Specifications
AI Processor
1x Huawei Ascend 310P (Da Vinci architecture)
Form Factor
Half-height half-length PCIe card
Host Interface
PCIe 4.0 x16
AI Performance
140 TOPS INT8 / 70 TFLOPS FP16
Memory
24GB LPDDR4X with ECC
Memory Bandwidth
204 GB/s
Video Decode
H.264/H.265 hardware decode, up to 32-ch 1080p30
Video Encode
H.264/H.265 hardware encode, up to 2-ch 1080p30
Precision Support
INT8, FP16
Power Consumption
Up to 72W
Cooling
Passive heatsink
Software Stack
Huawei CANN, MindSpore, AscendCL, PyTorch Adapter (torch-npu)
Operating System
openEuler, EulerOS, Ubuntu 20.04/22.04 LTS
Certifications
CE, FCC, RoHS
Overview
The Huawei Atlas 300I Pro is the single-chip counterpart to the Atlas 300I Duo, using one Ascend 310P processor to deliver 140 TOPS INT8 and 70 TFLOPS FP16. It is a compact, power-efficient inference card suited to dense multi-card deployment where accelerator power and thermal budgets are tight.
With 24GB of ECC LPDDR4X at 204 GB/s, the 300I Pro handles computer vision, video structuring, OCR and mid-size model inference while remaining within Huawei's low-power envelope. It slots into CANN-supported servers and edge platforms and shares the Ascend software stack with the rest of the Atlas family.
QS Compute supplies the Atlas 300I Pro for intelligent video surveillance, industrial quality inspection and edge inference deployments with a preference for high energy efficiency.
Key Benefits
140 TOPS INT8 from a single Ascend 310P. Low ~72W power envelope for dense, energy-efficient nodes. 24GB ECC LPDDR4X at 204 GB/s. Shared Ascend software stack with CANN and MindSpore.
Applications
Intelligent video surveillance and analytics, industrial and machine-vision inference, OCR and document processing, edge AI appliances, and datacenter inference scale-out.
Request a Quote — HUAWEI ATLAS 300I PRO — ASCEND 310P 24GB AI INFERENCE CARD
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →