ASUS ESC NB8-E11
Overview
Performance Metrics
| Precision | Peak (PFLOPS) | Per GPU (TFLOPS) | TFLOPS/W |
|---|---|---|---|
| FP64 | 0.3 | 37.0 | — |
| FP32 | 0.6 | 75.0 | — |
| TF32 | 8.8 | 1100.0 | — |
| FP16 | 18.0 | 2250.0 | — |
| BF16 | 18.0 | 2250.0 | — |
| FP8 | 36.0 | 4500.0 | — |
| FP4 | 72.0 | 9000.0 | — |
| INT8 | 36.0 | 4500.0 | — |
All figures are dense. Vendors commonly headline the sparse number, which is twice the dense one.
Configuration
- Accelerator
- B200 SXM 180GB
- Count
- 8
- Memory per GPU
- 180 GB
- Bandwidth per GPU
- 7.7 TB/s
Power & Efficiency
- Aggregate Power
- —
- TDP per GPU
- 1000 W
- FP8 Efficiency
- —
- Aggregate Bandwidth
- 64.0 TB/s
System Details
Interconnect
NVIDIA NVLink with NVSwitch, 1,800 GB/s GPU-to-GPU; one-GPU-to-one-NIC topology with up to 8 NICs
Notes
A 10U eight-GPU server built on the NVIDIA Blackwell HGX B200 baseboard, with dual 5th generation Intel Xeon Scalable processors at 350 W and NVLink giving 1,800 GB/s directly between GPUs, twice the bandwidth of the Hopper-generation ESC N8-E11. It uses a dedicated one-GPU-to-one-NIC topology supporting up to eight NICs, and offers up to 11 PCIe slots, 32 DIMM slots, 10 NVMe bays and dual 10Gb LAN, with 5+1 80 PLUS Titanium supplies. Unlike many B200 platforms it is air-cooled, using dedicated CPU and GPU airflow tunnels, which is what accounts for the 10U height. ASUS publishes no FLOPS figure of its own, so the compute shown is NVIDIA's for the HGX B200 baseboard, including the sparse figures NVIDIA publishes for Blackwell. ASUS's detailed specification tab does not render server-side, so chassis dimensions, weight and power supply wattages could not be read and are deliberately absent.