Huawei Atlas 300I Duo 96GB vs Huawei Ascend 910A
The Ascend 910A delivers 83% more FP16 throughput than the Atlas 300I Duo (256 vs 140 TFLOPS dense). The Atlas 300I Duo counters with 64 GB more memory (96 GB vs 32 GB).
Huawei Atlas 300I Duo 96GB
Full specs →Huawei Ascend 910A
Full specs →Specifications
Performance (TFLOPS)
FLOPS by Precision
What actually differs
The Atlas 300I Duo is the newer part: Da Vinci, launched in 2022, against the Ascend 910A's Da Vinci from 2019. Newer architectures typically add lower-precision formats and better throughput per watt, so check the precision rows your workload actually uses.
Dense throughput for the Atlas 300I Duo against the Ascend 910A: FP16 140 vs 256 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.
Memory is 96 GB against 32 GB, fed at 408 GB/s versus 1.2 TB/s. More memory per GPU means larger models fit before you have to shard across cards, which often matters more than raw TFLOPS for inference.
For LLM training and inference, weight the FP8 and FP16 rows and memory capacity most heavily. For scientific and HPC workloads, the FP64 row is the one to read.
Frequently asked questions
Is the Atlas 300I Duo faster than the Ascend 910A?
At FP16 precision the Ascend 910A reaches 256 TFLOPS dense against 140 TFLOPS for the Atlas 300I Duo. The performance table on this page lists every published precision for both GPUs.
Which has more memory, the Atlas 300I Duo or the Ascend 910A?
The Atlas 300I Duo carries 96 GB of memory versus 32 GB for the Ascend 910A. Memory bandwidth is 408 GB/s for the Atlas 300I Duo and 1.2 TB/s for the Ascend 910A.
How much power do the Atlas 300I Duo and the Ascend 910A draw?
The Atlas 300I Duo is rated at 150 W TDP and the Ascend 910A at 310 W. On FP32 throughput per watt, the Ascend 910A is the more efficient part.
Get Comparison Updates
New GPUs added weekly. Be the first to see how they compare.