NOTE This system currently has limited availability.
El Capitan resides in the CZ while undergoing usability testing. When fully deployed, it will be moved to the SCF (see documentation on the LC Zones page).
A comparison of ATS-4 machines
| El Capitan | Tuolumne | El Dorado* | RZAdams | |
|---|---|---|---|---|
| Nodes | 11,136 | 1,152 | 384 | 128 |
| MI300As per node | 4 | 4 | 4 | 4 |
| MI300As total | 44,544 | 4,608 | 1,536 | 512 |
| Node Peak (DP TFLOP/s) | 250.8 | 250.8 | 250.8 | 250.8 |
| System Peak (DP PFLOP/s) | 2,792.9 | 288.4 | 96.3 | 32.1 |
| System Peak (SP PFLOP/s) | 3,688.2 | 381.5 | 127.2 | 42.4 |
| System Peak (HP PFLOP/s) | 17,639.4 | 1,824.8 | 608.2 | 202.8 |
| Node Memory (GiB) (All HBM3) | 512 | 512 | 512 | 512 |
| System Memory (TiB) | 5,568 | 576 | 192 | 64 |
| Compute cabinets | 87 | 9 | 3 | 1 |
| Peak Power (MW) | 34.8 | 3.6 | 1.2 | 0.4 |
| Total Rabbit Modules | 696 | 72 | 24 | 8 |
| November 2024 Top500 position | 1 | 10 | 20 | 49 |
*El Dorado is sited at Sandia National Labs
El Capitan In-Depth
| Zone |
SCF
|
| Vendor |
HPE Cray
|
| User-Available Nodes |
Login Nodes*
32 nodes: elcap[1001-1016,12121-12136]
Batch Nodes
11,039
Debug Nodes
64
Total Nodes
11,103
|
| APUs |
APU Architecture
AMD MI300A
|
| CPUs |
CPU Architecture
4th Generation AMD EPYC
Cores/Node
96
Total Cores
1,065,888
|
| GPUs |
GPU Architecture
CDNA 3
Total GPUs
44,412
GPUs per compute node
4
GPU peak performance (TFLOP/s double precision)
68.00
GPU global memory (GB)
512.00
|
| Memory Total (GB) |
5,684,736
|
| Clock Speed (GHz) |
2.0
|
| OS |
TOSS 4
|
| Interconnect |
HPE Slingshot 11
|
| Parallel job type |
multiple nodes per job
|
| Recommended location for parallel file space |
/p/lustre4
|
| Class |
ATS-4, CORAL-2
|
| Password Authentication |
OTP, Kerberos, ssh keys
|
| Year Commissioned |
2024
|
| Compilers |
See Compilers page or the El Capitan Systems user guide |
| Documentation |