A100'.
- gpu100: 32 cores, 375G RAM, 4 Nvidia V100 GPUs with 16G VRAM
- gpun20{0,1,2,3): 32 cores, 500G RAM, each has 4 Nvidia A100 with 80G VRAM
- gpuxn200: 64 cores, 500G RAM, 2 Nvidia RTX A6000 with 48G VRAM
- node301: 192 cores, 2.25T RAM, 4 Nvidia L40S with 46G VRAM
AMD GPUs
Available in the gpu_amd partition. Requested the same way as NVIDIA GPUs, using the gpu_amd partition.
- gpua201: 32 cores, 375G RAM, 4 AMD MI210 with 64G VRAM