| Name | Node type | Node Count | Core Manufacturer | Core Type | Cores per node | Total Cores | Memory¹ | GPU | GPUs per node | Total GPUs |
|---|---|---|---|---|---|---|---|---|---|---|
| login1 | login | 1 | Intel | Haswell | 24 | 24 | 256 GB | N/A | 0 | 0 |
| login2 | login | 1 | Intel | Haswell | 20 | 20 | 256 GB | N/A | 0 | 0 |
| sn[001-056] | CPU compute | 156 | Intel | Haswell | 28 | 4,368 | 128 GB | N/A | 0 | 0 |
| cn-nvidia | GPU compute | 1 | Intel | Haswell | 12 | 12 | 64 GB | Nvidia P100 | 2 | 2 |
| sn-nvda[1,2] | GPU compute | 2 | Intel | Haswell | 28 | 56 | 128 GB | Nvidia V100 32GB | 2 | 4 |
| sn-nvda[3-8] | GPU compute | 6 | Intel | Haswell | 28 | 168 | 128 GB | Nvidia K80 | 4 | 24 |
| milan[1,2] | login | 2 | AMD | Milan | 64 | 128 | 512 GB | N/A | 0 | 0 |
| xeonmax | login | 1 | Intel | Sapphire Rapids | 96 | 96 | 512 GB | N/A | 0 | 0 |
| dg[001-48] | CPU compute | 48 | AMD | Milan | 96 | 4,608 | 256 GB | N/A | 0 | 0 |
| xm[001-044,049-094] | CPU compute | 90 | Intel | Sapphire Rapids | 96 | 8,640 | 128 GB HBM + 256 GB DDR5 | N/A | 0 | 0 |
| xm[045-048] | CPU compute | 4 | Intel | Sapphire Rapids | 96 | 384 | 1 TB DDR5 | N/A | 0 | 0 |
| dn[001-064] | CPU compute | 64 | Intel | Skylake | 40 | 2,560 | 192 GB | N/A | 0 | 0 |
| a100-[01-11] | GPU compute | 11 | Intel | Ice Lake | 64 | 704 | 256 GB | Nvidia A100 80GB | 4 | 44 |
| dg-mem | large memory | 1 | Intel | Cooper Lake | 96 | 96 | 3 TB | AMD MI210 | 2 | 2 |
| cn-mem | large memory | 1 | Intel | Haswell | 72 | 72 | 3 TB | Nvidia V100 16GB | 1 | 1 |
¹A small subset of node memory is reserved for the OS and file system and is not available for user applications.
Understanding SeaWulf's Hardware
SeaWulf is a high-performance computing (HPC) cluster built with cutting-edge components from AMD, Dell, HPE, IBM, Intel, Nvidia, and other leading technology providers. Its name is a blend of "Seawolf" and "Beowulf," referencing one of the earliest HPC clusters.
The cluster consists of over 400 nodes and approximately 23,000 cores, delivering a peak performance of around 1.86 PFLOP/s for research computations. The nodes are interconnected via high-speed InfiniBand® (IB) networks by Nvidia®, enabling data transfer speeds ranging from 5 to 50 gigabytes per second.
Computing Nodes
- 164 compute nodes from Penguin Computing: Each equipped with two Intel Xeon E5-2683v3 CPUs ("Haswell"), offering 14 cores per CPU at a base speed of 2.0 GHz on an FDR InfiniBand network.
- 8 of these nodes contain GPUs:
- 28 Nvidia Tesla K80 24GB Accelerators, totaling 159,744 CUDA cores (64x GK210 cores).
- One node with 2 Tesla P100 GPUs.
- One node with 2 Tesla V100 GPUs.
- Each node has 128 GB of DDR4 RAM (20 GB reserved for the system) with a combined memory bandwidth of 133.25 GB/s.
- 8 of these nodes contain GPUs:
- 64 compute nodes from Dell: Each with two Intel Xeon Gold 6148 CPUs ("Skylake"), offering 20 cores per CPU at a base speed of 2.4 GHz and 192 GB of RAM, connected via an FDR InfiniBand network.
- 48 compute nodes from HPE: Each equipped with two AMD 7643 CPUs ("Milan"), featuring 48 cores per CPU at a base speed of 3.2 GHz and 256 GB of RAM, connected via an HDR100 InfiniBand network.
- 11 GPU compute nodes from Dell: Each featuring 4x Nvidia A100 80GB GPUs and two Intel 6338 CPUs ("Ice Lake") with 32 cores per CPU at a base speed of 2.0 GHz and 256 GB of RAM, connected via an HDR100 InfiniBand network.
- 94 compute nodes from HPE: Each equipped with two Intel Xeon Max 9468 CPUs ("Sapphire Rapids w/ HBM"), offering 48 cores per CPU at a base speed of 2.6 GHz, with 256 GB of DDR5 RAM plus 128 GB of HBM2 RAM, connected via an NDR InfiniBand network.
Large Memory Nodes
SeaWulf also features two large memory nodes with 3 TB of RAM:
- DDR4 System: 4 Intel E7-8870v3 processors, each with 18 cores at 2.1 GHz, for a total of 72 cores and 144 threads (via Hyper-Threading).
- DDR5 System: 4 Intel 8360H processors, each with 24 cores at 3.0 GHz, for a total of 96 cores.
Storage System
SeaWulf's storage infrastructure is built on a GPFS storage solution, offering a total capacity of approximately 4 petabytes of SAS spinning disk and 50 terabytes of SSD. The SSD component provides exceptional performance, capable of sustaining over 4 million 4K random read IOPS and achieving sequential read speeds exceeding 36 gigabytes per second.
Connectivity
SeaWulf features five login nodes for user access and task management. The nodes are interconnected via high-speed InfiniBand® networks by Nvidia®, supporting data transfer speeds of 5-50 GB/s.
Performance Calculation
- Peak Performance: Cores per node × number of nodes × base clock per core × double precision FLOPS per cycle + Nvidia GPU Teraflops.
- Memory Bandwidth Calculation: Memory Clock (1,066 MHz) × 2 (DDR) × 64-bit Memory Bus width × 4 Memory interfaces per CPU (Quad-channel) × 2 CPUs per node / 8 bits per byte.