Up to 8 GPUs
BullSequana AI 620
Inference and fine tuning
For inference and fine tuning of small and medium AI models, with NVIDIA or AMD GPU options and air cooling.
Compare the BullSequana AI 620 and AI 640 product families by workload, GPU capacity, processor, accelerator and cooling options.
Up to 8 GPUs
Inference and fine tuning
For inference and fine tuning of small and medium AI models, with NVIDIA or AMD GPU options and air cooling.
8 GPUs
Training and large model inference
For training AI models and running inference with large models, with NVIDIA or AMD accelerators and air or hybrid cooling.
Review the processor, GPU and cooling options available for each server configuration.
| Product family | Configuration | Processor | GPU | Cooling |
|---|---|---|---|---|
| BullSequana AI 620 | 622 A03 | AMD EPYC™ | NVIDIA or AMD GPUs | Air cooling |
| BullSequana AI 640 | 641 A02 or H02 | Intel Xeon | NVIDIA HGX H200 | Air or hybrid cooling |
| BullSequana AI 640 | 642 A01 or H01 | AMD EPYC™ | AMD Instinct™ MI300X | Air or hybrid cooling |
| BullSequana AI 640 | 642 H02 | AMD EPYC™ | AMD Instinct™ MI325X | Hybrid cooling |
| BullSequana AI 640 | 642 H03 | AMD EPYC™ | AMD Instinct™ MI355X | Hybrid cooling |
BullSequana AI 620
BullSequana AI 640
BullSequana AI 640
BullSequana AI 640
BullSequana AI 640
