The META Lab operates a dedicated HPC cluster consisting of 40 CPU compute nodes and 1 GPU workstation node, along with individually assigned research laptops — providing computing resources across all scales of our computational and autonomous experimental workflows. Total theoretical peak: ~172.0 TFLOPS CPU (FP64) and ~207.2 TFLOPS GPU (FP32), with significant expansion incoming.
Lab-owned servers
40
CPU Nodes
2,164
Total CPU Cores
~172.0
TFLOPS CPU FP64 (peak)
4
GPU Units
~207.2
TFLOPS GPU FP32 (peak)
Computing Resources
CPU Cluster
| Count | CPU Model | Cores | RAM | Peak FP64 / node | Subtotal FP64 |
|---|---|---|---|---|---|
| 23 | Intel Xeon Gold 6342 * 2 | 48 | 188.2 GB | 4.30 TFLOPS | 98.9 TFLOPS |
| 7 | Intel Xeon Gold 6342 * 2 | 48 | 251.4 GB | 4.30 TFLOPS | 30.1 TFLOPS |
| 10 | Intel Xeon Gold 6530 * 2 | 64 | 503.4 GB | 4.30 TFLOPS | 43.0 TFLOPS |
| 40 | Total | ~172.0 TFLOPS | |||
GPU Cluster
| Units | GPU Model | VRAM | Peak FP32 |
|---|---|---|---|
| 1 | NVIDIA RTX 6000 Ada | 48 GB | 91.1 TFLOPS |
| 3 | NVIDIA RTX A6000 | 48 GB | 38.7 TFLOPS |
| 4 | Total | ~207.2 TFLOPS |
Incoming Equipment
Planned for H2 2026. Details subject to change.
| Units | Status | GPU Model | Architecture | VRAM |
|---|---|---|---|---|
| 2 | Planned for H2 2026 | NVIDIA H200 | Blackwell | 141 GB |
A high-end GPU cluster is planned for order in H2 2026. The exact model and specifications are not yet finalized. Intended for LLM-based scientific computing and autonomous laboratory research.
Personal Laptops
| Device Type | GPU Range | VRAM | Assigned To |
|---|---|---|---|
| Research Laptop (per member) | RTX 3080 – RTX 5070 | 8 – 16 GB | Individual lab members |
KENTECH GPU Cluster
Korea Institute of Energy Technology (KENTECH) · Free access.
| Units | GPU Model | VRAM | Peak FP32 | Peak FP64 | Reservation Unit |
|---|---|---|---|---|---|
| 8 | NVIDIA H100 80GB SXM5 | 80 GB HBM3 | 67.0 TFLOPS | 34.0 TFLOPS | 1 GPU / 28 cores / 256GB RAM |
| 8 | NVIDIA H100 80GB SXM5 | 80 GB HBM3 | 67.0 TFLOPS | 34.0 TFLOPS | 0.2 GPU / 6 cores / 50GB RAM |
| 16 | Total | ~1,072 TFLOPS | ~544 TFLOPS |
META Lab members also have access to KISTI supercomputing resources for large-scale HPC workloads.
Autonomous Laboratory
Physical automation systems for self-driving experiments and high-throughput materials discovery.
Universal Robots
Collaborative Robotic Arm · 6-DOF- Autonomous sample preparation, manipulation & handling for self-driving experiments
- Standardization of experimental procedures and automated synthesis workflows