Aiserver Manufacturer
1 / 5
| Introduction | The Atlas Atlas 300V 24G is a domestically produced Ascend series dual-chip AI inference acceleration card, designed specifically for large language models (LLM) and video analysis scenarios. It offers up to 96GB of massive memory and 280 TOPS of INT8 computing power, delivering a cost-performance ratio far superior to similar products. |
| Name | Atlas 300V 24G |
| INT8 | 100 TOPS |
| FP16 | 50 TFLOPS |
| VRAM type | HBM2E (High Bandwidth Memory, with approximately 30% higher bandwidth compared to HBM2, suitable for the high bandwidth demands of large model inference) |
| Chip fabrication process | TSMC 7nm FinFET process, paired with 2.5D CoWoS advanced packaging (ensuring core heat dissipation and signal stability) |
| DirectX support | Compatible with DirectX 12 (including 12_1 features), suitable for mainstream graphics rendering and video processing scenarios |
| Video memory frequency | Effective frequency 3600MHz (native high-frequency characteristics of HBM2E, achieving low-latency data transfer with 800GB/s bandwidth) |
| Memory Bus Width | 2048-bit (HBM2E standard bus width design, providing sufficient data channels for 32GB VRAM to avoid bandwidth bottlenecks) |
| Video memory speed | 0.55 ns (based on the physical characteristics of HBM2E memory chips, corresponding to an equivalent frequency of 4000 MHz, with latency better than GDDR6) |
| Output Interface | No traditional video output interfaces (DP/HDMI/VGA), focused on data center inference/computing scenarios |
| Output interface type | The core interface is PCIe Gen4.0 x16 (bidirectional bandwidth 64GB/s), supporting x86/ARM host interconnection. |
| Form Factor | FHFL dual - slot |
| Weight | 1064g |
