Datacenter88· High Benchmark verified8 of 8 available
8× NVIDIA H100 SXM
Ashburn, US · us-eastby Graniteline Systems
- VRAM / GPU
- 80 GB
- vCPU
- 192 cores
- System RAM
- 2,048 GB
- NVMe disk
- 30,720 GB
- Bandwidth
- 100,000 Mbps
- Architecture
- Hopper
- FP16 reference
- 989 TFLOPS
- GPUs on machine
- 8
Price history
MZG per GPU-hour set by the provider
810
+8.0% vs. catalogue reference 750
Benchmark
993FP16 TFLOPS measured
100.4% of the 989 TFLOPS catalogue reference · accepted band 70–115%
Last benchmark4d ago
VerificationPassed
Spot checks2% canary jobs
Provider
Graniteline Systems
0xeA0E2B…930177
Reliability score87.7 / 100
Uptime · 30d100.00%
Jobs completed260
Member sinceNov 11, 2026
Recent jobs on this offer
anonymised| Template | Duration | Status | Ended |
|---|---|---|---|
| ComfyUI + SDXL / FLUX.1-schnell | 1h 29m | completed | 1h ago |
| PyTorch 2.5 + CUDA 12.4 JupyterLab | 4h 53m | completed | 1h ago |
| Llama 3.1 8B Instruct | 4h 0m | completed | 1h ago |
| vLLM OpenAI-compatible | 56m 1s | completed | 3h ago |
| PyTorch 2.5 + CUDA 12.4 JupyterLab | 12h 57m | completed | 3h ago |
| vLLM OpenAI-compatible | 2h 11m | completed | 6h ago |
| Llama 3.1 8B Instruct | 18m 45s | completed | 8h ago |
| vLLM OpenAI-compatible | 2h 10m | completed | 11h ago |
| vLLM OpenAI-compatible | 0s | cancelled | 21h ago |
| vLLM OpenAI-compatible | 43m 3s | completed | 21h ago |