Huawei Atlas · Accelerator Card
Atlas 300I A2 Inference Card
The Atlas 300I A2 inference card delivers strong inference performance for generative LLM inference and content moderation, supporting up to 560 TOPS INT8 and 280 TFLOPS FP16.
Key Highlights
- Leading 560 TOPS INT8 and 280 TFLOPS FP16 with an 8-core 2.0 GHz CPU
- up to 1.6 TB/s memory bandwidth
- PCIe 5.0 interface.
Technical Specifications
| AI Compute Power | 560 TOPS INT8, 280 TFLOPS FP16 |
|---|---|
| On-chip Memory | 32GB (0.8TB/s) or 64GB (1.6TB/s), ECC |
| Interconnect | PCIe Gen5.0 |
| CPU | 8-core 2.0 GHz |
| TDP | max 350W |
| Form Factor | dual-slot full-length full-height PCIe card |
| Operating Temp | 5℃~45℃ |
Typical Use Cases
Generative LLM inference, content moderation and data-center AI inference.
翻译稿；官网未公开FP8/FP4精度算力数值；img1: main; 官网无独立公开技术白皮书链接。
Related Products
Huawei Atlas
Atlas 200I DK A2 Developer Kit
Easy-to-use AI development kit for fast innovation View details →
Huawei Atlas
Atlas 200I A2 Acceleration Module
Edge AI accelerator module embedded in smart devices, robots and drones View details →
Huawei Atlas
Atlas 300I Pro Inference Card
Inference accelerator card integrated into servers View details →
Huawei Atlas