Skip to main content

NVIDIA A800 vs Huawei Ascend 910B: Spec Comparison & Buyer's Guide

In AI infrastructure selection, NVIDIA A800 and Huawei Ascend 910B are two accelerators frequently compared. This article contrasts them item by item — architecture, compute, memory, power, and release cadence — to help you quickly judge which fits training or inference workloads.

Spec Comparison Table​

VendorNVIDIA A800Huawei Ascend 910B
VendorNVIDIAHuawei
ArchitectureNVIDIA Ampere (GA100)Da Vinci
ProcessTSMC 7nmTSMC 7nm+
Release Date20222022-11-08
FP8 Compute——
FP16 Compute312 TFLOPS256 TFLOPS
FP32 Compute19.5 TFLOPS51.2 TFLOPS
INT8 Compute624 TOPS512 TOPS
Memory TypeHBM2e—
Memory Capacity80 GB HBM2e64 GB HBM2e/ HBM3e
Memory Bandwidth2.0 TB/s600 GB/s/ 1,200 GB/s
TDP Power400 W310 W

Key Differences​

  • Power: Huawei Ascend 910B has a TDP of 310 W, lower than NVIDIA A800's 400 W, friendlier to datacenter PUE and cooling.
  • Memory capacity: NVIDIA A800 packs 80 GB HBM2e, more than Huawei Ascend 910B's 64 GB HBM2e/ HBM3e, more comfortable for single-card hosting of very large models.

Selection Advice​

  • When chasing extreme single-card compute and a mature toolchain, prioritize NVIDIA A800; if budget, power wall, or local support are hard constraints, Huawei Ascend 910B often fits better. Use this site's AI Compute Card Comparison Tool to validate multiple chips side-by-side before deciding.

FAQ​

What are the main differences between NVIDIA A800 and Huawei Ascend 910B?​

The core difference is architecture and compute density: NVIDIA A800 uses NVIDIA Ampere (GA100), FP8 ~No public FP8 data, memory 80 GB HBM2e; Huawei Ascend 910B uses Da Vinci, FP8 ~No public FP8 data, memory 64 GB HBM2e/ HBM3e. See the comparison table above.

What is the TDP (power) of NVIDIA A800?​

NVIDIA A800 has a TDP of 400 W; actual whole-system power also includes board, fans, and PUE.

Which is better for large-model training / inference?​

Training values memory capacity, bandwidth, and multi-card interconnect; inference values single-card throughput and power efficiency. Combine the "Key Differences" and "Selection Advice" above with your batch size, model size, and SLA.

How much do NVIDIA A800 and Huawei Ascend 910B differ in memory capacity?​

NVIDIA A800 is 80 GB HBM2e, Huawei Ascend 910B is 64 GB HBM2e/ HBM3e; the gap directly affects loadable model size and context length.